The Long Term Impact of Adolescent Risky Behaviors and Family Environment

Publication Date

Jul 31, 2001

Submitted by:
Michael R. Pergamit, Ph.D. Lynn Huang, Ph.D. Julie Lane, Ph.D.
National Opinion Research Center (NORC)
University of Chicago

Submitted to:
Office of the Assistant Secretary for Planning and Evaluation
U.S. Department of Health and Human Services

Executive Summary

Statement of the Problem

The course of human development is not a series of random events. The lives of adults at any point in time are the result of previous choices and environmental influences. Primarily due to lack of good data, insufficient attention has been paid to the relationship between early life behaviors, the context in which they occur, and outcomes in later adulthood. This report seeks to examine several of these relationships to form a broad basis for further research. Although lifetime outcomes are undoubtedly shaped from birth (if not before), we specifically look at occurrences during adolescence and relate them to a set of adult outcomes.

Adolescence is often a period during which individuals try on new attitudes, roles, and behaviors. Some adolescents choose to engage in risky behaviors. For some, the experience will be one of experimentation, a passing phase. For others, it will be the beginning down a path to problems that follow them into adulthood. Every year millions of dollars are channeled into efforts to curtail adolescent risky behaviors. The premise behind these initiatives is that risky adolescent behaviors put youth in danger for the occurrence of deleterious short- and long-term outcomes. Research to date has tried to explain who is likely to engage in these behaviors and whether they suffer negative consequences. For the most part, the consequences examined are typically short-term. This study is a departure from most of the existing literature in focusing on longer-term adult outcomes. In particular, it is one of the few studies to use a large, nationally representative sample to examine a wide variety of adult outcomes.

We seek to establish whether there is a relationship between engaging in risky behaviors as an adolescent and negative consequences later in life. We explore adulthood along several domains: health, economic success, family formation, and incarceration. We also seek to examine the relationship between family environmental factors and these adult outcomes in the presence of risk taking behavior. Specifically, we examine the roles of family structure, family socioeconomic status (as measured by parents' education), and the presence of an alcoholic parent.

In this report we explore the following questions:

Do youths engaging in risky behaviors face worse outcomes as adults?
Does the relationship between adolescent risky behaviors and adult outcomes vary by the type of behavior and the type of outcome?
What is the relationship between family environment and adult outcomes?
Given that a youth chooses to engage in a risky behavior, does family structure help reduce the likelihood of a bad adult outcome?
Within a given family structure, does socioeconomic status (SES) as measured by parents' education impact the likelihood of a bad adult outcome?

We examine five adolescent risky behaviors: alcohol usage, marijuana usage, cocaine usage, sexual activity, and delinquency. Each of these is measured using age of initiation except for delinquency, which is a measure of the total number of delinquent and/or criminal acts in 1980. A significant contribution of this study is that outcomes are measured well into adulthood and not immediately at or near adolescence. The outcomes we study are measured generally in the late twenties or early thirties.

Data

This study uses data from the National Longitudinal Survey of Youth--1979 cohort (NLSY79). The NLSY79 is a large, nationally representative, omnibus survey sponsored by the U.S. Bureau of Labor Statistics. Over 12,000 youths ages 14-22 were first interviewed in 1979. They have been re-interviewed annually through 1994 and biennially since. The ongoing sample includes over-samples of blacks and Hispanics. The sample has seen remarkably low attrition with over 84 percent having been interviewed in 1998 (the most recent year of data available at the time of this study).

The NLSY79 focuses on labor market behavior with information collected on aspects of the respondents' lives which are thought to influence, or be influenced by, their labor market behavior. The survey routinely collects information on education, job training, marriage, fertility, household composition, health status, income, and assets. In selected years, additional information on such things as alcohol consumption and drug usage has been collected.

The advantage of using the NLSY79 is the availability of measures of long-term adult outcomes in a continuous context. The youngest respondent in the NLSY79 turned 34 in 1998. Over the years of the survey, we see the respondents (generally) complete their education, develop their careers, and form families. Also, we have observations of certain behaviors such as alcohol use and drug use at multiple points in time.

Due to data limitations, we restrict our measures of adolescent risky behaviors to retrospective reports on age of initiation. The one exception to this is in reports of delinquent and criminal behaviors. In this case, questions were included in 1980, when the respondents were 15-23 years old. While the analysis must be limited in certain ways, the NLSY79 strength is its inclusion of a wide variety of long-term adult outcomes.

Main Findings

There is a fairly consistent pattern that engaging in risky behaviors as a teenager is associated with less successful adult outcomes. In most cases, the earlier one engages in the behavior, the more likely one faces a bad outcome as an adult. The most consistent predictor of a bad adult outcome is age of initiation into sexual activity. Alcohol usage, on the other hand, is perhaps the one teenage behavior least associated with bad adult outcomes. Age of initiation into alcohol usage is, however, associated with adult alcohol abuse or dependence. None of our results, including these findings for sex and alcohol initiation should be interpreted as causing the adult outcomes. These are statistical associations, not causal relationships. Age of initiation into any particular risky behavior may be associated with unmeasured adolescent characteristics or circumstances that are related to the transition into adulthood. In other words, there may be a personal or family characteristic which influences both early sex initiation (for example) and a bad adult outcome. By not having measured this relevant characteristic, we would incorrectly attribute the cause of the bad adult outcome to early sex initiation.

Our findings indicate that family structure effects differ by outcome domain measured. Adolescents who reside in intact families at age 14 clearly have the least likelihood of a bad economic outcome and are less likely to spend time in jail. This is less clear for health and family formation outcomes. Adolescents living in single mother headed households at age 14 do not fare as well as those in intact families in economic outcomes, but compare favorably along other domains. Interestingly, the presence of a biological father in the household at age 14 is associated with lower levels of adult alcohol disorders or drug usage. Adolescents living with a single mother or single father at age 14 are less likely to have married by age 33 compared with those who lived in either intact or re-married families.

We found similar associations when examining the relationship of parents' education to long-term adult outcomes. Having more educated parents is associated with better economic outcomes and less likelihood of going to jail. However, there are contradictory results for the health domains where mothers' education is associated with lower likelihood of adult alcohol problems, but greater likelihood of adult drug usage. Fathers' education had exactly the opposite associations across the two outcomes. These family structure and parents' education relationships generally held even when restricting the sample to those who initiated early into risky behaviors. The results for parents' education held even when further restricting the sample to either intact families or single mother households. There is a sense that although the family may not have prevented the youth from starting down a "wrong" path, it can help them from having that choice lead to bad consequences.

Our results suggest that the ways in which parents help prevent bad outcomes for their children differ across different domains. For economic outcomes, parents can use their resources (financial, networks, etc.) to send their children to college, help them get jobs, and serve as a fallback to prevent financial problems. There is probably a similar mechanism for keeping their children out of jail. However, for non-economic domains, particularly those of substance abuse, a different set of family processes contributes to an eventual healthy adulthood.

Acknowledgements

This report was prepared for the U. S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation under Contract # HHS-100-99-0007. Other contributors to the report include S. Courtney Bickert and Alice Lapray. We gratefully acknowledge the thoughtful guidance of ASPE staff Kelleen Kaye, Sonia Chessen, and Meredith Kelsey as well as former ASPE Division Director Matthew Stagner. We also thank Michele Zimowski for helpful comments on an earlier draft. All opinions are those of the authors and do not necessarily reflect the positions of the Department of Health and Human Services or its staff.

Chapter I: Statement of the Problem

Adolescence is often a period during which individuals try on new attitudes, roles, and behaviors. Some adolescents choose to engage in risky behaviors. For some, the experience will be one of experimentation, a passing phase. For others, it will be the beginning down a path to problems that follow them into adulthood. Every year millions of dollars and a multitude of effort are directed toward curtailing adolescent risky behaviors. Examples of these efforts include the recent multi-million dollar media campaign sponsored by the Office of National Drug Control Policy with the intention of reducing the use and abuse of illegal substances among America's youth and the funding of abstinence-only education programs. The premise behind these initiatives is that risky adolescent behaviors put youth in danger for the occurrence of deleterious short- and long-term outcomes. Research to date has tried to explain who is likely to engage in these behaviors and whether they suffer negative consequences. For the most part, the consequences examined are typically short-term such as high school graduation, college enrollment, and teenage pregnancy. A few studies examine more adult outcomes. This study is a departure from most of the existing literature in focusing on longer-term adult outcomes. In particular, it is one of the few studies to use a large, nationally representative sample to examine a wide variety of adult outcomes.

We seek to establish whether there is a relationship between engaging in risky behaviors as an adolescent and negative consequences later in life. We also seek to examine the relationship between family environmental factors and adult outcomes in the presence of risk taking behavior. Specifically, we examine the roles of family structure, family socioeconomic status (as measured by parents' education), and the presence of an alcoholic parent.

We explore adulthood along several domains: health, economic success, family formation, and incarceration. By examining outcomes well beyond the adolescent years, this study provides a background to further study of the pathways through which youths pass to make successful or less than successful transitions into adulthood. It is important to measure outcomes well into adulthood. Adolescent outcomes may in no way represent whether a successful transition to adulthood will take place. Short-term adult outcomes may not represent individual resilience. Also, the measurement of these outcomes may be deceptive. For example, a drug user may not attend college, but instead enter the labor force full time. In the short run, this person may appear to have higher income than those who spend their time acquiring higher education.

Past research has shown that family factors influence the choice to engage in risky behavior as an adolescent. Other research has established that family environment is important for a successful transition into adulthood. Those who grow up in intact families and those coming from higher socioeconomic status (SES) families typically fare better in many dimensions, especially economic. If engaging in risky behaviors as an adolescent has long-term consequences, then one can ask if the family environment's impact on the transition to adulthood is through its influence on risk taking or whether there is an independent effect. In this report we do not model the relationship between family structure and the choice to engage in a risky behavior. However, unlike other literature, we allow engaging in risky behaviors as an adolescent to have a direct impact on adult outcomes. We contribute to the literature by exploring the relationship between adolescent family environment and adult outcomes while holding constant the adolescent's risk taking behavior.

Given a set of relationships between adolescence and adulthood, the natural step is to investigate the pathways between these two points. Why do some youths use drugs or engage in sexual activity and go on to lead successful lives while others encounter problems? Given the choice to engage in a behavior that has a higher likelihood of bad outcomes, what factors might restore the individual to a path to a successful transition? In a later chapter of this report, we address this question by focusing exclusively on those who have chosen to initiate into risky behaviors at early ages. Early initiators into risky behaviors are at the most risk for deleterious adult outcomes. Essentially, they have started down the "wrong" path. We re-introduce family structure and SES to address whether family environment matters for these early initiators and contributes to the successful avoidance of the potential downside effects of these behaviors. The findings of this chapter should lay the groundwork for future research into the factors that influence the pathways that lead to a successful transition from adolescence to adulthood.

In this report we explore the following questions:

Do youths engaging in risky behaviors face worse outcomes as adults?
Does the relationship between adolescent risky behaviors and adult outcomes vary by the type of behavior and the type of outcome?
What is the relationship between adolescent family environment and adult outcomes?
Given that a youth chooses to engage in a risky behavior, does family structure help reduce the likelihood of a bad adult outcome?
Within a given family structure, does socioeconomic status (SES) as measured by parents' education impact the likelihood of a bad adult outcome?

The next chapter of this report reviews the literature associated with adolescent risk taking and family environment, with particular attention to their relationships with adolescent and adult outcomes. Chapter III describes the data and methods used in our analysis. One of the major strengths of the analysis is the use of the National Longitudinal Survey of Youth--1979 (NLSY79), a large nationally representative sample of individuals who have been interviewed regularly since they were adolescents in 1979. The NLSY79 is an omnibus survey, rich in details about the lives of these individuals during their adolescence and well into their adulthood. It offers multiple measures of adolescent risky behaviors and adult outcomes. Chapter IV provides context for how the measures of adolescent risky behaviors relate to the family environment measures for the NLSY79 sample.

The main analysis of the study begins in Chapter V. In this chapter, we report results of regressions relating adult outcomes to adolescent risky behaviors, family structure, parents' education, and parental alcoholism. Given a general conclusion that early initiation into risky behaviors is associated with poor adult outcomes, we turn in Chapter VI to focus on those who have chosen to initiate into one or more risky behaviors at early ages. These adolescents have chosen to follow a path that has a higher likelihood of negative consequences. We examine whether family structure and parents' education are associated with the adult outcomes obtained by these early initiators. That is, can family environment help a youth headed down the "wrong" path avoid the negative consequences associated with that path. This analysis is intended to lay the groundwork for further studies of the pathways adolescents follow in their transition to adulthood. We conclude the report in Chapter VII with a discussion of the main implications of our findings and a set of future directions for continued analysis.

Chapter II: Literature Review

In this chapter, we begin with an examination of the literature on youth risky behaviors. Although the purpose of this report is to focus on the adult consequences of youthful risky behavior, it is important to put this into perspective by examining literature on the risk and protective factors for adolescent risky behaviors and the adolescent outcomes of these risky behaviors. Literature on the relationships of three family variables--family structure, socioeconomic status (parental education), and parental alcoholism--with successful or failed adult outcomes is also reviewed.

A. Predicting Adolescent Risky Behaviors

Much of the research examining adolescent risky behaviors is centered on the factors that predict or co-vary with their occurrence. Risk factors are those variables that increase the likelihood that a certain negative outcome, in this case, risky adolescent behavior, will occur. Protective factors buffer the influence of risk factors on outcomes. The literature covers a vast array of risk factors that are thought to act as precursors to adolescent risky behavior. Examples of risk and protective factors related to the adolescent risk behaviors we examine in this report--early sex, alcohol, marijuana, and cocaine initiation and delinquency--are discussed in the rest of this section. The literature is diverse in terms of the factors studied, but a consistent set of relationships emerge.

Using data from the National Longitudinal Study of Adolescent Health (Add Health), Resnick, et al. (1997) examined the relationships of family, school, and individual risk and protective factors with the adolescent risky behaviors of suicide ideation and attempts, violence, cigarette use, alcohol use, marijuana use, age of sexual initiation, and pregnancy history. Among the many results reported, there was evidence that low grade point average and being held back a grade in school were associated with more substance use and sexual behavior. Family-related variables, such as parent or family connectedness, as well as school connectedness served as protective factors against all adolescent risk behaviors except for pregnancy. Parental attitudes also played a protective role in initiation of sex. Parents who were more disapproving of early sex initiation were more likely to have children with later age of onset of sexual behavior.

Review of the male delinquency and crime literature by Loeber and Dishion (1983) and Loeber and Stouthamer-Loeber (1987) revealed several factors that stood out as predictors of male offending. The important variables that were linked to higher levels of male offending were poor parental child management style, childhood antisocial behavior, parental and sibling criminality, low intelligence, low educational attainment, and separation from parents.

Analysis of risk and protective factors data from the youth module of the 1997 National Household Survey on Drug Abuse (NHSDA) showed a few factors that were particularly predictive of past-year marijuana use in adolescents (OAS, 2001). Controlling for demographic and other factors related to drug use, logistic regression analyses showed that the variables with the est relationship to past year use of marijuana were easy availability of marijuana, perceptions of close friends' more positive attitudes toward monthly marijuana use, actual marijuana use by close friends, and perceptions of low risk from marijuana use. For past-year alcohol use by 12 to 17 year olds, a slightly different set of factors arose in the final model. Past-year shoplifting, perceptions of parents' more negative attitudes toward binge drinking weekly, and anyone offering marijuana had the est relationship with alcohol use, although the predictive power of variables in this model was small in comparison with those from the final model predicting marijuana use.

Kosterman, et al. (2000) also found some similarities and some differences when comparing risk and protective factors for alcohol and marijuana initiation. Respondents in their Seattle Social Development Project were initially surveyed as fifth-graders and then again in almost every subsequent year until they were 18 years old. These researchers found that peer use of a given substance was directly predictive of initiation of that substance, for both alcohol and marijuana. For alcohol initiation, parents' alcohol use norms served as a protective factor. For marijuana initiation, parents' proactive family management was the key protective factor. Respondents own personal norms for substance use were predictive of marijuana initiation but not for alcohol initiation.

There is some evidence that the risk factors for initiation of risky behaviors may be distinct from risk factors for more regular use or abuse of these behaviors (Scheier and Newcomb, 1991). Weber, et al. (1989), for example, categorize two distinct pathways of adolescent alcohol use. In this view, normally socialized adolescents consume alcohol at a more steady pace while those who are "problem prone" show more rapid acceleration of alcohol involvement after initiation occurs. Scheier, et al. (1997) suggest that social learning factors such as peer and adult models and normative expectations are important ingredients in predicting initial stages of adolescent alcohol use. Personality components may be a key part of alcohol abuse later in young adulthood. These researchers found that several psychological factors--behavioral control, depression, anxiety, external locus of control, antisocial behavior, and low self-esteem--were significant predictors of alcohol consumption and change in drinking patterns from onset to more problematic drinking.

The Problem Behavior Theory presents one way of categorizing the risk factors predictive of adolescent risky behavior. Leading theorists, Jessor, Donovan, and Costa (1991) describe the three major systems of psychosocial risk and protective factors that are responsible for occurrence of risky behavior as the personality system, the perceived environment system, and the behavior system. When the variables in a given system are geared up for the occurrence of a problem, that system is in a state of proneness. When all three systems are in this state, then an individual shows overall psychosocial proneness toward a particular problem behavior.

This approach implies that adolescent risky behaviors such as early substance use, precocious sexual behavior, and delinquency are symptoms of an underlying trait (Jessor and Jessor, 1977). Using longitudinal data, Donovan, Jessor, and Costa (1988) concluded that a single common factor was responsible for the positive associations among a number of adolescent antisocial behaviors, including problem drinking, marijuana use, precocious sexual intercourse, and delinquency. Similarly, Patterson (1993) discusses a core antisocial trait that makes its appearance in various risk behaviors. Problem Behavior Theorists refer to these risky behaviors as a syndrome of problem behavior with a high degree of interrelatedness among the behaviors.

In their comprehensive review of the literature on predictors of adolescent drug abuse, Hawkins, et al. (1992) came up with a classification scheme that helps break down the overwhelming number of risk factors into manageable categories. They divided risk factors into two main groups: contextual and individual/interpersonal. Contextual factors include laws and norms encouraging substance use, easy availability of a substance, economic disadvantage, and neighborhood disorganization. Individual/interpersonal factors consist of: physiology, family substance use and attitudes, poor and inconsistent family management, high family conflict, low family bonding, early and persistent problem behaviors, academic failure, low commitment to school, peer rejection early in school, affiliation with substance-using peers, rebelliousness and alienation from society's values, pro-drug use attitudes, and early initiation of substance use. From their review, Hawkins, et al. drew some general conclusions about risk for drug abuse. Among their conclusions they reported that risk factors showed consistency over time. The same risk factors have been identified for different cohorts. They also determined that the greater number of risk factors, the greater the risk of drug abuse.

B. Adolescent Consequences of Risky Behaviors

Most literature on adolescent consequences of risky behaviors focuses on risky behaviors as outcomes, with some studies examining other outcomes such as educational attainment and employment. Risky behaviors beget other risky behaviors. The fact that adolescent risky behaviors often co-occur makes engaging in any one risky behavior a risk factor for engaging in another. Lindberg, Boggess, and Williams' (1999) results somewhat support the interconnectedness of adolescent risk behaviors. These researchers studied the co-occurrence of adolescent risk behaviors using data from 7th to 12th graders in the 1995 Add Health survey. They found that 28 percent of 7th to 12th graders indicated engaging in two or more risk behaviors, including regular or recent substance use, fighting, carrying a weapon, suicidal thoughts and ideation, and unprotected sexual intercourse. More than one-quarter of students, though, only participated in a single risky behavior. For all of the risk behaviors they examined, except for one, at least three-quarters of the students engaging in it were also involved in another risk behavior.

In a study using two large national data sets, the Youth Risk Behavior Survey (YRBS) and Add Health, the Center on Addiction and Substance Abuse (CASA 1999) found that teenagers who consume alcohol or take illicit drugs are more likely to engage in sex, to do so at a younger age, and to have several partners. For adolescents who are 14 and younger, consuming alcohol or using drugs doubles and quadruples, respectively, the likelihood that sexual intercourse has ever been experienced compared to adolescents who have never used these substances. Moore, et al (1995) reported that early onset of alcohol, tobacco, and other drugs; school problems; delinquency; and physical aggression are significantly associated with early onset of sexual behavior. Alcohol use in adolescence has also been found to be related to more frequent sexual activity and less frequent use of condoms (Cooper, Peirce, and Huselid, 1994).

Another example of adolescent risky behaviors occurring in conjunction with each other comes from research using the 1998 NHSDA (OAS, 2000). In this report, about 40 percent of alcohol users age 12 to 17 were also current illicit drug users. This percentage increased to 58 percent of binge drinkers and 69 percent of heavy drinkers who had used an illicit drug in the past month. Past-year criminal activity also co-occurred with substance use. Adolescents who had consumed alcohol in the previous month were more likely than those who had not to also have engaged in criminal activity in the past year. In general, the heavier the alcohol use, the greater the probability of criminality. Adolescents who drank in a heavy manner were more likely than lighter drinkers to be involved in a variety of delinquent acts and to show aggressive behavior, such as physically attacking people or destroying property.

Johnson, Arria, et al (1995) identified a connection in preadolescence between early, unsanctioned alcohol use (without permission from their parents) and higher levels of conduct problems. Earlier alcohol use was also associated with accelerated growth of conduct problem behaviors during the transition to early adolescence. The literature on adolescent conduct problems and substance use supports the notion of an "externalizing (behavioral) path" leading to substance use, particularly for males (Hussong, Curran, and Chassin, 1998). Windle (1990) found that even when holding early substance use constant, early adolescent delinquency predicted later substance use.

Each adolescent risk behavior also has some very specific consequences. We present these consequences, categorized under the relevant adolescent risk behavior.

Risky sexual behavior

The logical concerns arising from adolescent sexual behavior are pregnancy, parenthood, infection with a sexually transmitted disease, and exposure to the human immunodeficiency virus (HIV). Individuals who begin having sex at earlier ages are exposed to risk for a greater length of time, are less likely to use contraception, have more sexual partners, and are involved in high risk sexual behavior, such as substance use before intercourse (Moore, et al., 1995). Moore, et al also reported that another possible concern about early sexual behavior is that first sexual experiences are often coercive. An astonishing majority of first sexual experiences that occurred before age 15 among females were not voluntary. Coercion is damaging in itself, but it is also associated with improper or no use of contraception.

Delinquency

Conduct problems in childhood and early adolescence have been found to be associated with substance use problems later in adolescence (Lynskey and Fergusson, 1995; Windle, 1990). Results from Bergmark and Andersson's (1999) longitudinal study that followed Swedish participants from childhood to adulthood revealed that the earlier the conduct problems for boys, the more frequent the occurrence of adolescent drunkenness. Other research has found a significant relationship between delinquency and school attachment. Liska and Reed (1985) used data from the first two waves of the Youth in Transition study to examine this issue. They found that attachment to school has no influence on adolescent violence, but the reverse of this relationship, that violence decreases school attachment, was significant.

Substance Use

A possible consequence for adolescents who engage in substance use behavior is that this risky behavior can lead to increased, problematic use of a given substance. Early initiation of alcohol use was predictive of later adolescent problem drinking (Fergusson, Lynskey, and Horwood, 1994; Hawkins, et al, 1997; Pedersen and Skrondal, 1998). Kosterman, et al. (2000) found a connection between age of initiation of alcohol use and alcohol misuse later in adolescence. Hawkins, et al. (1997) also found that the younger the age of alcohol initiation, the greater the level of alcohol-related problems in late adolescence (see also Gruber, et al 1996). In this study, age of initiation served as a mediator of effects of ethnicity, parents' alcohol consumption, proactive parenting, school bonding, friends' alcohol initiation, and perceptions of alcohol's harmfulness on alcohol misuse in late adolescence (Hawkins, et al., 1997). In other words, when age of initiation was entered into the model predicting adolescent alcohol misuse, these formerly significant variables no longer showed any effect on the outcome. Fergusson, Lynskey, and Horwood (1997) reported that onset of marijuana use before age 15 had a relationship with later marijuana use.

There is also legitimate concern that adolescent substance use can lead to use of "harder" substances. Yamaguchi and Kandel (1984) provided evidence for a gateway model of drug use. In their model, use of softer substances, such as alcohol and cigarettes, open the gates to the use of marijuana, which in turn makes the use of other illicit drugs more likely.

Earlier adolescent marijuana use has been found to increase the risk of a variety of negative outcomes in later adolescence. Early marijuana use is predictive of not graduating from high school; delinquency; mental health problems; having multiple sex partners; inconsistent condom use; perception of drugs as not harmful; having problems with alcohol, tobacco, and other drugs; and having deviant friends (Brook, Balka, and Whiteman, 1999; Fergusson and Horwood, 1997). Fergusson and Horwood (1997) examined the issue of marijuana use and adjustment in adolescence with data from a longitudinal study in which a sample of New Zealand children were surveyed regularly from birth to age 18 years old. Even after controlling for childhood, family, and other potential risk factors, those respondents who indicated early marijuana use, especially more frequent use, were more likely to have a marijuana use disorder, use other substances, be unemployed, engage in delinquent acts, and drop out of school early by age 18.

C. Adult Consequences of Adolescent Risky Behaviors

Although a large number of studies examining consequences of adolescent risky behavior look at these consequences in adolescence, some research has focused on adult outcomes. The relevant research is presented below according to the type of risky adolescent behavior investigated.

Delinquency

Childhood delinquency has been found to be related to various negative adult outcomes. Research points to a distinction between short-term delinquent behavior that the individual outgrows and "career" criminality. Sampson and Laub (1990, 1993), in some of the most comprehensive studies on this subject to date, used data from the Gluecks' (1950) study, which looked at 500 delinquent boys and 500 non-delinquent boys born between 1924 and 1935 and followed them for 18 years. They found that, in adulthood, delinquents were more likely to have various negative outcomes including: charged with offenses in the military, excessive alcohol use, general deviance, arrest, economic dependence (welfare), unstable employment, divorce and separation. They were also less likely to have graduated from high school. Sampson and Laub also determined that job stability in young adulthood, commitment to educational and occupational goals, and attachment to spouse all have a large inverse relationship with measures of adult crime and deviance, and are predictive of later behaviors. Delinquents who later develop social bonds, such as attachments to spouses or work, tended to have fewer problems in adulthood than did other delinquents. While there is substantial evidence that criminal behavior continues, social ties in adulthood can explain changes in criminality over the life span.

Males with childhood conduct disorder are more likely than other males to have antisocial personality disorder as adults and to suffer with alcohol and drug dependence (Offord and Bennett, 1994). They are also likely to commit more crime in adulthood and are more likely to suffer premature death (Kratzer and Hodgins, 1997). Among women, those with a conduct disorder are more likely to have an internalizing (emotional) psychiatric disorder as adults (Offord and Bennet, 1994). Like men, they also are more likely to commit crimes and abuse substances as adults; however, most girls with conduct disorders do not experience any of these problems. When assessing the links between delinquency and adult crime, Robins (1978, p. 611) notes, "[The diagnosis of] Adult antisocial behavior virtually requires childhood antisocial behavior [yet] most antisocial youths do not become antisocial adults". Although adult crime rates may be higher for those who were involved in delinquency, most delinquents do not commit crimes as adults.

Bardone, et al (1998) found that, after controlling for numerous confounding variables, conduct disorder at age 15 predicted several health outcomes at age 21 including more medical problems, lower self-reported overall health, lower body mass index, alcohol and/or marijuana dependence, tobacco dependence, daily smoking, more lifetime sexual partners, sexually transmitted disease, and early pregnancy. Studies examining the predictive power of childhood aggression have determined that children rated as aggressive at ages 8-10 were more likely to be rated as aggressive at age 32. Aggression in childhood also predicted conviction of a violent crime, unemployment, use of tobacco and illicit drugs, and driving while intoxicated. This predictive power decreased over time, but was still significant at age 32 (Farrington, 1991).

Vitelli's (1997) study of prison inmates found that those who were early starters (first arrest before age 14) had higher rates of substance abuse than late starters or inmates who were never juvenile delinquents. Late starters, however, had a higher rate of lifetime violence. Females and males who were identified as juvenile delinquents were also significantly more likely to perpetuate abuse in intimate relationships (Giordano, et al 1999). Although the specific longitudinal links between delinquency and adult outcomes have not been fully elucidated, Hagan (1997) has shown that involvement in delinquent subculture causes strain on school and work roles, which increase the likelihood that delinquents drop out of high school. Although school dropout did not appear to negatively affect them when interviewed in their early 20s, by the time they reached their mid-30s, they were more likely to be unemployed and to feel despair. Monk-Turner (1989) also looked at education and found that holding several background variables constant, high school delinquents complete fewer years of schooling. However, she also showed that, after controlling for years of schooling and other background variables, involvement in delinquency during high school did not significantly shape adult occupational status.

The few economic studies that explore the critical issue of inter-temporal linkage between youth risky behaviors and adult outcomes produce somewhat mixed results. Anderson, Mitchell, and Butler (1993) studied the effect of deviance during adolescence on the choice of jobs as adults. They analyzed data from the Epidemiologic Catchment Area Program surveys to ascertain whether deviance during adolescence increases the likelihood an individual will develop mental health disorders in adulthood and simultaneously has negative effects on educational attainment. They also investigated whether deviance during adolescence has indirect effects, through education and mental health disorders in adulthood, on the probability of working and occupational choice. Their results indicated that deviance during adolescence has significant negative effects on future labor market outcomes. Levitt and Lochner (2000) attempted to explore the effect of criminal participation status at young ages on educational outcomes, labor market outcomes, and family measures at age 30 using the National Longitudinal Survey of Youth-1979 cohort. There was only a small negative correlation between youth crime and adult work, which might be the result of unobserved heterogeneity. There were significant differences in educational attainments between criminals and non-criminals. They found no difference in marriage and fertility patterns.

Substance Use

Early substance use has been associated repeatedly with later substance misuse in adulthood. Using 1988 National Health Interview Survey (NHIS) data, Chou and Pickering (1992) found that early onset drinking poses increased risk for lifetime alcohol-related problems. Having a first drink at age 15 or younger increased the odds of later having 3 or more alcohol-related problems, which are similar to criteria for alcohol dependence. A delay in drinking until age 20 or 21 sharply reduced risk of developing alcohol-related problems. Grant and Dawson (1997), using cross-sectional data from the nationally representative National Longitudinal Alcohol Epidemiological Survey (NLAES), found for each additional year that passed before initiation of drinking, the risk for development of alcohol dependence and alcohol abuse decreased by 14 percent and 8 percent, respectively. ¹ Even when controlling for family alcoholism, earlier age of initiation into alcohol consumption was associated with increased likelihood of alcohol dependence (Grant, 1998). In a longitudinal study, Guy, Smith and Bentler (1994) found that a general drug use factor in adolescence predicted drug use 12 years later in young adulthood. This study helps to confirm the idea that there is some stability of drug use across adolescence and young adulthood.

Prescott and Kendler (1999), using twin study structured psychiatric interviews, also found evidence for an association between early drinking onset and risk for alcohol dependence but less evidence for an association with alcohol abuse. They suggest that the relationship between alcohol initiation and diagnosis of alcohol dependence is non-causal so any attempts to prevent dependence by delaying the onset will probably not work. They argue that both early initiation of alcohol use and adult alcohol dependence are manifestations of vulnerability. Their shared vulnerability hypothesis claims that both behaviors tap into the underlying dimension of proneness to problematic alcohol involvement.

Jessor, et al.'s (1991) study revealed a relationship between youthful substance use and some, but not all, adult outcomes. These researchers studied the impact of a large array of risky adolescent behavior on adult outcomes with the use of longitudinal data drawn from two samples of young people in a single city. One sample consisted of junior high students first surveyed in 1969 and followed up several times until 1981 when they were ages 25-27 (the High School Study). The other group was made up of college freshman first surveyed in 1970 and then five more times in the next 11 years (the College Study). Their results provided support for the Problem Behavior Theory's relevance to problem behavior in young adulthood, rather than just in adolescence. Roughly the same proportion of variance was accounted for in explaining adolescent and young adult problem behavior using psychosocial variables related to this perspective.. Results provide evidence for a syndrome of problem behavior in young adulthood as well as in adolescence. Results showed that involvement in problem behaviors in adolescence was related to later engagement in problem behaviors in adulthood.

Outcome measures other than those related to problem behaviors were also examined, namely educational and occupational attainment. Deviant behaviors (behaviors that violate societal or legal norms apart from substance use) in adolescence were related to later decreased educational attainment (for all except the College Study women). For the High School Study men and women, multiple problem behaviors in adolescence were also predictive of decreased educational attainment in adulthood. They found no significant relationship between measures of psychosocial proneness from youth and later occupational attainment (composite measure of occupational prestige). When education was held constant, psychosocial proneness to problem behavior did account for variation in occupational attainment for the High School Study men.

Newcomb and Bentler's (1988) "precocious transitions" theory provides a challenge to Problem Behavior Theory in that the latter would suggest that the effects of adolescent drug use should be the same as the effects of general proneness to deviance. In an important study, Newcomb and Bentler found very specific consequences connected to different types of substances, rather than just one typical adult outcome. Newcomb and Bentler found that, consistent with their theory, illicit drug use during adolescence speeds up the typical developmental process and forces the adolescent user into adult roles without appropriate skills to handle them. An important step of development is missed when an adolescent is not allowed to gain any practice at these new roles. These "precocious" movements into adult roles increase the likelihood of failing at these roles in the long run.

Newcomb and Bentler (1988) approached the issue of adolescent risky behavior's connection with adult outcomes using a longitudinal design in which they surveyed junior high school students from 11 schools in Los Angeles County through their early twenties (1976-1984). These researchers were concerned primarily with the adolescent risky behaviors of alcohol and illicit drug use. In this highly influential study, they found a multitude of adult consequences associated with risky teenage behavior.

In their study, general drug use (including alcohol, marijuana, and hard drugs) as an adolescent directly decreased the likelihood of attending college, but was associated with an increase in income in young adulthood. In addition, general drug use in adolescence was directly related to job instability. A few of the other young adulthood outcome variables of which adolescent general drug use was predictive were: earlier marriage, earlier childbirth, divorce, involvement in drug crimes, stealing, and psychoticism.

Each specific type of adolescent drug use revealed a different network of associations with young adult outcomes. In their final confirmatory factor analysis (CFA) model predicting adult outcomes from 12 substance use measures, hard drug use as a teenager was related to earlier family creation, less likelihood of graduating from high school, increased income, adult suicidal ideation, loneliness, and less social support. Marijuana use as an adolescent was positively associated with the number of times unemployment compensation was collected as an adult. Hashish use as an adolescent was only significantly related to more stealing episodes and job instability later in life. The influence of adolescent alcohol use appeared to work substantially differently than other substance use. Adolescent alcohol use apart from general drug use predicted earlier marriage. Adolescent alcohol use also predicted lower levels of property crimes, confrontational acts, loneliness, and college involvement; greater likelihood of full-time employment or being in the military; more social support; and happiness with one's sex life. They reported that adolescent alcohol use was also related to decreased social conformity and religious commitment in adulthood. These researchers speculated that early use of alcohol may help individuals become part of a social network by reducing inhibitions, thus enabling them to learn appropriate social competencies.

Newcomb and Bentler (1988) reported that cocaine use during early and late adolescence was associated with increased number of relationships, increased number of aggressive or confrontational acts, reduced number of theft episodes, reduced degree of happiness with being close to someone and increased chances of divorce in young adulthood. In another report, Newcomb and Bentler (1993) found that frequency of cocaine use in young adulthood was uniquely predicted by early illicit drug use and late adolescent alcohol use. The only unique outcome from adolescent cocaine use was dealing cocaine later in life.

Teenage drug use in Newcomb and Bentler's (1988) study showed very few direct effects on the young adult sexual behavior and relationship variables for women. The impact of the general drug use factor on young adulthood sex outcomes was mediated through teenage sexual behavior and social conformity. For women the only direct effects from adolescence to young adulthood were: adolescent hard drug use related to less happiness with being close to someone and greater number of relationships, and alcohol use related to greater number of steady partners. Similar to the results for women, there was a lack of a direct influence of early drug use on young adult male sex and relationship outcomes. The effects of drug use were again mediated through social conformity and early sexual involvement. The only direct effect of drug use on young adulthood sex and relationship variables for men was that cannabis use as an adolescent was related to an increased number of steady partners in adulthood.

Other researchers have also found evidence for a link between early substance use and precocious transitions to adult roles. Krohn, Lizotte, and Perez (1997) collected data using a 10-year wave panel study (Rochester Youth Development Study), starting with seventh and eighth graders. For females in their study, early substance use was associated with parenthood, living apart from parents, and the total number of precocious transitions. Early alcohol and drug use for males was related to getting someone pregnant, becoming teenage parents, dropping out of school, leaving the parental home, and accumulation of precocious transitions. In turn, those respondents who experienced off-time transitions to adult roles were more likely to be engaging in substance use in their early twenties, even when controlling for earlier substance use, peer use and other related variables.

Sexual Behavior

Another adolescent risk behavior that was examined as part of Newcomb and Bentler's (1988) study was early sexual involvement. Early sexual involvement ly predicted many outcome measures in young adulthood. Early sexual involvement for women was correlated with more dating competence, increased number of relationships, more frequent intercourse, greater likelihood of an abortion, and greater likelihood of contracting venereal disease. All of these outcome variables were also correlated with men's early sexual involvement (except for occurrence of an abortion), and additionally, less effectiveness of birth control and greater satisfaction with intimacy.

Precocious sexual behavior often has as its consequence teenage pregnancy, which in turn is related to several negative adult outcomes. Teenage parents, in comparison to their counterparts, are more likely to receive less education, be poor and receive welfare as adults (Hayes, 1987; Rosenheim, 1992). There is some evidence though that a particular subset of women may actually find some success later in adulthood even when following an alternative life course to the traditional route of high school graduation, employment, marriage, and childbirth (Furstenberg, Hughes, and Brooks-Gunn, 1992; Hamburg and Dixon, 1992). African American young women who are from extremely disadvantaged situations and who see few available employment options may start this series of life events with childbirth. kinship networks appear to be a necessity for success in this approach, functioning to support the young woman in establishing a household and caring for children. If child bearing is complete by age twenty, the young woman following this path can enter the work force at a young age without taking time off for childbirth or paying for childcare.

D. Family Structure

ast research has emphasized the important role that family-related variables play in the prediction of various adolescent risky behaviors (e.g., Hawkins, et al., 1992; Kandel, 1996). It also seems likely that the impact of adolescent risk factors on adult outcomes is influenced, for the better or worse, by the adolescent's family unit. Family structure research has shown that divorce can be a major force in shaping children's lives. Marital disruption is associated with cognitive, emotional, and behavioral problems and lowered academic achievement in children who have undergone the dissolution (Amato and Keith, 1991a; Hetherington, 1989; Wallerstein, 1988). Moore, et al. (1995), for example, reported that disruption of parents' marriage and living with a single parent are related to earlier onset of adolescent sexual behavior. These researchers speculate that this finding may be explained by lower family incomes, disadvantaged neighborhoods, less supervision and parental modeling, and more permissive attitudes in single parent families.

In their meta-analysis of the influence of divorce on children's adjustment, Amato and Keith (1991a) discovered that effect sizes were modest (mostly due to the large variability within any type of family structure), but the largest impact occurred in the arena of behavior problems. Children of divorce were twice as likely as children from intact families to display outcomes of dropping out of school, teenage pregnancy, teenage idleness, and truancy. Wells' and Rankin's (1991) meta-analysis focusing on the relationship of broken homes and delinquency, clarified this issue further. They found that broken homes were indeed more likely than intact homes to have delinquent adolescents. They found that the association between family structure and delinquency was er for more minor offenses rather than serious types of crime (see also Nye, 1958; Rankin, 1983; Wilkinson, 1980). They also could find no consistent evidence that step-parent families were more likely than single-parent families to include delinquents.

Although there is some debate as to the extent of repercussions for children of divorce ², it is generally accepted that most children whose parents divorce eventually grow into relatively well-functioning adults (Hetherington and Clingempeel, 1992). There are, however, in some cases, consequences from parents' divorce that will result in adjustment problems even in adulthood. Adults from divorced families of origin have shown more behavior problems, lower feelings of well-being, lower socioeconomic attainment, higher marital instability and divorce, and more difficulties in workplace and family relationships (Amato and Keith, 1991b; Amato, Loomis, and Booth, 1995; Booth and Amato, 1994; Hetherington, 1999; McLanahan and Sandefur, 1994). On the other hand, some research finds no lasting substantive effects of divorce into adulthood. Lang and Zagorsky (2001), for instance, using 1979-1993 NLSY79 data, found that once background characteristics are controlled for, the influence of having an absent parent during childhood on adult economic attainment is not substantial. The exceptions to this rule are the significant effects that a father's presence has on adult sons' and daughters' cognitive performance and education and that a mother's presence has on these outcome variables for adult daughters.

Amato (1999) presents research consistent with the idea that the negative effects of divorce may last into adulthood, drawing on life course and risk/resiliency perspectives. A basic assumption of the life course approach is that the effects of one's family of origin are long-lasting--influencing an individual even after he or she leaves the fold (Elder, 1994). The risk and resiliency perspective holds that children's reactions to a stressful event such as parental divorce can be substantially shaped by the amount and the quality of resources in their lives. Some children are more resilient than others in the face of life's stressors. The idea that parental divorce can have long-term effects on individuals is consistent with both approaches.

Amato (1999) compared adults from various types of family structure on four measures of adjustment: socioeconomic, marital quality, relationship with parents, and subjective well-being. Analysis of education and employment measures from the National Survey of Families and Households showed that parental divorce decreased educational attainment for white men and women and black women by about one half year. The disparity in earnings for white men whose parents divorced compared to those whose parents remained married was about $4000 less every year; this difference was $2000 for white women. Education differences accounted for most of this financial disproportion. These results were consistent with other studies that have found children or young adults from broken homes more likely than individuals from intact families of origin to drop out of high school, to not attend college, to be unemployed, and to be at a comparative financial disadvantage (Keith and Finlay, 1988; Krein, 1986; McLanahan and Sandefur, 1994; McLeod, 1991; Wadsworth and McLean, 1986).

Amato (1999) used data from the Marital Instability Over the Life Course study which sampled married adults over a 12 year period and found that adults who were raised in happy intact families were the happiest with their own marriages. Individuals from unhappy intact and divorced families of origin reported the highest levels of conflict and instability in their own marriages. Previous research has shown that the likelihood for divorce is the lowest when neither partner comes from a divorced family of origin and the likelihood for divorce is highest when both spouses come from this type of family (Amato, 1996). In other findings, Amato (1999) found that having divorced parents compared to having an intact family of origin weakened the relationships with parents, especially with fathers. Same-sex bonds (father/son and mother/daughter) did not show this pattern if the divorce occurred during late adolescence. One final realm--well-being--also showed signs of being affected by divorce in one's family of origin. Happiness and satisfaction levels were lowest of all for adults from divorced families. Related research has found that, even after controlling for pre-divorce measures of behavioral and academic problems, parental divorce is related to psychological problems in adulthood (Chase-Lansdale, Cherlin, and Kiernan, 1995).

McLanahan's (1999) findings are consistent with Amato's. In her examination of labor market detachment (neither working nor in school) using several large, nationally representative surveys, she found that young men at ages 23 to 26 years old from single-parent families were about 1.5 times as likely to be out of school and out of work as men from intact homes. This same difference in detachment existed even when dropout rates are held constant. Blacks were especially sensitive to the effects of family structure on labor market success, with growing up in a single-parent home increasing idleness incidence rates by 40 percent for blacks and only 30 percent for whites.

McLanahan (1999) also examined the relationship between family structure and early family formation among women. The results showed that the proportion of young women who experienced teenage pregnancy was significantly higher for respondents coming from a non-intact family compared to women from two-parent households.

NLSY79 data revealed that white females from non-intact families compared to their counterparts had a 14 percentage point greater risk of becoming a mother during the teenage years. For white women from families with higher SES, the risk of teenage pregnancy was 5 times greater for those with family disruption; for blacks from higher SES families, the risk was twice as high. An interesting note to this research is that neither the amount of time a child spent in a non-intact family nor the timing of the marital dissolution had any effect on the long-term consequences studied. Remarriage also did not seem to make a difference in the child's likelihood of avoiding teenage parenthood.

McLanahan (1999) posits three main reasons that children growing up in a single-parent home suffer in comparison to those from intact homes: (1) fewer financial resources, (2) less available time and energy to monitor and care for children, and (3) reduced access to community resources that can act as an extra support to parents. Furstenberg and Cherlin (1991) also emphasized the importance of economic disadvantage in powering the negative effects of divorce on children. McLanahan (1999) suggested that economic instability is the culprit that accounts for half of the disadvantage that often accompanies being raised in a single-parent home. Parenting factors may be responsible for half the increased high school dropout rate among children from one-parent homes. Adjusting for parental resources did not make much of a difference in accounting for teenage birth risk, but it did entirely close the disparity between those from intact and single-parent families in the area of labor force detachment.

It is interesting to note that even though remarrying is likely to put intact families and reconstituted families in the same financial position, this benefit does not translate into equal footing for the children in other areas (Furstenberg and Cherlin, 1991; Hanson, McLanahan, and Thomson, 1996; Hetherington, 1993). McLanahan and Sandefur (1994) reported that often stepchildren living in blended families do not outperform children from single-parent families, and in some cases they even perform worse. The lack of a positive effect and sometimes even a harmful effect of remarriage on educational outcomes has been shown repeatedly in the literature (Boggess, 1998). Wojtkiewicz' (1993) found that length of time exposed to a stepfather was related to lowered probability of graduating from high school. Ginther and Pollack (2001) similarly found that, in general, children from stable blended families had shown less positive educational outcomes than children in intact families. Children from blended families had outcomes similar to children from single-parent families. These researchers also found that within stable blended families there is no significant difference between the performance of stepchildren and the biological children.

E. Socioeconomic Status

Family socioeconomic status touches many aspects of an adolescent's life. The general idea that socioeconomic status has far-reaching influences can be seen in the sheer variety and number of studies in which it serves as a background factor. Socioeconomic status of family of origin can affect factors ranging from community or neighborhood characteristics to types of discipline used (Avenevoli, Sessa, and Steinberg, 1999).

The status attainment literature has focused on determining factors that impact achieving educational, occupational, and other socioeconomic success in youth and adulthood. Hill and Duncan (1987) found that parents' education, especially father's education, as a measure of socioeconomic status, plays an important part in children's educational attainment. However, Sewell and Shah (1967) provided evidence that even though the majority of lower class students did not show high levels of educational attainment, that some students managed to "make it." They also found that mother's education was a more important factor than father's education in predicting educational attainment for women from lower class origins.

More recently, Krohn, et al. (1997) found that being from a lower class family of origin was associated with precocious transitions. In their study, if the household was on welfare, under the poverty line, or if the household primary wage earner was unemployed, then the family was classified as lower class. For females, lower class status was related to dropping out of school, living independently early, and experiencing more precocious transitions. Lower class males were also more likely to drop out of school and have more precocious transitions. Later alcohol use was significantly negatively correlated with lower class for females.

Recent data from the youth module of the 1997 NHSDA reveal that household socioeconomic status as measured by family income is associated with adolescent substance use (2001, OAS). For adolescent past-year marijuana use, there does not appear to be much of an overall variation with household income. Examination of heavier marijuana use, though, shows a different pattern with socioeconomic status. Adolescents 12 to 17 years old from poorer households (incomes less than $20,000) were more likely than those from the wealthiest households (incomes $40,000 and greater) to have used marijuana at least 51 times in the past year (6.5 percent vs. 3.9 percent, respectively). Household income does not show an association with heavier alcohol use. It does, however, show a significant relationship with overall past-year alcohol use, with the higher an adolescent's family income, the higher the likelihood that an adolescent used alcohol. Consistent with this finding, Zucker and Harford (1983) found a positive relationship of teenage drinking with parental occupational prestige and education. The relationship between socioeconomic status and delinquency appears to work differently, with a negative correlation found between these two factors (see Hawkins, et al., 1992).

Jessor, et al. (1991) reported that socioeconomic variables related to a respondent's family of origin showed virtually no relationship to young adult problem behaviors. Some exceptions to this were the inverse relationship of mother's education with young adult women's past-month marijuana use in the College sample and the positive correlation of mother's education with men's past-month marijuana use and past 6-month intoxication in the High School sample. When young adult educational attainment was the dependent variable of interest, family socioeconomic measures showed significance more consistently. Father's education, father's occupational status, and family socioeconomic status were positively associated with young adult educational attainment in all samples except the College Study men. Mother's education was also found to have a positive correlation with young adult education in the two High School Study samples.

F. Parental Alcoholism

A branch of research has developed around the issue of parental or familial alcoholism's effects upon children and adults (Russell, 1990). Research shows that children of alcoholics (COAs) are at risk for a plethora of negative outcomes, including early onset of alcohol and drug use and lowered academic achievement (Chassin, et al, 1993; Hill and Yuan, 1999; McGrath, Watson, and Chassin, 1999). Individuals with family members who abused alcohol were also more likely to show alcohol and hard drug abuse or dependence in adolescence (Kilpatrick, et al 2000). Perceived control and cognitive coping buffered adolescents' initiation of substance use from parents' alcoholism (Hussong and Chassin, 1997). Other research has shown that family cohesion is important in mitigating the relationship between stress that accompanies a parent's substance use disorder and adolescent substance use (Su, et al., 1997).

Adult children of alcoholics (ACOAs) are still affected by this factor in terms of their close relationships. Having an alcoholic parent is associated with earlier marriages, increased marital problems, and greater likelihood of divorcing in adulthood (Dawson, Grant, and Harford, 1992; Goodwin, et al 1977; Parker and Harford, 1988).

The connection between parental or familial alcoholism and later alcohol-related problems is well established. Dawson, Harford, and Grant (1992), after adjusting for age, race, sex, and poverty, found that compared with respondents with a negative family history, the odds of alcohol dependence increased by 45 percent among persons with alcoholism only in 2^nd or 3^rd degree relatives (i.e., biological grandparents, aunts, uncles, cousins, nieces, nephew, or other blood relatives) and by 86 percent among those with alcoholism only in 1^st degree relatives (i.e., biological parents, siblings, or children) (see also Grant, 1998). Effects of parental alcoholism on young adult substance use disorder and accelerated heavy alcohol use in adolescence is in part mediated by conduct problems (Chassin, et al 1999; Hussong, Curran, and Chassin, 1998). There is also some evidence that ACOAs are at increased risk for a drug use disorder (Gotham and Sher, 1996).

Endnotes

1. Alcohol abuse is a chronic disease in which a person continues using alcohol despite problems caused or worsened by its use and in dangerous situations. Alcohol dependence, or alcoholism, is a chronic, progressive disease characterized by frequent use of excessive amounts of alcohol, increased tolerance, inability to cut down on use, and occurrence of withdrawal symptoms when a person attempts to give up alcohol use. Alcohol abuse and dependence are distinguished on the basis of the number and type of alcohol diagnostic criteria met as detailed in the Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised (DSM-III-R; American Psychiatric Association, 1987).

2. See Twaite (1998) for a discussion of the Disaster Theory of Divorce versus the Challenge Theory of Divorce. The former perspective paints a fairly bleak picture of children's adjustment post-divorce regardless of circumstances. The Challenge approach posits that the effects of a divorce can be either negative or positive, depending on circumstances and perceptions following the divorce.

Chapter III: Data and Methods

A. Data

This study uses data from the National Longitudinal Survey of Youth--1979 cohort (NLSY79). The NLSY79 is a large, nationally representative, omnibus survey sponsored by the U.S. Bureau of Labor Statistics. ³ Over 12,000 youths ages 14-22 were first interviewed in 1979. They have been re-interviewed annually through 1994 and biennially since. The original sample included over-samples of blacks, Hispanics, economically disadvantaged non-black non-Hispanics, and youth in the military. Over time the economically disadvantaged and military over-samples were discontinued due to budget reductions. The remaining sample has seen remarkably low attrition with over 84 percent having been interviewed in 1998, the eighteenth round of interviewing.

The NLSY79 focuses on labor market behavior with information collected on aspects of the respondents' lives which are thought to influence, or be influenced by, their labor market behavior. The survey routinely collects information on education, job training, marriage, fertility, household composition, health status, income, and assets. In selected years, funding from various government agencies has provided for collecting additional information on such things as alcohol consumption and drug usage.

The advantage of using the NLSY79 is the availability of measures of long-term adult outcomes in a continuous context. This requires a longitudinal survey so data sets such as the Youth Risk Behavior Survey (YRBS) or the Current Population Survey (CPS) are inadequate.

Also, surveys such as the Survey of Income and Program Participation (SIPP) do not follow respondents for a long enough period of time. Other longitudinal surveys of adolescents, such as Add Health, the National Survey of Adolescent Males, or the new NLSY97, began more recently and have not yet followed their respondents beyond the earliest years of adulthood. The youngest respondent in the NLSY79, on the other hand, turned 34 in 1998 (the most recent year of data available at the time of this study). Over the years of the survey, we are able to see the respondents (generally) complete their education, develop their careers, and form families. Also, the NLSY79 has collected information on certain behaviors such as alcohol use and drug use at multiple points in time.

The NLSY79 pioneered asking about sensitive activities in a large omnibus survey. However, these questions did not appear until the respondents were mostly out of their adolescent years. ⁴ Thus contemporary measures of frequency and intensity are not measured in the years needed for this study. For the most part, we are restricted to a retrospective report on age of initiation. The one exception to this is in reports of delinquent and criminal behaviors. In this case, questions were included in 1980, when the respondents were 15-23 years old. The weakness in this case is that some respondents were no longer adolescents, and it is a one-time measure so that cumulative effects differ across different aged respondents. ⁵ While these weaknesses limit the analysis in certain ways, the NLSY79 is very in its inclusion of a wide variety of long term adult outcomes.

Like all survey data, the NLSY79 relies on self-reports of socially undesirable behaviors. Thus there may be under-reporting of these behaviors. Mensch and Kandel (1988) find some evidence of under-reporting drug use in the 1984 NLSY79. Also, since we are limited to reports of age of initiation, the length of retrospective recall may distort the distribution. Individuals may recall inaccurately, putting events nearer or further in time than their actual occurrence. However, contemporary reports may include more reporting of less important events (e.g. a single cigarette smoked) and may be more affected by the social desirability of a given response. Thus, retrospective reports of age of initiation may not be worse than contemporaneous reports.

The sample used in this study includes all NLSY79 respondents who are part of the continuing sample. Table 1 shows sample sizes by sex and race/ethnicity.

Table 1. Sample Sizes by Sex and Race/Ethnicity: NLSY79⁶
Sex	Race/Ethnicity	Sample Size	Percent
Total		9986	100
Male	Non-Black non-Hispanic	2529	25.33
	Black	1524	15.26
	Hispanic	981	9.82
Female	Non-Black non-Hispanic	2495	24.98
	Black	1477	14.79
	Hispanic	980	9.81

It should be noted that this is the total available sample for the analysis. Any given behavior, outcome, or other variable may be missing for a particular observation. Thus sample sizes will differ depending on what relationship is being analyzed.

B. Methods

In this chapter we provide a general description of the variables and estimation techniques employed in this study. A complete description of how we created variables, the mean and sample size for each variable, and a discussion of our estimation techniques is included in Appendix A.

1. Variable Creation

a. Adult Outcomes

The NLSY79 offers a wide array of outcomes we can study. We define ten adult outcomes of interest and categorize them into four major domains. The outcome domains and the measures in each domain are:

Health

A measure of past-year alcohol abuse or dependence around age 30 ⁷, defined according to the DSM III-R ⁸. Alcohol abuse and alcohol dependence measure different problem behaviors. We combine them to capture either type of alcohol use disorder.
Whether used drugs (marijuana or cocaine) in the past month around age 30. ⁹ While simply having used marijuana or cocaine in the past month is not an indication of a drug problem per se, most adults do not use these drugs routinely. We specifically chose not to capture drug use in the past year, a common measure, in order to reduce the likelihood of capturing occasional recreational use.

Economic

Ever under the poverty line between ages 25-29. Data limitations restricted the highest age we could consider to 29. We purposely cut off the bottom age to prevent college students from appearing to be in poverty (see Appendix A).
The number of years in poverty between the ages of 25-29. This variable distinguishes those who may have temporarily, for any number of reasons, fallen below the poverty line from those who are more regularly in poverty.
Ever on welfare (AFDC/TANF or food stamps) between ages 21-33. As an alternative to the poverty measure, we capture welfare receipt. The data allow us to look over a wider range of ages than for poverty.
The number of years on welfare between ages 21-33. This variable distinguishes transitory welfare receipt from longer term dependence.
The percent of employable time spent employed between the end of formal schooling and age 33. This variable captures attachment to the labor force. In order not to "penalize" women for time spent out of the labor force for childbearing, we calculate the time spent employed as a percent of time in the labor force. We identify the end of formal schooling before starting the count.
The age at which achieved at least 2 years working for a single employer since leaving formal schooling. Unlike general labor force attachment, this measure captures how long it took, since leaving formal schooling, to hold a "steady" job. There is not a standard measure of labor market success commonly used, but these two measures are consistent with the literature (Pergamit 1995).

Family Formation

We create a single outcome comprised of six combinations of marriage and fertility outcomes measured at age 33. These are:

never married without children,
never married with children,
currently married without children,
currently married with children,
previously married but now divorced without children, and
previously married but now divorced with children.

Crime

Whether ever been in jail by age 33. This variable is based on the location of the interview (respondents in the NLSY79 are followed anywhere, even into institutions). Only incarcerations as an adult are counted. Since some jail episodes would have occurred between interviews, this outcome is underestimated. However, no national data set measures the lifetime prevalence of serving time in jail; the NLSY79 provides an opportunity not found elsewhere (Freeman 1996).

b. Adolescent Risky Behaviors

The NLSY79 offers us a set of risky behaviors that typically begin during or near the teenage years, if they begin at all. We examine five of these behaviors:

alcohol usage,
marijuana usage,
cocaine usage,
sexual activity, and
delinquency.

Each of these is measured using age of initiation except for delinquency, which is a measure of total number of delinquent and/or criminal acts in 1980. We divide ages of initiation into four age categories:

11-15 (early adolescence);
16-17 (middle adolescence);
18-19 (late adolescence); and
those who did not initiate as an adolescent (either initiated after age 19 or never initiated).

Combining those who initiate after age 19 and those who never initiate reflects our interest in the effects of adolescent behavior. These two groups include all individuals who did not initiate as an adolescent. Further, since we do not observe respondents' entire lifetimes, we cannot know if they "never" initiate. One caveat throughout our analysis is that "late initiators" and "never initiators" may be very different groups.

Two primary ways of analyzing delinquency appear in the literature. One divides delinquent acts into "personal" and "property" (e.g. Greenberg 1985). Another common method is to break down delinquency by levels of frequency. Nye and Short (1957) suggest four categories: (1) did not commit the act, (2) committed the act once or twice, (3) committed the act several times, and (4) committed the act very often. We adopt this latter approach and define four categories of delinquency based on the number of delinquent/criminal acts engaged in: 0, 1-2, 3-8, and 9 or more.

c. Family Environment

For our investigation into the impact of adolescent family structure, we use a measure of living arrangements at age 14. While the impact of family structure is likely a dynamic process, differing depending on when and how many changes occurred, we simplify the analysis to a single age around the time that most adolescents are making their choices whether to undertake a risky activity. ¹⁰

We identify six different family types:

two biological parents,
single mother,
single father,
mother with stepfather (or other man),
father with stepmother (or other woman), and
no biological parents (living with relatives, non-relatives, or in institutions).

To explore the impact of SES, we examine mother's and father's education, categorized as: less than 12 years, 12 years, some college, and college graduate. To explore the impact of having a parent with alcohol problems, we create a single variable that captures if either biological parent was an alcoholic or a problem drinker during the respondent's childhood. ¹¹

2. Estimation

We employ a variety of estimation methods as dictated by the type of outcome variables examined. For the continuous variable measuring the percent of time employed since leaving school, we are able to use ordinary least squares (OLS). For the dichotomous variables alcohol abuse or dependence, used drugs in the past month, ever in poverty between ages 25-29, and ever on welfare between ages 21-33, we employ logistic regressions. In the case of marriage and fertility outcomes, where there are multiple combinations of a single outcome variable, we estimate a multinomial logistic. The variables, "number of years in poverty" and "number of years on welfare," are count variables that are estimated using a negative binomial regression. Finally, to estimate the time since leaving school that is needed to acquire a steady job (lasting at least two years) requires estimating a survivor function which we approximate with the Weibull distribution.

A more detailed description of the estimation methods is included in Appendix A. What is important is the interpretation of the values reported from these regressions. They fall into two categories: marginal effects and odds ratios. We report marginal effects for OLS, negative binomial, and Weibull regressions. For each of these methods, the marginal effect measures the change in the outcome variable for a unit change in an explanatory variable. In the case of a categorical variable, each marginal effect is relative to the effect of an omitted category. For example, we divide age of initiation into four categories. In the regressions, initiation between ages 11-15 is omitted and serves as the reference category.

Odds ratios are reported for logistic and multinomial logistic regressions. Odds ratios measure the relative probability of the estimated outcome among one group relative to the reference group. The choice of reference group is the same across all types of estimations. Odds ratios for multinomial logistics are difficult to interpret because there are multiple equations. As an alternative we create predicted values for each outcome combination for each category of the relevant explanatory variable. Thus, we would have a predicted probability of observing each of the six adult family formation outcomes given each of the four age of initiation categories.

Endnotes

3. For a complete description of the NLSY79 and the entire NLS program, see Bureau of Labor Statistics, NLS Handbook 2000, Washington, DC.

4. The youngest members of the sample may have gotten questions about some risky behaviors in their later adolescent years. For example, the first alcohol questions appeared in 1982 when the respondents were 17-25 years old.

5. The measurement error introduced most likely increases the standard errors of any of our estimates, making it more difficult to detect significant relationships.

6. Note: The analysis sample excludes the military over-sample dropped after 1984 and supplemental economically disadvantaged non-black, non-Hispanic over-sample dropped after 1990.

7. The respondents were 29-32 at the time the outcome was measured.

8. DSM-III-R refers to the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised. This tool is used by practitioners to help diagnose and treat mental disorders. The manual was updated in 1994 with publication of DSM-IV.

9. The respondents were 28-32 at the time the outcome was measured.

10. This simplification seems reasonable given the scope of this study, however we clearly do not capture the full family context. For example, for an adolescent living with a single parent at age 14, we do not know whether the parents never married; whether a marriage ended in divorce or death, or whether a divorce (or death) occurred recently or earlier in the adolescent's life.

11. We require that the respondent must have lived with the biological parent at least one year.

Chapter IV: The Relationship Between Adolescent Risky Behaviors and Family Environment

Most existing literature on adolescent risky behaviors concentrates on explaining the causes and correlates of the behaviors. A subset of the literature explores the relationship between the behaviors and long-term outcomes. Before we examine the relationship between engaging in adolescent risky behavior and subsequent adult outcomes, we show the distribution of ages of initiation for our sample by sex, race/ethnicity, and educational attainment (Table 2). With the exception of cocaine use, a majority of individuals have engaged in all of the behaviors by age 19. Only about 16 percent have not used alcohol and about 20 percent have not yet engaged in sex, whereas 37 percent have not committed any delinquent acts and 42 percent have not tried marijuana. Males are more likely than females to engage in any of the behaviors, and they start at younger ages.

Teenage alcohol consumption, marijuana use, and cocaine use are more prevalent among whites than either blacks or Hispanics. However, whites have the lowest rates of sexual initiation. Blacks, on the other hand, have the highest rates of sexual initiation with nearly 42 percent having initiated sex by age 15. Only 8.5 percent have not engaged in sex by the age of 19, compared with 22 percent of whites and 19.5 percent of Hispanics. While whites are the most likely to have committed nine or more delinquent acts, blacks have the highest rates of committing any delinquent acts.

Not surprisingly, most adolescent risky behavior is inversely correlated with educational attainment. High school dropouts have the highest rates of early alcohol use, though each educational group has about the same rate of alcohol initiation by age 19. High school dropouts are the most likely to use marijuana and engage in sex and do both at earlier ages than more educated individuals. They are also the most likely to commit delinquent or criminal acts and commit the most such acts. Only cocaine use does not appear to have any significant differences across educational group.

To determine if ages of initiation in the NLSY79 are consistent with those of other data sets, we examined means and medians for each risky behavior. It proved difficult to make exact comparisons because of the need to identify a similar age cohort from a similar time period, reporting at similar ages. However, the National Household Survey of Drug Abuse (NHSDA) provided a nearly perfect match for comparing ages of initiation for alcohol, marijuana and cocaine. We used the 1985 NHSDA and selected respondents ages 20-28, the same ages NLSY79 respondents were in that year. NLSY79 respondents first reported their age of alcohol initiation in 1982 and ages of marijuana and cocaine initiation in 1984.

The two surveys compare favorably. Mean ages of alcohol, marijuana, and cocaine usage in the NHSDA (for those who had initiated) were 16.17, 16.51, and 19.65, respectively. The comparable means from the NLSY79 were 16.67, 16.41, and 19.79. Medians also tracked well across the two surveys. Median ages of initiation in both surveys were roughly 16 for alcohol, 16 for marijuana, and 19 for cocaine. ¹²

Making comparisons for age of sex initiation proved more of a challenge. Most studies examine teenage sexual behavior and generally only for women. The best source of comparison is the National Survey of Family Growth (NSFG). Our mean for age of sex initiation of 17.62 compares favorably to those found in the 1995 NSFG. In that year, NLSY79 respondents were 30-38 years old. Abma, et. al. (1997) report means from the 1995 NSFG of 17.8 for women ages 30-34 and 18.0 for women ages 35-39.

The National Survey of Adolescent Males (NSAM) provides perhaps the only possible measure for male sexual debut. However, NSAM cohorts are from a later time period and the respondents have not been followed far enough into older ages to make a good comparison. Using the NSAM, Ku, et. al. (1993) report a mean age of sex initiation for males of 15.4 in 1998 and 15.2 in 1991, measured when these males were between ages 17.5-19. Given that the NLSY79 capture later ages of initiation, our mean of 16.38 seems reasonable. When we restrict the NLSY79 sample to those who initiated before age 20, our mean is 15.9, closer to the NSAM results.

Table 2. Weighted Percent of Adolescent Risky Behaviors by Sex, Race/Ethnicity, and Educational Attainment
Adolescent Risky Behaviors	Sample Size	Total	Sex		Race/Ethnicity			Educational attainment
Adolescent Risky Behaviors	Sample Size	Total	Male	Female	Non-black, non-Hispanic	Black	Hispanic	High School Dropout	High School Graduate	Some college and up
Age of Alcohol Initiation
11-15	1830	20.4	25.25	15.36	21.61	13.61	20.40	30.05	20.97	17.97
16-17	3454	37.2	39.14	35.29	38.29	33.83	31.95	32.79	36.76	38.56
18-19	2585	26.9	24.80	29.03	27.03	26.51	25.72	21.15	27.21	27.70
not by age 19	1848	15.5	10.81	20.33	13.05	26.05	21.92	16.01	15.05	15.77
Age of Marijuana Initiation
11-15	1949	23.2	26.11	20.22	24.00	19.03	22.65	32.34	23.80	20.87
16-17	1813	22.3	24.17	20.37	23.22	19.62	16.95	21.27	23.98	21.00
18-19	1005	12.1	12.81	11.30	12.12	11.95	11.66	8.19	11.14	13.68
not by age 19	3828	42.4	36.90	48.11	40.65	49.41	48.74	38.21	41.09	44.45
Age of Cocaine Initiation
11-15	114	1.2	1.49	0.90	1.19	0.91	1.95	1.98	1.10	1.14
16-17	379	4.7	4.90	4.52	5.21	2.35	3.79	5.73	5.32	3.97
18-19	622	7.5	9.25	5.79	8.12	4.36	7.50	6.82	7.61	7.63
not by age 19	7454	86.5	84.36	88.79	85.48	92.37	86.76	85.47	85.97	87.26
Age of Sex Initiation
11-15	2487	21.7	29.68	13.49	17.98	41.98	22.89	41.59	26.13	13.76
16-17	3295	34.7	35.91	33.39	34.90	33.81	33.71	38.78	37.72	31.10
18-19	2167	23.8	18.90	28.95	25.30	15.70	23.90	13.31	22.79	26.90
not by age 19	1691	19.8	15.52	24.17	21.82	8.51	19.50	6.32	13.36	28.24
Number of times committed crimes/delinquencies in 1980
9+	1318	15.1	23.13	6.79	15.66	12.64	12.86	22.16	16.41	12.52
3-8	2288	25.3	29.61	20.80	25.42	24.98	23.76	26.25	25.95	24.45
1-2	2238	22.7	21.25	24.24	21.45	29.70	23.40	24.71	23.35	21.81
0	3476	37.0	26.01	48.17	37.47	32.68	39.98	26.88	34.29	41.21

Table 3 shows the distribution of age of initiation into each behavior by family type at age 14. Adolescents living in two biological parent (intact) families have the lowest rates of earliest initiation and the highest rates of latest initiation of all family types for nearly all behaviors. Nowhere is this more distinct than in sex initiation. About 17 percent of those in intact families initiate sex between ages 11 and 15 whereas most other family types range between 30-35 percent (although 41 percent of those living without either biological parent initiate in this age range). As late as age 19, 23 percent of those living in intact families have not yet initiated whereas only 7-12 percent of those in other family types has waited.

Interpreting behaviors for other family types requires caution. Those not living in families that include their biological mother are unusual. Only about 6 percent of the sample weren't living with their biological mothers at age 14. During the 1970s, when these individuals were teenagers, placement of a youth with his/her father was unusual. It may have occurred due to the death of the mother, due to an unfit mother, or perhaps because of the youth's behavior. A mother who is having difficulty with a teenager, particularly a son, may send the child to live with his father. Thus adolescents living only with their father may differ in ways not measured which could impact results. As a result, it would be pure speculation why we see that those living with fathers and stepmothers have the highest rates of early alcohol initiation, have the lowest rates of late initiation into marijuana use, and commit the most delinquent acts. Earliest initiation into marijuana occurs among those living with their mother and a stepfather. This group also is the most likely to commit at least 3 delinquent acts. On the other hand, those who are least likely to have not committed any delinquent acts are those living with only their fathers.

Table 3. Weighted Percent of Adolescent Risky Behaviors by Type of Family Structure at Age 14
Adolescent Risky Behaviors	Family type at Age 14
Adolescent Risky Behaviors	Two biological parents	Single mother	Single father	Mother with stepfather	Father with stepmother	No biological parents (relatives, non-relatives, institutions)
Sample Size	6763	1732	118	768	167	299
Age of Alcohol Initiation
11-15	18.91	23.42	30.31	24.15	32.21	29.15
16-17	37.40	36.46	30.83	39.77	37.04	31.28
18-19	28.72	21.03	23.49	22.86	18.53	18.90
not by age 19	14.98	19.08	15.36	13.22	12.22	20.68
Age of Marijuana Initiation
11-15	20.99	28.80	24.68	31.31	29.23	34.77
16-17	22.31	20.91	23.92	24.77	24.47	19.31
18-19	12.55	11.08	8.70	8.86	14.80	11.58
not by age 19	44.15	39.22	42.71	35.06	31.50	34.34
Age of Cocaine Initiation
11-15	0.94	2.09	0.00	2.11	0.97	1.90
16-17	4.02	6.06	10.93	8.42	6.40	2.35
18-19	7.25	7.06	6.06	10.74	10.00	8.33
not by age 19	87.79	84.80	83.01	78.73	82.62	87.42
Age of Sex Initiation
11-15	17.46	35.16	31.43	30.33	34.43	40.61
16-17	34.29	33.64	36.76	40.93	33.55	32.25
18-19	25.14	19.44	20.15	21.13	22.16	18.07
not by age 19	23.11	11.75	11.66	7.61	9.86	9.07
Number of times committed crimes/delinquencies in 1980
9+	14.07	16.74	18.23	19.98	22.78	14.06
3-8	25.34	23.92	26.60	28.39	17.13	26.86
1-2	21.92	25.28	32.62	21.42	28.07	28.30
0	38.66	34.06	22.55	30.21	32.02	30.78

Table 4 reveals some very interesting, and perhaps counter-intuitive, relationships between parents' education and ages of initiation. For alcohol, marijuana, and cocaine use, there is generally an inverse relationship between the age of initiation and the parents' educational attainment, i.e. the more education either parent has, the earlier their children engage in substance use. This stands out in several places. Individuals whose mothers are college educated have the highest rates of early alcohol initiation (23.14 percent between ages 11-15) and the lowest rates of post-teenage initiation (11.47 percent). Those with fathers who are high school dropouts have the lowest rates of adolescent and teenage alcohol usage (nearly 19 percent have not used alcohol by age 19 compared with 12-14 percent for the other three education levels). Rates of teenage usage of marijuana rise consistently with mothers' education and nearly so with fathers' education. Rates of teenage usage of cocaine also rise consistently with both mothers' and fathers' education (little cocaine usage occurs before ages 18-19).

This positive relationship between adolescent substance use and parents' education has been found previously in the literature (e.g. Zucker and Harford, 1983), but no adequate explanation has been put forward. It is possible that parents' education is a proxy for parental income. Other literature has found a relationship between family income and adolescent substance abuse (OAS, 2001). However, since parents' education should capture factors besides income including parental attitudes that have been found to be protective factors, this implies a very B dominating income effect. We are not able to control for income to determine if there are separate, perhaps countervailing impacts. Further study into this relationship would be valuable.

A much different pattern emerges for sex initiation, with sexual debut occurring later for children of more educated parents. There is a consistent relationship such that children of more educated parents are the least likely to initiate sex at young ages and are the most likely to wait until after their teenage years. For example, nearly 30 percent of those with college educated mothers have not initiated sex by age 19 compared with under 14 percent for those whose mothers are high school dropouts. Similarly, nearly 31 percent of those with college educated fathers do not initiate as a teenager compared with under 17 percent whose fathers are high school dropouts. Delinquency does not show consistent patterns, though in general children of more educated parents are least likely to have committed any delinquent or criminal acts.

Table 4. Weighted Percent of Adolescent Risky Behaviors by Parental Educational Attainment
Adolescent Risky Behaviors	Mother's Educational Attainment				Father's Educational Attainment
Adolescent Risky Behaviors	High School Dropout	High School Graduate	Some College	College Graduate	High School Dropout	High School Graduate	Some College	College Graduate
Sample Size	4020	3729	860	718	3569	2909	797	1203
Age of Alcohol Initiation
11-15	19.88	19.67	20.29	23.14	20.62	19.70	22.91	19.56
16-17	35.58	38.07	36.37	38.17	34.89	39.31	37.99	38.76
18-19	25.65	27.59	28.67	27.21	25.60	27.49	27.18	29.00
not by age 19	18.89	14.67	14.37	11.47	18.89	13.50	11.93	12.68
Age of Marijuana Initiation
11-15	23.07	23.07	21.52	23.34	21.98	23.81	22.40	23.96
16-17	22.40	22.45	20.36	23.29	21.79	23.50	22.30	22.03
18-19	9.39	12.19	17.46	13.10	10.34	11.68	12.59	16.45
not by age 19	45.14	42.29	40.66	38.26	45.89	41.01	42.70	37.56
Age of Cocaine Initiation
11-15	1.31	1.22	0.52	1.26	1.24	1.01	1.29	1.25
16-17	4.28	4.53	5.20	5.81	4.46	5.14	4.98	4.20
18-19	6.27	7.29	7.88	12.75	5.39	8.00	8.72	10.39
not by age 19	88.13	86.96	86.40	80.18	88.91	85.86	85.01	84.16
Age of Sex Initiation
11-15	28.65	18.45	16.54	15.22	25.61	20.46	16.76	12.40
16-17	36.38	35.71	30.59	28.67	35.12	36.35	35.39	29.68
18-19	21.08	25.72	24.18	26.34	22.35	24.74	23.95	27.00
not by age 19	13.90	20.12	28.69	29.77	16.92	18.45	23.89	30.91
Number of times committed crimes/delinquencies in 1980
9+	14.74	16.22	13.25	12.78	14.51	15.99	15.56	13.96
3-8	26.03	24.71	26.39	23.79	24.73	24.52	29.56	24.94
1-2	24.85	21.02	21.56	22.56	25.57	21.60	18.72	21.86
0	34.38	38.04	38.79	40.87	35.19	37.88	36.15	39.24

Although there may be some ambiguity about the degree to which other family environment measures are positive or negative influences on children, parental alcoholism is probably never viewed as good. The cross-tabulations in Table 5 clearly confirm this view. Across all measures, children of alcoholic parents engage in risky behaviors at younger ages and commit more delinquent acts. By age 17, over 64 percent have used alcohol, compared with 55 percent of those without an alcoholic parent. Fifty seven percent have used marijuana, nearly 9 percent have used cocaine, and 63 percent have initiated sex by this age. The comparable rates for children without alcoholic parents are 42 percent, 5 percent, and 53 percent, respectively. In 1980, over 45 percent had committed at least 3 delinquent or criminal acts, compared with 38 percent for the comparison group.

Table 5. Weighted Percent of Adolescent Risky Behaviors by Parental Alcoholism
Adolescent Risky Behaviors	Parent with Drinking Problem
Adolescent Risky Behaviors	Yes	No
Sample Size	1545	6549
Age of Alcohol Initiation
11-15	24.54	18.76
16-17	40.04	36.63
18-19	21.97	28.82
not by age 19	13.45	15.79
Age of Marijuana Initiation
11-15	32.26	20.69
16-17	25.15	21.33
18-19	10.60	12.52
not by age 19	32.00	45.46
Age of Cocaine Initiation
11-15	1.89	0.92
16-17	6.81	3.95
18-19	11.09	6.71
not by age 19	80.21	88.42
Age of Sex Initiation
11-15	23.98	20.04
16-17	39.03	32.81
18-19	23.86	24.48
not by age 19	13.13	22.67
Number of times committed crimes/delinquencies in 1980
9+	17.06	13.68
3-8	28.34	24.38
1-2	23.22	22.77
0	31.39	39.18

Chapter Summary

Males are more likely than females to engage in adolescent risky behaviors, and they start at younger ages.
Whites are more likely than blacks or Hispanics to engage in substance use, but have the lowest rates of sexual initiation. They are the most likely to commit nine or more delinquent or criminal acts, but blacks have highest rates of committing any delinquent acts.
Most adolescent risky behavior is inversely correlated with educational attainment.
Adolescents living in intact families have the lowest rates of early initiation and the highest rates of late initiation for nearly all risky behaviors, especially sex initiation.
Higher levels of parents' education are associated with earlier substance use among adolescents, but also the latest sex initiation and the lowest likelihood of committing any delinquent acts.
Adolescents with alcoholic parents initiate risky behaviors at earlier ages and commit more delinquent acts than other adolescents.

Chapter V: The Relationship of Adolescent Risky Behaviors and Family Environment to Adult Outcomes

We turn now to the central focus of this study, relating adolescent risky behaviors and family environment to long-term adult outcomes. Chapter III described the ten outcomes of interest in this study. We first show the distributions of these outcomes by sex, race/ethnicity, and educational attainment in Table 6.

A. Distributions

Nearly 14 percent of adults suffer from alcohol abuse or dependence as measured around the age of 30 while about 11 percent are using drugs at that age. The prevalence rate of past-year alcohol use disorder is comparable to that (14 percent) found by Harford and Grant (1994) using 1989 NLSY79 data. Both past-year alcohol use disorder and recent drug use are associated more with males than females and are inversely related to educational attainment. Whites have the highest rates of alcohol problems, blacks the lowest. Hispanics use drugs the least at age 30.

By the age of 33, 3.7 percent of adults have spent some time in jail, primarily males who have a rate of about 6.5 percent. Blacks are the most likely to have been in jail at nearly 9 percent, with Hispanics second at 5.8 percent. Going to jail is inversely associated with educational attainment.

Approximately one-quarter of the adults were in poverty between ages 25-29 and about the same fraction were on welfare between ages 21-33. Poverty and welfare prevalence is higher for females than males and again inversely associated with educational attainment. Blacks are more likely than Hispanics, who, in turn, are more likely than whites to have been in poverty or on welfare. Similarly, employment outcomes are typically best for males, whites, and more educated individuals.¹³

More than three-quarters of thirty-three year olds have been married (either currently or in the past) and about two-thirds have had children. While on average 6.4 percent have never married but have children at that age, the rate for blacks is nearly 26 percent. Although marriage is not very different among educational groups (a little higher for high school dropouts), having children is much less prevalent as education increases. About one-fifth of high school dropouts have not had children by age 33 compared with about one-quarter of high school graduates, and over two-fifths of those who have had some college education. Since marriage and fertility are not complete for everyone at age 33, this comparison reflects both delayed childbearing and childlessness among more educated people. Some of this difference will probably erode as the sample continues to age.

It should be noted that the category "never married, with children" is a subset of all non-marital births. Since we measure family status at age 33, those who may have had a non-marital birth, but married before age 33, will not be in this category. Although the percentage who at age 33 had a child but had never been married is only 6.4 percent, the percentage of those whose first birth preceded their first marriage is 19.3 percent.¹⁴

Table 6. Weighted Percent/Means of Adult Outcomes by Sex, Race/Ethnicity, and Educational Attainment
Adult Outcomes	Sample Size	All	Sex		Race/Ethnicity			Educational Attainment
Adult Outcomes	Sample Size	All	Male	Female	Non-black non-Hispanic	Black	Hispanic	High School Dropouts	High School Graduates	Some college and up
Health
Percent alcohol abuse or dependence ~ age 30	9986	13.8	19.31	8.07	14.26	11.54	12.92	20.40	15.07	11.32
Percent drug use in the past month ~ age 30	8803	11.1	13.94	8.18	11.26	11.81	7.60	14.89	12.31	9.27
Incarceration
Percent ever being in jail ages 18-33	9986	3.7	6.46	0.77	2.56	8.84	5.80	11.11	4.56	1.37
Economic
Percent ever in poverty, ages 25-29	9986	23.5	20.24	26.85	18.91	43.84	34.80	52.80	26.16	15.22
Years in poverty, ages 25-29	9430	0.511	0.388	0.638	0.383	1.11	0.792	1.329	0.591	0.275
Percent ever on welfare, ages 21-33	9986	25.0	19.68	30.52	20.51	45.97	33.92	55.10	30.41	13.92
Years on welfare, ages 21-33	9986	1.084	0.615	1.570	0.798	2.44	1.615	2.997	1.291	0.516
Percent time employed-end of school to age 33	9801	0.909	0.913	0.904	0.928	0.812	0.885	0.802	0.893	0.944
Age achieved steady job since leaving school	8467	24.97	24.82	25.13	24.79	25.85	25.39	24.97	24.098	25.71
Family Formation
Percent in marital and fertility status at 33:
Never married without children	9144	15.3	19.41	10.96	15.19	16.64	13.24	10.94	12.68	18.40
Never married with children	9144	6.4	6.22	6.54	2.71	25.93	8.78	13.60	8.12	3.43
Married without children	9144	11.9	13.25	10.45	13.27	5.79	8.03	5.16	7.91	16.70
Married with children	9144	43.6	41.60	45.73	44.97	33.88	48.40	40.56	44.55	43.40
Married but divorced without children	9144	5.2	5.23	5.17	5.81	2.65	3.28	3.99	4.89	5.71
Married but divorced with children	9144	17.7	14.29	21.15	18.05	15.11	18.27	15.75	21.85	12.36

By measuring marriage and fertility status at age 33, we are implicitly considering a pre-marital birth that is later followed by a marriage to be a transition comparable to those who had their children after marriage. We selected this measure to be consistent with our goal to look at long-term outcomes.

Table 7 shows the relationship between the age of initiation into risky behaviors and adult outcomes. A clear relationship is seen between each adolescent risky behavior and each adult outcome. The earlier one engages in any given risky behavior or the more delinquent acts committed, the greater prevalence of bad adult outcomes. For both health outcomes and for spending time in jail, the relationship is perfectly consistent for all five adolescent risky behaviors. For example, the rates of adult drug use are 17.88 percent, 14. 19 percent, 7.62 percent, and 3.01 percent for initiation between ages 11-15, 16-17, 18-19, and after 19, respectively.

For economic outcomes the relationship generally holds though there are some anomalies. In particular, those who report not initiating alcohol, marijuana, or cocaine use by age 19 do not necessarily have the best outcomes; nor is the number of delinquent acts perfectly connected with bad adult outcomes in all cases. In fact, for the outcome, "age at which obtained a steady job", those who commit more crimes find a steady job sooner. The notion of adolescent risky behavior leading to precocious transitions into adult roles may account for part of this result. For example, Newcomb and Bentler (1988) found that adolescents who used substances were more likely to enter the workforce early and thus had higher incomes as young adults, albeit they also showed more job instability. In our study, adolescents who initiated alcohol early showed a slight tendency toward this precocious transition pattern on the same employment outcome, however, later cocaine and sex initiation both were associated with earlier maintenance of a stable job.

Excluding those who have not initiated by age 19, we nearly always see later initiation associated with better adult outcomes. For example, 29.86 percent of those who initiate into alcohol between ages 11-15 spend time on welfare between ages 21-33 compared with 24.04 percent of those who initiate at ages 16-17 and 20.87 percent of those who initiate at ages 18-19. Similar welfare participation rates exist for age of marijuana initiation at 31.37 percent, 26.41 percent, and 20.39 percent, respectively.

Certain relationships stand out in the adult family formation outcomes. The later one initiates a risky behavior (or the fewer delinquent acts one commits), the more likely one will be married with children at age 33 and the least likely one will have been married with children but divorced at age 33. Not surprisingly, early sex initiation is associated with never being married with a child (13.12 percent) and least associated with never being married without children (12.01 percent).

Section Summary

Adult alcohol disorders and drug usage is more prevalent among males, whites, and those with less education.
Males, blacks, and less educated individuals are the most likely to spend time in jail.
Poverty and welfare receipt are more prevalent among females, blacks, and those with less education.
Adult employment outcomes are best for males, whites, and those with more education.
Those with more education are less likely to have children by age 33.
Earlier initiation into adolescent risky behaviors or committing more delinquent acts is associated with having more bad adult outcomes.

Table 7. Adult Outcomes by Adolescent Risky Behaviors
Adolescent Risky behaviors	Adult Outcomes
	Health		Percent ever been in jail ages 18- 33
	Percent alcohol abuse or dependence ~ age 30	Percent drug use in the past month ~ age 30	Percent ever been in jail ages 18- 33
Age of Alcohol Initiation
11-15	22.94	16.68	6.40
16-17	16.39	13.15	3.31
18-19	9.38	8.44	2.91
not by age 19	3.24	3.51	2.17
Age of Marijuana Initiation
11-15	20.30	21.10	6.68
16-17	17.27	13.30	3.60
18-19	17.11	12.65	3.49
not by age 19	7.47	3.81	2.04
Age of Cocaine Initiation
11-15	28.66	26.89	21.97
16-17	25.87	25.49	9.47
18-19	23.14	24.84	6.72
not by age 19	12.08	8.78	2.78
Age of Sex Initiation
11-15	19.51	17.88	10.09
16-17	16.01	14.19	3.08
18-19	11.60	7.62	0.93
not by age 19	6.48	3.01	0.82
Number of times committed crimes/delinquencies in 1980
9+	25.31	19.28	9.49
3-8	17.77	15.19	3.75
1-2	13.15	10.20	3.29
0	6.71	5.45	1.07

Adult Outcomes by Adolescent Risky Behaviors
Adolescent Risky Behaviors	Adult Outcomes
	Economic
	Percent ever in poverty, ages 25-29	Years in poverty, ages 25-29	Percent ever on welfare, ages 21-33	Number of Years on welfare, ages 21-33	Percent time Employed from end of schooling to age 33	Age achieved steady job since leaving school
Age of Alcohol Initiation
11-15	25.71	0.536	29.86	1.251	0.903	24.984
16-17	21.60	0.462	24.04	1.017	0.913	24.693
18-19	21.13	0.454	20.87	0.909	0.918	25.025
not by age 19	28.55	0.701	28.35	1.352	0.893	25.292
Age of Marijuana Initiation
11-15	27.37	0.562	31.37	1.286	0.899	24.929
16-17	22.02	0.473	26.41	1.086	0.913	24.896
18-19	19.38	0.411	20.39	0.851	0.922	24.880
not by age 19	23.22	0.529	22.01	1.033	0.908	25.040
Age of Cocaine Initiation
11-15	45.69	0.921	46.07	1.899	0.855	25.589
16-17	31.60	0.657	36.41	1.418	0.888	24.686
18-19	22.66	0.459	25.44	0.988	0.922	24.724
not by age 19	22.71	0.499	23.94	1.057	0.910	24.994
Age of Sex Initiation
11-15	33.89	0.771	38.06	1.771	0.867	25.195
16-17	25.23	0.544	29.29	1.288	0.904	24.902
18-19	18.73	0.399	19.40	0.799	0.926	24.802
not by age 19	15.56	0.314	10.51	0.360	0.944	24.874
Number of times committed crimes/delinquencies in 1980
9+	27.12	0.557	27.45	1.113	0.891	24.721
3-8	22.38	0.506	26.42	1.188	0.907	24.888
1-2	25.96	0.559	27.27	1.241	0.908	24.924
0	20.04	0.435	20.74	0.874	0.920	24.961

Adult Outcomes by Adolescent Risky Behaviors
Adolescent Risky Behaviors	Adult Outcomes
	Family Formation: Marital and Fertility Status at age 33 (in percent)
	Never married without children	Never married with children	Married without children	Married with children	Married but divorced without children	Married but divorced with children
Age of Alcohol Initiation
11-15	14.92	6.06	10.00	40.49	6.61	21.93
16-17	15.01	6.54	12.31	43.43	5.31	17.39
18-19	15.27	5.90	12.93	44.86	5.12	15.92
not by age 19	16.74	7.14	11.52	46.06	3.08	15.46
Age of Marijuana Initiation
11-15	14.91	7.48	11.29	38.02	6.16	22.14
16-17	16.01	6.16	11.92	42.24	4.82	18.85
18-19	14.79	5.10	12.12	44.97	6.75	16.26
not by age 19	15.22	6.13	12.12	46.98	4.51	15.04
Age of Cocaine Initiation
11-15	12.54	9.66	8.40	37.56	8.96	22.87
16-17	18.40	5.79	12.30	28.32	7.55	27.64
18-19	19.10	6.85	12.30	38.61	6.37	16.77
not by age 19	14.79	6.28	11.85	45.02	4.94	17.11
Age of Sex Initiation
11-15	12.01	13.12	8.17	36.61	5.48	24.62
16-17	12.68	6.62	10.04	42.80	5.80	22.05
18-19	14.94	4.11	12.25	48.26	5.42	15.02
not by age 19	23.57	1.55	18.34	46.99	3.38	6.18
Number of times committed crimes/delinquencies in 1980
9+	17.61	8.78	10.73	35.98	5.99	20.91
3-8	14.47	6.83	10.74	42.55	6.11	19.31
1-2	15.84	6.74	11.95	42.68	4.99	17.81
0	14.40	4.39	13.20	48.15	4.60	15.27

B. Regression Analysis

1. Introduction

These cross-tabulations do not hold anything constant, including the effect of the other adolescent behaviors on each adult outcome. The anomalies observed may reflect some other factor correlated with late initiation or delinquent behavior. To hold constant other factors that influence these adult outcomes, we estimate a set of regressions. These regressions take the general form

y = f(B,F,X)+ g,

where y is one of the ten outcomes of interest, B represents the five adolescent risky behaviors, F represents the three family environment variables, X represents all the other control variables and is a random error term.

The other control variables include sex, race/ethnicity, educational attainment, and Rosenberg's measure of self-esteem (measured in 1980).¹⁵ The regressions are estimated unweighted and each regression includes all the behaviors, family environmental variables, and control variables. The regressions allow us to examine the relationship of the adolescent risky behaviors, family structure, parents' education, and parental alcoholism to the adult outcomes while holding each set of variables constant. The regressions also allow us to control for participation in all five adolescent risky behaviors so that the impact of any particular behavior can be identified separately from the others. We might expect, similar to Newcomb and Bentler's (1988) findings, that each adolescent behavior may show a unique pattern of influence on adult outcomes. We now examine sequentially the relationship of the four sets of adolescent factors to the adult outcomes. The full set of regressions with all variables is shown in Appendix B.

2. The Relationship of Adolescent Risky Behaviors to Adult Outcomes

With ten regressions, examining the impact of five different adolescent risky behaviors with four different age categories each can be daunting. Seeming inconsistencies must be considered in the broad scheme; many of these are due to small cell sizes. We interpret the findings with an eye toward the big picture. To draw conclusions, we look at patterns of relationships across the different outcome domains. Most of the findings we cite in this report have p-values of 0.1 or less, a conventional level of statistical significance. However, while we consider statistical significance important, we do not rely on statistical significance alone. Because this study is exploratory in nature, we use p-values of 0.1 as a guideline.

Our techniques generally compare outcomes for three age of initiation groups (or numbers of delinquent/criminal acts) with a reference group (ages 11-15 or 9 or more acts). We base our judgments on whether a set of odds ratios or marginal effects generally go in the same direction when compared with the reference group, i.e. if the outcomes for each group are all better or worse than the reference group. Generally, if all the values are going in the same direction, then if the comparison for at least one of the other groups achieves significance we consider the relationship to hold. In the economic domain, where there are multiple outcomes, we see common relationships across the different outcome variables. When they all are nearly the same, we consider this confirming evidence of a relationship for the entire domain. We attempt to notify the reader when we are describing a general pattern, but where statistical significance does not always hold. It is important to note that all estimated relationships are associations between variables; nothing should be construed as a causal relationship. To understand causality would require an extensive model of behavior with many more measures than are available to us.

Despite some occasional oddities, a clear picture emerges as to the relationship of these five behaviors to the various outcomes. We report our findings by looking at the relationship of various adolescent risky behaviors to each adult outcome.

Adolescent alcohol usage. As seen in Figure 1, age of initiation into alcohol consumption is very much associated with having an alcohol abuse or dependence problem as an adult in the past year. To interpret Figure 1, recall that ages 11-15 are the reference group. Their odds ratio is set equal to one. Odds ratios for other age groups are relative to those who initiated between ages 11-15. Thus, those who initiated at ages 16-17 are only 88 percent as likely to have an adult alcohol abuse or dependence problem compared with those who initiated between ages 11=15. Those who initiated at ages 18-19 are only 56 percent as likely, and those who had not initiated by age 19 are only 23 percent as likely.

It is not too surprising to find such a strong relationship since adolescent use of alcohol is clearly the most closely related behavior to the possible consequence. Quite clearly, the earlier one initiates into alcohol use, the more likely one has an alcohol problem as an adult. This result is consistent with other research that has found a link between early alcohol initiation and adult alcohol problems (Chou and Pickering, 1992; Grant and Dawson, 1997).

Figure 1.Odds Ratios for Age of Alcohol Initiation As it Relates to Alcohol Abuse or Dependence

The use of alcohol and drugs are different behaviors among adolescents. Adolescents consuming alcohol may never use drugs. Although early adolescent alcohol use may increase the likelihood of an adult alcohol problem, it may imply nothing about adult drug problems. In our findings, there are no real differences in the likelihood of an adult drug problem among the three lower age groups. However, not using alcohol by age 19 is associated with much lower likelihood of having a drug problem as an adult. Thus those who do not engage in alcohol use as teenagers are those most likely to avoid both long-term alcohol and long-term drug problems.

Although alcohol consumption in adolescence is associated with negative health outcomes, it generally is not associated with bad economic, family formation, or incarceration outcomes. Compared with all the other behaviors we study, adolescent alcohol usage has the least association with long-term outcomes. These findings together with those from Newcomb and Bentler (1988) point to the idea that the impact of early alcohol use may portend less serious adult consequences than does engaging in other forms of risky behavior as an adolescent. On the other hand, we have measured only the direct effects of adolescent alcohol initiation. If early alcohol use increases the likelihood of other risky behaviors such as sexual activity, then there may be an indirect relationship with adult outcomes.

Adolescent marijuana usage. Like adolescent alcohol use, engaging in marijuana use as an adolescent has a strong relationship with the adult outcome of most direct consequence. That is, earlier use of marijuana implies a greater likelihood of using drugs as an adult. As seen in Figure 2, those who initiate marijuana usage at ages 16-17 are 76 percent as likely to use drugs as an adult compared with those who initiated before age 16. Those who initiated at ages 18-19 are 88 percent¹⁶ as likely and those who did not initiate by age 19 are only 28 percent as likely.

Figure 2. Odds Ratios for Age of Marijuana Initiation As it Relates to Past Month Drug Use at Age 30

Initiation as a teenager implies greater likelihood of adult alcohol abuse or dependence, but within the teenage years there is little distinction. Unlike alcohol, engaging in marijuana use as an adolescent is generally associated with bad economic outcomes. For example, Figure 3 shows the relationship between age of marijuana initiation and the likelihood of being in poverty between ages 25-29.

Figure 3. Odds Ratios for Age of Marijuana Initiation as it Relates to Ever Being in Poverty Between Ages 25-29

Those who initiate in the 18-19 age group are 79 percent as likely to be in poverty as those who initiate between ages 11-15. This age group is consistently less prone to incur bad outcomes for the other economic measures than those who initiate earlier. Oddly, those who have had not initiated by age 19 are statistically no different than the earliest initiators. This is true nearly always across the economic outcomes. We do not have an explanation for this. Since late initiators are combined with those who never initiate, there may be some impact of late initiation that is pulling down the average outcome for this group. Late initiation is likely a different phenomenon than early initiation, perhaps resulting from bad economic circumstances rather than preceding them.¹⁷

The pattern for the likelihood of spending time in jail, however, is consistent; the earlier one initiates marijuana use, the more likely they spend time in jail.

For the relationship between age of initiation into marijuana use and adult family formation, only one odds ratio achieves a reasonable level of statistical significance. Overall this implies that the relationship is the same regardless of age of initiation.

Adolescent cocaine usage. Many young people experiment with both alcohol and marijuana, but cocaine use is rarer (see Table 2). Its usage also tends to begin at later ages. As a result some of the cell sizes in our estimations are small. However, a clear pattern emerges showing that those who begin using cocaine as teenagers have worse economic, health, and incarceration outcomes. It is more difficult to interpret the results for family formation. Teenage cocaine users are no more likely than non-users to end up as single parents or divorced. However, it is clear that they are the most likely to end up as never having married without children and least likely to end up married with children. This implies that teenage cocaine users do not have non-marital births or get divorced primarily because they are least likely to be able to form solid relationships in the first place. Overall they are the least likely to marry and the least likely to have children (regardless of marital status) than those who do not use cocaine as teenagers.

Adolescent sexual activity. Age of initiation into sexual intercourse is perhaps the most consistent predictor of poor adult outcomes across nearly every domain. Unlike any other behavior studied, the results have near perfect consistency with no anomalies. Not only is engaging in sex as a teenager associated with bad outcomes in general, earlier ages of initiation are consistently increasingly associated with poorer outcomes. This is true for all six measures of economic success and for the probability of spending time in jail. Similarly, early sex initiators have poorer family formation outcomes. The predicted probabilities of each outcome for each age group are shown in Figure 4. The earlier a youth begins engaging in sexual activity, the more likely he or she ends up never married with a child. Specifically, we predict 11.2 percent of those who initiated sex between ages 11-15 to be never married with children at age 33. Of those who initiate at ages 16-17, we predict 8.7 percent will be never married with children at age 33; 7.5 percent of those who initiate at ages 18-19 and 4.2 percent of those who do not initiate by age 19. In general, early sex initiators are more likely to marry by age 33 than are those who wait. Over 79 percent of those who initiated by age 17 had married while about 76 percent of those who initiated at ages 18-19 and 67.4 percent of those who had not initiated by age 19 had married by age 33. Of those who marry by age 33, divorce rates are predicted to be higher the earlier there had been sex initiation. 36.2 percent of those who initiated between ages 11-15, 31.1 percent of those who initiated at ages 16-17, 23.5 percent of those who initiated at ages 18-19, and 14.2 percent of those who had not initiated by age 19 are predicted to be divorced by age 33.¹⁸ Not surprisingly, those who initiated sex earlier are predicted to be more likely to have children by age 33, regardless of their marital history. By age 33, 77.1 percent of those who initiated sex between ages 11-15 are predicted to have children; 72.6 percent of those who initiated at ages 16-17; 67.4 percent of those who initiated at ages 18-19; and 53.3 percent of those who had not initiated by age 19.

Figure 4. Predicted Probabilities of Marriage-Fertility Outcomes At Age 33 as They Relate to Age of Sex Initiation

The health domain as measured by adult alcohol and drug problems is least related to early sexual activity. Those who do not engage in sex as teenagers are less likely to have an adult alcohol problem, but there is no distinction in the likelihood across the teenage years. Those who initiate after age 17 are less likely to use drugs as an adult; those who initiate after 19 are least likely. There is no distinction for those who initiate before age 18. Although there isn't as strong a relationship between age of sex initiation and the two health measures, it is still the case that teenage sex initiators are more likely to have adult alcohol and drug problems.

Adolescent delinquency. Committing delinquent and criminal acts as an adolescent is associated with mostly bad adult outcomes. As expected, the more delinquent acts committed, the more likely the youth spends time in jail by age 33. A similar result is found for the likelihood of having an adult alcohol or drug problem. These results are comparable with Windle's (1990) where, even when controlling for early substance use, early delinquency was predictive of later substance use. These three adult outcomes have consistent results across the levels of delinquency.

Some of the other results bounce around a bit, possibly a consequence of noise in our measure. However, isolating those who committed nine or more delinquent or criminal acts shows they definitely have worse economic outcomes than do others. This holds true across all six economic measures. One of the stronger relationships is with the likelihood of ever being on welfare between ages 21-33, shown in Figure 5. Those who did not commit any delinquent acts are 69 percent as likely to ever be on welfare during those ages compared to those who committed nine or more delinquent acts; those who committed one or two delinquent acts are 80 percent as likely; and those who committed 3-8 delinquent acts are 83 percent as likely. All three differences are statistically significant at the 90 percent confidence level or higher. Past research on the connection between delinquency and economic outcomes has been mixed. After controlling for education and other background variables, Monk-Turner (1989) could find no relationship between delinquency and adult occupational status. In our study, when holding education and other variables constant, we were able to observe a significant impact of delinquency on a variety of economic measures.

Figure 5. Odds Ratios for Number of Delinquent Acts as it Relates to Ever Being on Welfare Between Ages 21-33

In terms of family formation, there are some consistent results. The more delinquent acts performed, the more likely the person ends up never married with a child and the least likely they are to end up married with children. In general, they are less likely to marry and more likely to divorce if they marry.

To summarize, we have examined the relationship between five adolescent risky behaviors and ten adult outcomes. There is a fairly consistent pattern that engaging

in risky behaviors as a teenager is associated with less successful adult outcomes. For many, the relationship is such that the earlier one engages in the behavior, the more likely they face a bad outcome as an adult. The most consistent predictor of a bad adult outcome is age of initiation into sexual activity. Alcohol usage, on the other hand, is perhaps the one teenage behavior least associated with bad adult outcomes. Age of initiation into alcohol usage is, however, associated with adult alcohol abuse or dependence. These relationships should not be interpreted as causal, but only as associations. Other factors that are not controlled for may influence these outcomes. Age of initiation may be capturing unmeasured adolescent characteristics or circumstances that are related to the transition into adulthood. The summary below provides a more detailed description of the findings in this section.

Section Summary

Age of initiation into adolescent alcohol use is associated with adult alcohol disorders and teenage use is associated with adult drug usage. However, there is no relationship between adolescent alcohol use and adult economic, family, or incarceration outcomes.
Age of initiation into marijuana use is associated with adult drug usage, bad adult economic and incarceration outcomes. Teenage usage is associated with adult alcohol disorders but adolescent marijuana use is not associated with any specific adult family formation outcomes.
Teenage cocaine use is associated with worse economic, health, and incarceration outcomes. They are the least likely to marry or have children by age 33.
Age of initiation into sexual intercourse is the most consistent predictor of poor adult outcomes across nearly every domain, especially economic and incarceration outcomes. They are the most likely to have never married but have a child by age 33. Those who marry by age 33 are the most likely to divorce.
Committing more delinquent acts is associated with bad adult alcohol, drug, and incarceration outcomes. Committing nine or more delinquent acts is associated with worse economic outcomes. Those committing nine or more delinquent acts are also the least likely to marry, the most likely to divorce, and the most likely to have never married but have a child at age 33.

3. The Relationship of Adolescent Family Structure to Adult Outcomes

In the previous section we found that engaging in adolescent risky behaviors generally is associated with diminished adult outcomes; the earlier the initiation, the worse the outcome. In this section we examine the role played by adolescent family structure in the relationship to the adult outcomes. As noted in Chapter II, much of the research on family structure has focused on the association between family structure and adolescent risky behaviors where participating in the behavior is the outcome. Few other outcomes have been studied. Those that have been tend to be adolescent outcomes such as high school graduation. Very little has been examined with regard to long-term adult outcomes. We estimate the relationship of adolescent family structure to long-term adult outcomes, holding constant the adolescent's engagement in risky behaviors.

Family structure may have impacts on children that follow them well into adulthood and perhaps their entire lives. The reduced economic circumstances following a divorce can reduce the resources available to invest in the child's future. Acrimony between divorced parents, the divorce itself, changes in living arrangements, or the death of a parent may contribute to psychological problems for the children. Re-marriages may improve economic conditions, but create additional stress on family relationships.

Living in an intact family (two biological parents) is almost universally associated with having a better chance at economic success, and our results are no exception. In general, we find that living in an intact family at age 14 is associated with better outcomes in the economic domain. This is consistently true across all six economic measures. For example, Figure 6 shows the relative likelihood of being in poverty between ages 25-29 for each family type. Those living with both biological parents at age 14 have the lowest probability of being in poverty between ages 25-29. The odds ratios for other family types range between 1.12 for those living with their father and a stepmother to 2.73 for those living in single father families. Individuals living with a single mother at age 14 are 32 percent more likely to spend time in poverty between ages 25-29 than individuals in intact families.

Figure 6

Two biological parent households also see their children with the lowest probability of spending time in jail. There is not much of a clear pattern for either the health or family formation measures. Intact families are neither the best nor worst at preventing poor outcomes in these domains. However, they are generally neutral, not having strong effects either way. For example, they have neither the highest nor lowest rates of marriage, divorce, or childbearing.

Single mother households, the second most common type of family structure, fare worse on the economic outcomes. Compared with intact families, those who lived only with their mother at age 14 do not do as well in the labor market and are more likely to spend time in poverty and on welfare. This is not surprising given that single mother households typically have lower income than intact families. Therefore the likelihood of not achieving economic success as an adult is probably more a function of fewer resources than it is a function of single motherhood per se. This general pattern of the negative relationship between single-mother families and successful economic outcomes is consistent with past research (Amato, 1999; McLanahan, 1999). Even when holding school dropout constant, McLanahan (1999) found higher labor market detachment rates among men from single-parent homes as compared to those from intact homes.

As with intact families, there is not much of a pattern in adult family formation for adolescents living in single mother families. However, those who lived with a single mother (or single father) at age 14 are less likely to have married by age 33 when compared with those who lived in either intact or re-married households. Adolescents in single mother households at age 14 are no more likely than adolescents in intact families to spend time in jail or use drugs as an adult. Interestingly, they are only 71 percent as likely to have an alcohol problem as an adult. These results imply that single mothers may lack the income of intact families, but are still capable of raising children who do not end up as alcoholics, drug users, or in jail.

Because of small cell sizes, we do not emphasize the results for other family types. However, certain findings are worth noting. Adolescents with the least likelihood of having an alcohol or drug problem as an adult are those who live with their father and stepmother at age 14 (42 percent as likely as those in intact families). They are also the most likely to be married, the most likely to be married with children, and the least likely to have divorced at age 33. Paradoxically, they are also the most likely to spend time in jail, spend the most years on welfare, and take the longest to find a steady job.

Those living in a single father household at age 14 are less likely than those in intact families to have adult alcohol or drug problems. They are the least likely to marry and the most likely to not be married but have a child at age 33. They have the highest likelihood of being in poverty, spend the most years in poverty, are the most likely to spend time on welfare, and take the longest to find a steady job. They also have a higher likelihood of spending time in jail than those in intact or single mother households.

In summary, like most literature we find that growing up in an intact family leads to the best outcomes overall, particularly in the economic domain. Growing up in a single mother household makes children more vulnerable to poor economic outcomes, but does not increase the likelihood of other bad outcomes. Those who grow up in either single mother or single father households have the least likelihood of marrying by age 33. In general, it appears that fathers are important for children avoiding adult alcohol and drug problems. But outside intact families, living with a biological father is associated with worse economic, family formation, and incarceration outcomes. One should be cautious about drawing conclusions from these associations. Children are not distributed across different family types randomly. In particular, adolescent behavior may influence reasons for living with a father but not a mother, rather than the other way around. The summary below provides a more detailed description of the findings in this section.

Section Summary

Adolescents living in intact families at age 14 have the best adult economic outcomes and are the least likely to spend time in jail. However, intact families are generally neutral in the adult health and family formation outcomes as the adolescents from these families have neither the best or worst outcomes in these domains.
Adolescents living in single mother families at age 14 have worse adult economic outcomes than those in intact families, but show no difference in adult drug usage or likelihood of spending time in jail. Adolescents from single mother families have lower likelihood of an adult alcohol disorder than adolescents from intact families.
Adolescents living with a single mother or single father at age 14 are less likely to have married by age 33 compared with those who lived in either intact or re-married households.
The presence of a biological father in the household at age 14 is associated with lower levels of adult alcohol disorders or drug usage.

4. The Relationship of Parents' Education (SES) to Adult Outcomes

Parents' socioeconomic status (SES) can play a major role in a variety of ways. Higher income, higher education, and increased access to resources allow parents to invest more in their children's future. They have more access to information about education, jobs, and other opportunities. Furthermore, higher SES parents likely have better information and better access to means of preventing negative consequences such as counseling, drug rehabilitation programs, or good lawyers.

Socioeconomic status is a combination of various attributes including education, occupation, and income. We proxy for SES by using parents' education. While data limitations motivate our choice of education as the variable to study, education is highly correlated with the other attributes and should proxy well for general SES. It should be recognized, though, that our results are specific to parents' education. Had measures of other dimensions of SES been available, we might have obtained different results.

The relationship between parents' education and adult outcomes is fuzzy, but basically follows expected patterns. Mothers' and fathers' education has similar effects for most outcomes, but contradictory for some. They are particularly contradictory for the health outcomes. Higher levels of mother's education are associated with lower likelihood of adult alcohol problems (Figure 7a). For example, those with college graduate mothers are only 69 percent as likely to have an adult alcohol abuse or dependence problem compared with those whose mothers are high school dropouts. Conversely, higher levels of fathers' education are actually associated with greater likelihood of an adult alcohol problem (Figure 7b). For example those with college graduate fathers are 33 percent more likely to have an alcohol abuse or dependence problem compared with those whose fathers are high school dropouts.

Figure 7a

Figure 7a

Figure 7b

While mothers' education has a negative relationship with adult alcohol abuse or dependence, it has a somewhat positive relationship with the likelihood of adult drug usage. Fathers' education has no relationship with adult drug usage. Conversely, while no relationship exists between mother's education and the likelihood of going to jail, higher levels of fathers' education are associated with a lower probability of spending time in jail.

The two parents' education levels have more similar associations in the economic domain. More education for either parent is generally associated with better economic outcomes, though there are some minor inconsistencies. It is clear that those with mothers who are high school dropouts nearly always fare worst and those with college graduate mothers typically fare best. For example, adolescents with mothers who are high school graduates are 71 percent as likely to be in poverty between ages 25-29 (Figure 8) compared to those whose mothers are high school dropouts; those with mothers who are college graduates are only 55 percent as likely. For fathers, whenever significance is achieved, there is a positive association between fathers' education and the economic outcome.

Figure 8

Few associations between either parent's education and the marriage/fertility outcomes achieve statistical significance. One notable exception shows children of college graduate mothers are less likely to be never married with children at age 33. Although generally not statistically significant, we see a consistent relationship between higher levels of parents' education and delayed childbearing. Although children of more educated parents are no less likely to have married by age 33, they are more likely to be never married without children. Children of more educated mothers show higher rates of being married without children, though this relationship does not show up for fathers' education. Individuals with more education typically show rates of later marriage and childbearing. Since there is a positive correlation between an individual's education and his/her parents' education, we are capturing a similar relationship.

Section Summary

Higher levels of mothers' education are associated with fewer adult alcohol disorders and better economic outcomes, but more adult drug usage. There is little relationship with either incarceration or adult family formation.
Higher levels of fathers' education are associated with good economic and incarceration outcomes, but more adult alcohol disorders. There is little relationship between fathers' education and adult drug use or family formation.

5. The Relationship of Parental Alcohol Problems to Adult Outcomes

As noted earlier, parent's alcoholism is unlikely to have positive effects on a child's development. The regressions bear this out strongly. Compared with individuals who do not grow up with an alcoholic parent, a child of an alcoholic parent is 33 percent more likely to have an adult alcohol problem, 21 percent more likely to use drugs as an adult, and 32 percent more likely to spend time in jail. They do not fare any better on the economic front. They are 22 percent more likely to spend time in poverty and 41 percent more likely to receive welfare at some point. Although they do not have lower percent time employed, they take nearly ten months longer to find a steady job. The p-values indicate that these results have relatively low probabilities of occurring by chance.

On the other hand, few relationships between parental alcohol use and marriage/fertility outcomes are significant. There is no difference in the likelihood of being married at age 33.¹⁹ However, consistent with the literature, we find that ACOAs are more likely to divorce (Parker and Harford 1988). In fact, the only category that achieves statistical significance is married and divorced with children. About 30 percent of those who marry are divorced by age 33, while about 27 percent of those without an alcoholic parent who marry are divorced.

Section Summary

Compared with individuals who do not grow up with an alcoholic parent, a child of an alcoholic parent is more likely to have an adult alcohol problem, is more likely to use drugs as an adult, and is more likely to spend time in jail.
Adult children of alcoholics are more likely to spend time in poverty and are more likely to receive welfare at some point. Although they do not have lower percent time employed, they take longer to find a steady job.
Few relationships between parental alcohol use and marriage/fertility outcomes are significant; however, adult children of alcoholics are more likely to divorce.

Endnotes

13. The one exception is the measure of age when first held a job for at least two years. The fact that this age is higher for those with some college most likely reflects the fact that these individuals were in school longer and began their transition to a steady job later.

14. This does not include pre-marital births which occurred before second or later marriages, nor does it include other non-marital births.

15. In earlier estimations we included several other variables. However, their inclusion did not change the results so we settled on a more parsimonious specification to aid interpretability.

16. Not statistically significant.

17. Alternatively, the result may merely reflect the timing of our measurement. If bad economic outcomes dissipate over time, then the measured impact for late initiators would have had less time to diminish.

18. Those who initiated sex later may also have married later, leaving less time in which to have gotten divorced.

19. Since we measure marriage at age 33, we are not able to replicate the finding in the literature that ACOAs are likely to marry early.

Chapter VI: Exploring Pathways of Early Initiators

A. Introduction

In the previous chapter, we have shown that in general, adolescent risk taking is associated with diminished adult outcomes. In particular, the earlier one initiates into a risky behavior, the worse the outcomes will be. However, not all youths engaging in risky behaviors fare poorly. Two youths both choosing to initiate early into a risky behavior may have different outcomes as adults. The divergent pathways may be the result of many factors. To understand why these outcomes may differ requires examining the pathways followed from adolescence to adulthood. Among the factors that may influence these pathways is the environment within the family. We have seen relationships between adolescent family structure and parents' education and adult outcomes. Yet these likely proxy for additional processes occurring within the family.

This chapter seeks to lay the groundwork for future research into the transition from adolescence to adulthood, focusing on one point along the path. We have not modeled the choice of when to engage in a risky behavior; instead we begin at a fork in the path where the adolescent has made that choice. Those choosing early initiation are taking themselves down a path with an increased likelihood of bad outcomes. Given that they have chosen this fork in the path, what factors can help prevent them from facing bad outcomes? What factors will return them to the "right" path?

To address this question, we limit our sample to "early initiators" within the various adolescent risky behaviors, i.e. those who have chosen the "bad" fork in the path. We estimate a set of regressions similar to those in Chapter V and continue our focus on family structure and parents' education. For this analysis, we address whether adolescent family structure and parents' education can help early initiators avoid bad outcomes.²⁰ Specifically, we examine the relationship between adolescent family structure and parents' education and long-term adult outcomes for early initiators.²¹

It is possible that the impact of family structure and parents' education is the same for early initiators and late initiators, or the effect could be in the same direction but of a higher or lower magnitude. While the relative importance of these variables between early and late initiators would be interesting, we do not compare them here. We want to know exclusively about early initiators. In particular, it is quite possible that family structure and parents' education have no impact on early initiators. As noted above, if they didn't prevent the early initiation, maybe they can't help after the fact. That is the question we address in this chapter, whether family structure and parents' education have an impact on the adult outcomes of adolescents who choose to initiate early into one or more risky behaviors.

In this analysis, we estimate a unique regression for each behavior-outcome pair. For example, there is a regression for the outcome "ever been in poverty between ages 25-29" for early initiators into alcohol use, another for early initiators into marijuana use, yet another for early sex initiators, and so on. The definition of an early initiator varies by both the behavior and the outcome being studied and is determined empirically from the regressions in Chapter V. For each behavior-outcome pair, we define early initiators by starting with the earliest age group (11-15) and then combining with any age groups for which there was no significant difference in the association of the behavior and the outcome. The combined group is then referred to in this context as "early initiators." The age groups that are not included are deleted from the analysis.²² This means that the definition of "early initiator" will be different for each behavior-outcome pair. For example, early initiators into alcohol use are defined as ages 11-17 when examining the outcome of adult alcohol abuse or dependency, but are ages 11-19 when examining adult drug usage.

No significant difference across the four age groups implies that early initiation into the behavior is not different than late initiation with regard to that particular outcome. Since we are concentrating on the "bad" paths chosen by early initiators, these behavior-outcome pairs were dropped from the analysis. For example, the age of alcohol initiation was not associated with any economic outcomes except ever being on welfare. Therefore, for the other five economic outcomes, we do not examine further the relationship between alcohol initiation and those outcomes.

Although delinquency is not measured by age of initiation, but rather by the number of delinquent or criminal acts in 1980, we refer for convenience to those who engaged in greater numbers of such acts as "early initiators." In this case, we begin with those who committed nine or more delinquent or criminal acts, then combine with lesser numbers of acts if there is no significant difference in the relationship of that number of acts and the outcome being studied. Appendix C shows the ages used to define "early initiators" for each analyzed behavior-outcome pair.

B. Does Adolescent Family Structure Affect the Pathways of Early Initiators?

In this section we examine family structure as a measure of family environment. Other literature has examined how family structure may influence the choices made by these youths. However, as before, we take as given the point at which youths initiate into a risky behavior. We frame the question of interest around early initiators. Specifically, does adolescent family structure affect the outcomes for early initiators into risky behaviors? Although certain family types may have reduced the likelihood of initiating a certain behavior, nevertheless some adolescents will choose to do so. If those family types did not prevent the behavior from beginning, can they still help guide the adolescent to avoid deleterious long-term effects from engaging in the behavior? It is possible that certain family types help late initiators, but do not help early initiators. We limit our analysis specifically to address whether family type matters for those who have taken a "wrong" fork in the path.

The full set of regressions is found in Appendix D. A summary table isolating the family structure relationships precedes the full set of regression tables. It is not possible to discuss all results, but certain patterns emerge regarding early initiators as a whole. For nearly any adolescent risky behavior, living in an intact family reduces the likelihood of a bad adult economic outcome for an early initiator. In 19 of the 24 behavior-outcome pairs examined in the economic domain, those living with both biological parents at age 14 have the least, or nearly least, likely bad outcomes of any family type. They also generally have among the lowest probabilities for spending time in jail.

As we saw in the previous section, two biological parent families are not the best for preventing adult alcohol and drug problems, though they are not much different from other family types except fathers with stepmothers. Remarried fathers are generally best for minimizing the likelihood of a bad health outcome. Adolescents in single father families also again have lower probabilities of adult drug usage, though no real difference on adult alcohol problems. Taking these together, it once again appears that fathers are important for reducing the probability of an adult substance abuse problem, even for those at most risk due to early initiation.

Adolescents living in single mother households at age 14 generally fare worse in economic outcomes than do those in intact families, but better than those living in other family situations. Similarly they fare among the best at keeping out of jail. As seen before, single mothers actually do better than two biological parents at reducing the probability of an adult alcohol or drug problem, particularly alcohol abuse or dependence, given early initiation.

Marriage rates are lowest among those who were living with only one parent, while fairly similar across other family types. Adolescents living with only one parent at age 14 are also the most likely to be never married with children. On the other hand, conditional upon having married, divorce rates are highest among those who were living with their mother and a stepfather (35 percent) or with no biological parent (37 percent). Rates for other family types are consistently lower (22-27 percent).

While the direction of effects of adolescent family structure for early initiators is similar to those found for the sample as a whole, many magnitudes are different. Because there are so many comparisons one could make and this is not our focus, we simply note that adolescent family structure is associated with adult outcomes for adolescents who engage in risky behaviors at early ages. The extent to which these relationships are similar to, or different from, those that would be found for late (or non-) initiators is not critical. It is important to know that there are family forces that can influence the pathways of these early initiators such that the likelihood of bad outcomes can be reduced.

Section Summary

Relationships between family structure and adult outcomes for early initiators are similar to those for the whole sample.
For nearly any of the adolescent risky behaviors studied, early initiators living in intact families at age 14 have lower likelihood of a bad adult economic or incarceration outcome, but have a neutral association with adult alcohol and drug problems.
Early initiators living in single mother families at age 14 have worse adult economic outcomes than those in intact families, but better than those in other family situations. They also have less likelihood of an adult substance use problem.
Presence of a biological father in the family at age 14 is associated with fewer adult substance abuse problems for early initiators.
Early initiators living with only one parent at age 14 is associated with the lowest levels of marriage and the highest levels of never having been married but having a child by age 33.

C. Does Parents' Education Affect the Pathways of Early Initiators?

We have addressed the question of whether adolescent family structure may influence the paths of early initiators and found that intact families are generally the best at helping adolescents avoid most negative consequences of risky behaviors. However, even in intact families, some adolescents will develop adulthood problems while others do not. In this section we go deeper into the family environment and consider another measure, the family's socioeconomic status, again represented by parents' education. Two adolescents from the same family type, both of whom initiate early into one or more risky behaviors, may face different consequences. One may end up with adulthood problems while the other doesn't. Within family types, what makes the difference? We seek to determine if parents' education plays a role in influencing the pathway for early initiators within a given family type.

In addition to limiting the sample to early initiators, we further restrict the analysis to sample members who lived in either of the two most common family types at age 14: two-biological parents and single mother households.²³ To study the impact of parents' education, we estimate the same regressions as above. We estimate each set of regressions separately for the two family types. This allows us to delve down a layer within the family, examining whether parents' education matters for early initiators within these family types. We use the same definitions of early initiators and again do not estimate any behavior-outcome pairs that do not have significant differences across the age of initiation categories. The full set of regressions appears in Appendix E (two biological parent families) and Appendix F (single mother families). A summary table isolating the relationships for each parent's education precedes the full set of regression tables in each appendix.

Restricting the sample to two-parent families only, we find that higher levels of parents' education generally reduce the likelihood of a bad adult economic outcome among early initiators. Small cell sizes make the numbers bounce around a lot and statistical significance frequently is not achieved. However, the direction seems consistent. The pattern seems to also hold for the likelihood of being incarcerated, but virtually none of the parameter estimates are statistically significant. Cell sizes dramatically impair our ability to draw conclusions in the family formation domain. About the only result that is consistent is that families with either parent having completed college have children who are less likely to be unmarried with children at age 33.

As with the full sample, the connection between parents' education and the likelihood of an adult problem with alcohol or drugs is contradictory. In general, in intact families mothers' education reduces the likelihood of an adult alcohol problem and increases the likelihood of an adult drug problem while fathers' education has exactly the opposite effect in both cases. We are reluctant to place too much emphasis on these findings since virtually none of the parameter estimates is statistically significant. The most important thing to consider, though, is how this fits with our earlier findings. Intact families were seen not to be the best for minimizing the likelihood of an adult alcohol or drug problem. Also, higher levels of parents' education were not always associated with better adult health outcomes. Now we observe seemingly quirky results for the impacts of mothers' and fathers' education when examining early initiators within intact families exclusively. Clearly there are processes within these families that are not captured here that influence the pathways to adult alcohol and drug problems.

When we restrict the sample to those who lived in single mother households when they were age 14, we examine the impact of education including the mother's education only.²⁴ Interestingly we see a pattern similar to the intact families. Greater mother's education for the most part reduces the likelihood of a poor economic outcome and has no clear effect on the probability of incarceration. As in intact families, the only adult family formation result that is clear is that college educated mothers are less likely to have offspring who are unmarried with children at age 33.

Once again, we also see that more education for the mother reduces the likelihood of an adult alcohol problem, but increases the likelihood of adult drug usage. It is interesting that this seemingly strange result holds for the case where we have limited the sample to those who lived in single mother households and we have included only the mother's educational attainment.

Section Summary

Relationships between parents' education and adult outcomes for early initiators are similar to those for the whole sample, even when restricting family type (at age 14) to either intact families or single mother families.
Higher levels of mothers' education are associated with less adult alcohol disorders, more adult drug usage, better economic outcomes, less incarceration, and lower likelihood of being unmarried with children at age 33 for early initiators regardless of family structure at age 14.
Higher levels of fathers' education is associated with more adult alcohol disorders, less adult drug use, better economic outcomes, less incarceration, and lower likelihood of being unmarried with children at age 33 for early initiators living in intact families at age 14.

Endnotes

20. An alternative question we could have posed would be whether family environment mitigates the impact of risky behaviors on adult outcomes for early initiators as compared with late initiators. We have chosen not to frame the question this way because we want to examine the effects of family environment given the choice to initiate at early ages. Those who initiate at later ages or never initiate are following a different path. Family environment may involve different processes for these groups. In addition, since our oldest group includes both late initiators and non-initiators, we cannot disentangle the influence of family environment for these two groups.

21. By limiting the sample to early initiators, the model is equivalent to fully inter-acting "early initiator" with all the variables in the regression, but only examining the results for the early initiators. Thus the estimated values for family structure or parents' education are similar to what would be obtained from interacting family structure and parents' education with age of initiation in the regressions in Chapter V. This functional form also allows all the other variables to have effects specifically for early initiators

22. This method of combining age groups to define early initiators makes the functional form only similar to, but not the same as, the fully interactive model described above.

23. Small sample sizes limit our ability to study other family types. Furthermore, this restriction provides the cleanest way to address the issue. In these two family situations, the measures of mothers' and fathers' education specifically relate to the parents with whom the youth has lived and do so exclusively. This would not be true in other family situations. For example, to study a family comprising the child's mother and a stepfather would require knowing the education of the mother and stepfather, and perhaps also the biological father. We would need to know the timing of the divorce (if there was one) and the re-marriage and the degree to which the father has stayed involved financially and directly with the child. Even knowing all of this, we would have to estimate a complicated relationship between the influence of different parents' education, biological and step, on the child. This places too many demands on the data. For single mother households, we face a similar, though not as complex problem. We do not know the financial or time contributions made by the absent father. However, because this family type is so common, we include it. Since we can only look at mothers' education, there may be significant omitted variable bias without a measure of father's education.

24. We also estimated the regressions including father's education to examine whether there were differences. However, some of the values lacked credibility, i.e. they were orders of magnitude different than previously observed and changed much too dramatically from the regression without father's education. At best, these estimates would be difficult to interpret since we do not have measures of the father's financial contribution or level of involvement with the youth.

Chapter VII: Discussion and Future Directions

In this study we have addressed a set of issues which have generally not been analyzed previously. We first examined the relationship between adolescent risky behaviors and long-term adult outcomes. Five adolescent behaviors were studied: alcohol use, marijuana use, cocaine use, sexual activity, and delinquency. Multiple outcomes were measured across four different domains: health (alcohol abuse or dependence, drug use), economic (poverty, welfare, employment), family formation (marriage, divorce, fertility), and incarceration.

There is a fairly consistent pattern that engaging in risky behaviors as a teenager is associated with less successful adult outcomes. For many, the relationship is such that the earlier one engages in the behavior, the more likely they face a bad outcome as an adult. The most consistent predictor of a bad adult outcome is age of initiation into sexual activity. Alcohol usage, on the other hand, is perhaps the one teenage behavior least associated with bad adult outcomes. Age of initiation into alcohol usage is, however, associated with adult alcohol abuse or dependence.

Age of initiation into any given risky behavior is highly correlated with the frequency and intensity of engagement in that behavior. However, in our analysis, those who merely experiment are combined with those who have more serious intentions. By not measuring frequency and intensity, we cannot distinguish between the effects of these measures from the effects of early initiation.

These findings represent associations and the relationships should not be interpreted as causal. Many are consistent with what might have been expected, given what is known from other literature. However, few relationships with long-term adult outcomes have been documented. This study provides new findings from a nationally representative survey and contributes to our understanding of the impacts of risk taking by young people. This study lays the groundwork for developing future work to understand these relationships.

The second avenue of investigation in this study was to document the relationship between family environment and adult outcomes. We select three measures of family environment for analysis: family structure, parents' education (as a proxy for socioeconomic status), and parental alcoholism. Most research on the impact of family structure has focused on how it relates to initiation and participation in risky behaviors. Other outcomes studied have generally reflected short-term outcomes such as high school graduation, college enrollment, and teenage pregnancy. This study focuses on a wide variety of long-term outcomes, providing some of the first estimates of such relationships.

Our findings indicate that family structure effects differ by outcome domain measured. Adolescents who reside in intact families at age 14 clearly have the least likelihood of a bad economic outcome and are less likely to spend time in jail. This is less clear for the health and family formation domains. Single mother headed households do not fare as well as intact families in economic outcomes, but compare favorably along other domains. Interestingly, fathers with stepmothers lead to the least likelihood of an alcohol or drug problem as an adult, but also lead to the highest likelihood of having bad economic outcomes or spending time in jail. Single father families have similar relationships implying that fathers are important for reducing the chances of an adult problem with alcohol or drugs.

We found similar associations when examining the relationship between parents' education and long-term adult outcomes. More educated parents are associated with better economic outcomes and less likelihood of going to jail. However, there are contradictory results for the health domains where mothers' education is associated with lower likelihood of adult alcohol problems, but greater likelihood of adult drug usage. Fathers' education had exactly the opposite associations across the two outcomes.

While interpreting any given relationship would be problematic without an underlying behavioral model, we believe our results suggest that the way in which parents help prevent bad outcomes for their children differs across different domains. For economic outcomes, parents can use their resources (financial, networks, etc.) to send their children to college, help them get jobs, and serve as a fallback to prevent financial problems. There is probably a similar mechanism for keeping their children out of jail. However, for non-economic domains, particularly those of alcohol and drug problems, a different set of family processes contributes to an eventual healthy adulthood.

Our third avenue of investigation involved exploring the influence on the pathway from adolescence to adulthood. We restricted our analysis to those who had already chosen to engage in one or more risky behaviors at an early age, i.e. those who had chosen a path more prone to poor outcomes. We then examined the relationships of family structure and parents' education to adult outcomes specifically for this sample. The results are comparable to those found when not restricting the sample. They indicate that these two family environmental measures are associated with changing the paths these early initiators have taken. There is a sense that although the family may not have prevented the youth from starting down a "wrong" path, it can help them from having that choice lead to bad consequences.

These findings perhaps raise more questions than they answer, laying the groundwork for future research. Several aspects of this study can be varied to create a richer analysis. First, we began our examination taking initiation into each risky behavior as given. However, a complete model would treat these decisions as endogenous and model the age of initiation. Such a model would be critical for teasing out the role family environment plays in the choices to engage in adolescent risky behaviors versus its impact after the choice has been made.

Second, although our regressions control for initiation into several behaviors, they do not explicitly account for multiple risk taking. The idea of multiplicity of risk factors is one discussed in the literature but not often tested. From the Problem Behavior Theory perspective, adolescents do not usually engage in one risky behavior, but rather the problems form a cluster of behaviors comprised of drinking, bingeing, smoking, marijuana use, delinquency, and premarital sexual intercourse (Donovan, et al., 1991). An important next step in our research may be to form a risk behavior composite index and examine its predictive ability in terms of adult outcomes.

Third, we have treated family structure at age 14 as exogenous. While some living situations are not reflective of the child's behavior (e.g. divorce), other situations may be the outcome of the child's behavior. For example, a boy who has become involved with drugs and delinquency may be difficult for a mother to handle so she sends him to live with his father. Therefore the endogeneity of family structure needs to be accounted for.

Fourth, we have used family structure at age 14. It was necessary for us to simplify our analysis and age 14 seems to reflect the time when adolescents are beginning to make decisions about whether to engage in various risky behaviors. However, the decision to engage in such behavior, and its impact, is probably the result of a more dynamic process. For example, changes in family structure that occur early in the child's life may have a different impact than those that occur later. And the number and type of changes may have differential impacts. A complete dynamic structural model would be necessary to account for the connection between risk taking and family changes.

Fifth, we have examined the relationships of family structure and parents' education on certain adult outcomes. These, however, are surely proxies for underlying family processes that we are not measuring such as parents' relations with each other for residential parents and with absent parents and parent-child interactions for residential and absent parents. Past research has shown the key role of certain family process variables as risk and protective factors for children whose parents have substance use disorders (Johnson et al., 1995). Some of these measures include family discord, hostility, care giving quality, family cohesion, supportive parenting and monitoring. Recent releases of new data sources such as Add Health and the NLSY97 provide more information on these family processes than was previously available. It will be quite some time, though, before we have observed long-term adult outcomes in those samples.

Sixth, recent cohorts may not behave similarly to the teenagers of the early 1980s. For instance, although adolescent drug usage has been slowly rising in the 1990s, the rates are considerably lower than in the early 1980s. There is evidence that alcohol use at early ages has become steadily more common in the last several decades, but early marijuana use has shown a different trend. Johnson and Gerstein (1998) found that the percentage of 15 year olds who had ever used alcohol or marijuana was higher for the 1961-1965 birth cohort compared to the 1956-1960 birth cohort (25 percent versus 20 percent, alcohol; 17 percent versus 12 percent, marijuana), the two groups which roughly comprise the sample used in this study. Those born in 1971-1975 showed higher rates of alcohol usage at age 15 at 33 percent but no increase in marijuana usage at 13 percent. As risk taking behavior changes across cohorts, it is an open question whether there is a change in the relationship between the behavior and adult consequences.

Various data sets can be used to compare inter-cohort risk taking. However, longitudinal data sets of more recent cohorts such as Add Health, the NLSLY97, and the National Survey of Adolescent Males have not existed long enough to provide long-term outcomes. As these data sets mature, research will be possible to determine if long-term relationships have changed across cohorts.

Seventh, a great deal could be learned about the impact of family structure and family process if one had longitudinal data on multiple generations. Most data sets have only a limited set of information about the respondents' previous or next generations. One exception is the Children of the NLSY79. This data set includes a set of assessments and interviews with the children of the female respondents in the NLSY79. Begun in 1986 and administered biennially since, the Children of the NLSY79 collects information to assess the physical, social, and cognitive development of these children. In 1994, sufficient numbers of these children had reached age 15. At that time, a full interview with these youths was conducted, similar to the one their mothers had received. This sample has become known as the Young Adult sample. The NLSY79 with its accompanying child supplement and young adult interview provides us with extremely rich detail on the lives of children and their parents. In each case information on the mothers exists for many years, extending back to the mother's own adolescence. With data collected through 1998, there are not only lengthy longitudinal data on mothers, but there are now multiple observations for a substantial number of youths. These data will provide an opportunity to study intergenerational relationships in adolescent risky behavior.

Finally, it is clear that research needs to differentiate among different outcome domains. We found important differences in the relationships of family environment across the different outcome domains studied. The underlying family process affecting the various outcomes is undoubtedly different. Introducing to a model such things as parents' income and assets, parent-child interactions, and contributions and involvement by absent parents would enrich our ability to understand these distinctions.

Appendices

Appendix A: Variables and Estimation Methods

Description of Variables

This section describes the creation of variables used in the analysis of this report. We group the variables into four categories: adult outcome variables, youth risky behavior variables, family environment variables, and other explanatory variables of adult outcomes.

Because of the longitudinal nature of the data, many variables were constructed using multiple waves of data from the NLSY79. Means and sample sizes for each variable used in the analysis are presented at the end of this section.

Adult Outcome Variables

The adult outcomes have four domains with a total of 10 measures.

The health domain includes two measures: alcohol abuse or dependence around age 30; drug use (marijuana or cocaine) in the past month around age 30.

A series of questions, asked during the 1982-1985, 1988-89, 1992, and 1994 surveys, elicited information on the development of drinking patterns, quantity of various alcoholic beverage consumed, frequency of use, impact of consumption on schoolwork and/or job performance, and types of physiological and behavioral dependency symptoms. Although types and wording of alcohol abuse or dependence questions vary over the years, the items in the1989 and 1994 surveys are almost identical. We use 1989 data for those born in 1957-1960 and the 1994 data for those born in 1961-1964. The respondents were between ages 29-32 during these years. Diagnosis of current alcohol abuse and dependence were derived using a virtually identical set of 29 symptom-item questions in the1989 and 1994 surveys designed to operationalize the DSM-III-R abuse and dependence. The creation of the measure of alcohol abuse or dependence follows a set of complicated criteria provided by Harford, and Grant (1994). Alcohol abuse and alcohol dependence are different measures of alcohol problems. We combine them to form a single variable that takes the value of one if either condition is met.

Questions on substance use were included in the 1984, 1988, 1992, 1994, and 1998 surveys. Among other usage information collected, these surveys collected information on the most recent use of marijuana and cocaine, from which the measure of past-month drug use was created. The 1988 data were used for those born in 1957-1960 and the 1994 data were used for those born in 1961-1964. Respondents were ages 28-32 in these years. Past-month drug use measures whether a respondent used marijuana or cocaine in the past month. We chose not to use a measure of drug use in the past year, another common measure, in order to reduce the likelihood of measuring occasional recreational use. We implicitly assume that adults who used these drugs in the past month are more likely to be regular drug users.

The economic domain has six measures: ever under the poverty line between ages 25-29; number of years in poverty between the ages of 25-29; ever on welfare (AFDC/TANF or food stamps) between ages 21-33; number of years on welfare between ages 21-33; percent of employable time spent employed between the end of formal schooling and age 33; and age when found a "steady" job, i.e. working at least 2 years for a single employer since leaving formal schooling.

Variables have been created in the NLSY database for each survey year (1979-1998) that indicate whether or not a respondent's total family income for the past calendar year was above or below the poverty level for a given family size. Our two poverty variables were created based on these variables from 1979-1994. The interview frequency of the survey changed to biennial after 1994 and the poverty information was not available for the missing years. As a result, the oldest we could observe all individuals is age 29. We chose 25 as the starting age to avoid classifying respondents in college as in poverty. If a respondent was in poverty in any of the years when he/she was 25-29 years old, a value of "1" was assigned to the respondent for the measure of ever under the poverty line between ages 25-29; otherwise, a value of "0" was assigned.

The "income" section of each year's questionnaire collects information on amounts and time periods during which cash and non-cash benefits were received from various sources of public assistance. The universe and type of data collected varies across survey years. This report focuses on AFDC/TANF and food stamps, since they are the major components of public assistance and were consistently collected over all survey years. Unlike poverty, information on welfare was collected using event histories and therefore information is available for every year from 1978-1997 even if the survey was not fielded in some years. Similar to the poverty measures, if a respondent received AFDC and/or food stamps in any of the years when he/she was 21-33 years old, a value of "1" was assigned to the respondent for the measure of ever on welfare between ages 21-33; otherwise, a value of "0" was assigned. The measure of years on welfare between the ages of 21-33 sums the years on welfare between the ages of 21-33.

The other two economic measures (percent time employed and tenure) were created using both the NLSY79 main and Work History Files. The NLSY79 Work History File was constructed from work experience data collected during the main NLSY79 surveys. It provides a week-by-week longitudinal work record of each respondent from January 1, 1978 through the current survey date. Since continued education may take away time from working, to give every respondent a consistent starting point in counting his/her percent time employed and tenure, we created the measure of percent time employed and the measure of tenure from the end of formal schooling to the age of 33. The age at which a respondent ended his/her formal education was created using two sets of created variables that summarize each respondent's school enrollment status and highest grade completed as of May 1 of each survey year from the NLSY79 main file. ²⁵

"Percent of time employed" for each respondent is calculated by the ratio of weeks employed over weeks in the labor force between the end of formal schooling and age 33. This measure is intended to measure attachment to the labor force. We chose to limit this measure to time in the labor force so as not to "penalize" women for childbearing. Although some individuals maintain an attachment to the labor force, they have difficulties maintaining a steady job. We define "steady job" as one lasting at lest two years. For a discussion of measuring the transition from school to work, see Pergamit (1995).

To calculate age when first reached two years of tenure after the end of formal schooling, we first calculated the age when started each job listed. Then the starting age was compared with age when formal schooling ended. If the job was started before formal schooling ended, then the starting point for the job was replaced by age when formal schooling ended. Tenure was calculated for each job listed by the difference between the starting and the ending point of each job. Age when first reached two years of tenure was obtained by choosing the minimum age at jobs where two years of tenure was reached. One limitation of this variable is that college graduates cannot achieve two years of tenure at as early ages as those with less education (see Table 6). Thus, they appear to do worse for this outcome. The impact on the analysis, however, is small since college graduates generally attain two years of tenure quickly with fewer such individuals in the right tail of the distribution.

The family formation domain has one measure, which is a measure of six combinations of marriage and fertility status at the age of 33. Marital status and fertility status at the age of 33 were first created separately.

A series of edited Supplemental Fertility File variables that reflects the beginning and ending dates of marriages was constructed for 1982 through 1998. This information is derived from the marriage section of the NLSY79 questionnaire. We used the information to create length of first marriage and assigned marital status at the age of 33. Marital status at the age of 33 has three categories: never married, married and stayed married, and married but divorced at the age of 33.

Beginning in 1982, every wave of the NLSY79 data release has included a created variable that tracks the age at which respondents first had a child. This information was used to create fertility status at the age of 33, that is, whether a respondent had a child by the age of 33.

The marital status variable and the fertility status variable were then combined to create a six-combination measure of marriage and fertility status at the age of 33. The categories of this measure are: never married without children, never married with children, married without children, married with children, married but divorced without children, and married but divorced with children.

The crime domain has one measure: ever been in jail by the age 33. NLSY79 respondents are followed and interviewed even when they enter an institution. Interviewers designate the "Type of Residence" which identifies those respondents who resided in jail at each interview date. This information was used to create a binary measure of ever being in jail by the age of 33. Since some jail terms will have occurred between interviews, this is an underestimate of the number of respondents who ever spent time in jail.

Youth Risky Behavior Variables

We explore five youth risky behaviors: alcohol usage, marijuana usage, cocaine usage, sex activity, and delinquency. The first four behaviors are measured by age of initiation. Age when started a risky behavior was asked in multiple waves of interviews. In general, the value usually taken was from the earliest year the question was asked. However, to overcome potential data entry errors, if the reported age of initiation is less than 11, then a later year entry was taken. Those who reported age of initiation less than 11 were included in the youngest age group. Age when started to use alcohol was asked in both the 1982 and the 1983 interview. Age of marijuana and cocaine initiation was created using the 1984 and 1998 questions on drug initiation. Age of sex initiation was created using the 1983, 1984, and 1985 questions on sex initiation. Age of initiation is grouped into four categories: 11-15, 16-17, 18-19, and not by age 19 (i.e. after age 19 or never initiated).

Delinquency was measured as the total number of delinquent and/or criminal acts reported in the 1980 interview. The 1980 NLSY9 included a self-reported section detailing respondents' participation in and income from delinquent or criminal activities such as skipping school, alcohol/marijuana use, vandalism, shoplifting, drug dealing, robbery, assault, or gambling during the previous twelve-month period. Alcohol and drug use were not included so as not to double-count with the other risky behavior measures. Measures of the total number of delinquent/criminal acts and the types of delinquent/criminal acts were both created. Since distributions of both measures are very similar, we opt to use only the first measure. This measure could be problematic because of the age differences among respondents in 1980. However, the chosen measure predicted very well the likelihood of spending time in jail. Further research could be considered to separate personal versus property crimes, frequently distinguished in the criminal behavior literature. The number of delinquent/criminal activities was grouped into four categories: no delinquent/criminal acts, 1-2, 3-8, and 9 or more.

Family Environment Variables

Family Structure: In 1979, data were collected on whom respondents lived with at the age of 14. Twenty-eight categories of living arrangements were collapsed into six family structure categories: living with both biological parents, with single mother, with single father, with mother and stepfather, with father and stepmother, and other relatives and/or non-relatives (including institutions). If a mother (father) was living with a man (woman) to whom she (he) is not married, we consider that person a stepfather (stepmother).

Parental Educational Achievement: Highest grade completed was collected for respondents' biological mothers and biological fathers in 1979. Parental educational achievement was broken down into four categories: high school dropout, high school graduate, some college, and college graduate.

Parental Alcohol Problems: The 1988 NLSY79 interview collected family history on alcoholism and problem drinking. Respondents were asked whether they had any relatives who had been alcoholics/problem drinkers at any time, their relationship to the alcoholic relatives, and the number of years they lived with the alcoholic relatives. Based on this information, we created a binary measure of whether a respondent had at least one alcoholic parent whom he/she lived with for at least one year.

Other Explanatory Variables

Sex: Individual respondent was assigned the value of "1" if he/she was identified as "male", otherwise the value of "0" was assigned.

Race and Ethnicity: The NLSY79 distinguishes among three mutually exclusive and exhaustive race-ethnicity groups. A respondent was designated as "non-black, non-Hispanic", "black", or "Hispanic" based on a racial/ethnic variable from the NLSY79 Screener File. Since Hispanic is a mutually exclusive group, this classification is different than used for Census Bureau tabulations where Hispanics can also be classified as "black" or "white."

Educational Achievement: Data have been collected during each NLSY79 survey on respondents' current school enrollment status, highest grade attended and highest grade completed. Information on highest grade completed from 1979 through 1998 was compiled to obtain a measure for the respondent's educational achievement. There are three categories for this measure: high school dropout, high school graduate, and some college and up.

Rosenberg Self-Esteem Scale: The Rosenberg Self-Esteem Scale administered during the 1980 interview was used to construct a summary score that measures the self-evaluation that an individual makes and customarily maintains. It describes a degree of approval or disapproval toward oneself (Rosenberg, 1965). The summary score ranges from 4 to 40, with higher scores designating higher self-esteem.

Table A1. Means and Sample Sizes of Variables Used in the Analysis
Variable	Sample Size	Weighted Mean
Youth Risky Behavior Variables
Age of Alcohol Initiation
11-15	9717	20.4%
16-17	9717	37.2%
18-19	9717	26.9%
not by 19	9717	15.5%
Age of Marijuana Initiation
11-15	8595	23.2%
16-17	8595	22.3%
18-19	8595	12.1%
not by 19	8595	42.4%
Age of Cocaine Initiation
11-15	8569	1.2%
16-17	8569	4.7%
18-19	8569	7.5%
not by 19	8569	86.5%
Age of Sex Initiation
11-15	9640	21.7%
16-17	9640	34.7%
18-19	9640	23.8%
not by 19	9640	19.8%
Number of Times Committed Crimes/Delinquencies
9+	9320	15.1%
3-8	9320	25.3%
1-2	9320	22.7%
0	9320	37.0%
Adult Outcome Variables
Alcohol Abuse or Dependence around Age 30	9986	13.8%
Past-month Drug Use around Age 30	8803	11.1%
Ever Being in Jail by age 33	9986	3.7%
Ever in Poverty between Ages 25-29	9986	23.5%
Ever on Welfare between Ages 21-33	9986	25.0%
Number of Years in Poverty at Ages 25-29	9430	0.51
Number of Years on Welfare at Ages 21-33	9986	1.08
Percent Time Employed Between the end of Formal Schooling and age 33	9801	0.91
Age when Reached 2 Years of Tenure after Formal Schooling	8467	24.97
Marital and Fertility Status at Age 33
Never married without children	9144	15.3%
Never married with children	9144	6.4%
Married without children	9144	11.9%
Married with children	9144	43.6%
Married but divorced without children	9144	5.2%
Married but divorced with children	9144	17.7%
Other Explanatory Variables
Male	9986	50.9%
Race/Ethnicity
Non-black non-Hispanic	9986	79.3%
Black	9986	14.2%
Hispanic	9986	6.6%
Educational Attainment
High school dropout	8399	9.5%
High school graduate	8399	42.8%
Some college and up	8399	47.6%
Rosenberg Self-esteem Scale (4-40)	9555	32.46
Family Type
Both biological parents	9847	75.2%
Single mother	9847	12.1%
Single father	9847	1.2%
Mother and step father	9847	7.6%
Father and step mother	9847	1.8%
Others (relatives. non-relatives, institution)	9847	2.2%
Mother's Education Attainment
High school dropout	9327	32.0%
High school graduate	9327	46.9%
Some college	9327	11.1%
College Graduate	9327	10.0%
Father's Education Attainment
High school dropout	8478	33.7%
High school graduate	8478	36.8%
Some college	8478	11.1%
College Graduate	8478	18.4%
Biological Parent with Drinking Problem	8094	20.3%

Description of Multivariate Regression Estimates

To obtain accurate estimates of the effects of youth risky behaviors and family environment on adult outcomes, different types of regression techniques were utilized for different adult outcomes. Many econometric textbooks provide good discussions of the principles and practice of the regression analyses used in this study, e.g. Greene (2000), Judge, et al (1985). These regression techniques were applied to assess the independent effects of youth risky behaviors and family environment on adult outcomes. By independent effects, we mean the effects after other important factors affecting adult outcomes have been controlled or adjusted. These "effects" are associations between the dependent and independent variables and do not necessarily reflect a causal relationship.

OLS Regression

Ordinary Least Squares (OLS) regression is used to estimate continuous outcome variables that are normally distributed. In this report, percent time employed between the end of formal schooling and the age of 33 is estimated using OLS regression.

Since OLS regression assumes a linear function, the interpretation of estimated coefficients is simple and straightforward. Estimated coefficients from OLS regressions measure changes in the outcome variable resulting from a unit change in an explanatory variable. For example, in estimating percent time employed, we obtained an estimated coefficient of 0.02 for males relative to females. This means that the percent time employed for male respondents was on average 0.02 higher than that of female respondents.

Most explanatory variables in our analysis are categorical variables. In estimating outcome models, one of the categories of each categorical variable has to be dropped due to co-linearity. The omitted category becomes the reference group. Any estimates for other categories become relative to the reference group. For example, age of marijuana initiation has four categories: initiated at ages 11-15, at ages 16-17, at ages 18-19, and at ages older than 19 or never initiated. If the reference group is "initiated at ages 11-15" and the estimated parameter for the group who initiated at ages 16-17 is 0.003 in the OLS regression of percent time employed, this means that those who initiated at ages 16-17 had on average percent time employed 0.003 higher than those who initiated at ages 11-15.

Logistic Regression

Logistic regression is used for modeling outcomes that are binary (1/0) variables. A linear probability model has a number of shortcomings in estimating binary dependent variables (Judge et al 1985, Cox and Snell, 1989). Adult outcomes that are binary in our report are alcohol abuse or dependence, past-month drug use, ever being in jail by the age of 33, ever being under poverty at ages 25-29, and ever being on welfare at ages 21-33.

Coefficient estimates from logistic regression do not allow an easy interpretation. Instead, odds ratios, an alternative and preferred measure, are used to present estimated results. Odds ratios measure the relative probability of the estimated outcome among one group relative to the reference group. For example, in estimating the probability of alcohol abuse and dependence, we obtained an odds ratio of 2.25 for males relative to females. Then the interpretation is that male respondents in the sample were 2.25 times as likely to develop alcohol abuse and dependence as female respondents in the sample. If an odds ratio for a group is greater than 1, then this group is more likely to end up with the outcome compared with the reference group (whose odds ratio is 1); if an odds ratio is less than 1, then this group is less likely to end up with the outcome compared with the reference group.

Multinomial Logistic Regression

Estimation of unordered-choice dependent variables requires a multinomial logistic model (Greene, 2000). It is intended for use when the dependent variable takes on more than two outcomes and the outcomes have no natural ordering. In our study, the family formation variable, "fertility and marital status at the age of 33", takes on six outcomes without natural ordering. Similar to logistic regression, the multinomial logistic regression provides a measure of the probability of one outcome relative to the reference outcome, known as relative risk. However, it is more difficult to interpret the relative risk from multinomial logistic regression since there are multiple equations. As an alternative, prediction is used to aid interpretation. We use the "method of recycled predictions", in which we vary characteristics of interest across the whole data set and average the prediction (STATA Reference, version 7). For example, in our data set we have those who initiated sex at ages 11-15, 16-17, 18-19, and those who had not initiated sex by age 19. We first assume that all respondents initiated sex at ages 11-15 but hold their other characteristics constant. We then calculate the probabilities for each fertility and marriage outcome. We repeat this exercise for the other three initiation groups. The difference between any two sets of calculated probabilities, then, is the difference due to different ages of sex initiation, holding other characteristics constant. For example, the predicted probabilities of "never married with children" for the four age categories of sex initiation are 0.112, 0.087, 0.075, and 0.042, respectively.

Negative Binomial Regression

Negative binomial regression is used to model count dependent variables. A count variable, for example, the number of years in poverty, is assumed to follow a Poisson distribution. The Poisson distribution has the feature that its mean equals its variance. Since the variance of a count variable is often empirically larger than its mean, a situation known as over-dispersion (Hausman, Hall and Griliches, 1984), in a negative binomial regression, the Poisson parameter is assumed to follow a Gamma distribution. In our study, two outcome variables are of the nature of count data: years in poverty between the ages of 25-29, and years on welfare between the ages of 21-33.

An estimated coefficient from a negative binomial regression is interpreted as percentage changes in the outcome variable given a unit change in an explanatory variable. To keep the interpretation comparable, we present results from negative binomial regressions in the form of marginal effects. Marginal effects measure changes in an outcome variable resulting from a unit change in an independent variable. For example, in estimating the number of years in poverty, we obtained a marginal effect of -0.231 for males relative to females (whose marginal effect is set to 0), which means male respondents on average spent 0.231 years less in poverty compared with female respondents. In essence, the interpretation of marginal effects is equivalent to estimated coefficients in linear models.

Weibull Regression

The Weibull distribution is one of the distributions used in modeling survival or duration data. The variable of interest in the analysis of duration is the length of time that elapses from the beginning of some event either until its end or until the measurement is taken, which may precede termination. In our study, we apply Weibull regression to the age when reached 2 years of tenure with one employer after ending formal schooling. The starting point is the age when formal schooling ended and the ending point is when two years of tenure was reached.

The Weibull distributional function is nonlinear. To keep the interpretation of results comparable, we present results in the form of marginal effects as well. Marginal effects measure changes in an outcome variable resulting from a unit change in an independent variable. For example, in estimating age when reached 2 years of tenure after ending formal schooling, we obtained a marginal effect of -1.003 for males relative to females (whose marginal effect is set to 0). This means it took male respondents on average 1.003 years less than female respondents to reach 2 years of tenure after ending formal schooling.

Statistical Significance of Estimates

The outcome variables in this study were assumed to follow different distributions and therefore a variety of regressions were estimated. To demonstrate statistical significance of estimated results consistently and as simply as possible, we chose to show p-values for all estimates, be it odds ratios or marginal effects. P-values range from 0 to 1 and are the observed "tail" probability of a statistic being at least as extreme as the particular observed value when the null hypothesis is true. Usually the null hypothesis in our study is that the estimated coefficient is 0. For example, if the estimated coefficient of male relative to female from the OLS regression of percent time employed yields 0.02 and the t-statistic is 3.45 with the associated p-value at 0.01, then the probability is as high as 99 percent (1 minus p-value) that the true effect of sex on percent time employed is not 0. The lower the p-value, the higher the statistical significance of the estimated result. The cutoff for statistical significance in this report is a p-value of 0.1.

Endnote

25. The creation of the variable "last time enrolled in school" takes into account the highest degree obtained by each individual respondent. An algorithm was developed to create this variable.

Appendix B: Regressions of Adult Outcomes on Adolescent Risky Behaviors, Family Environment, and Other Independent Variables

Table B1. Odds Ratios from Logistic Regressions of Adult Health and Crime Outcomes
Independent Variables	Adult Outcome
	Health		Crime
	Alcohol Abuse or Dependence Odds Ratio [P-value]	Past-month Drug Use Odds Ratio [P-value]	Ever Being in Jail Odds Ratio [P-value]
Number of observations	5571	5462	5571
Age of alcohol initiation
11-15	1	1	1
16-17	0.88 [0.21]	1.07 [0.58]	1.19 [0.40]
18-19	0.56 [0.00]	1.06 [0.71]	1.54 [0.07]
not by age 19	0.23 [0.00]	0.62 [0.03]	1.16 [0.62]
Age of marijuana initiation
11-15	1	1	1
16-17	1.07 [0.37]	0.76 [0.03]	0.77 [0.20]
18-19	1.30 [0.07]	0.88 [0.40]	0.62 [0.09]
not by age 19	0.61 [0.00]	0.28 [0.00]	0.52 [0.00]
Age of cocaine initiation
11-15	1	1	1
16-17	1.27 [0.50]	0.91 [0.79]	0.28 [0.01]
18-19	0.92 [0.82]	1.03 [0.92]	0.29 [0.00]
not by age 19	0.75 [0.37]	0.53 [0.05]	0.12 [0.00]
Age of sex initiation
11-15	1	1	1
16-17	1.06 [0.56]	1.00 [0.98]	0.84 [0.34]
18-19	1.04 [0.78]	0.69 [0.02]	0.49 [0.02]
not by age 19	0.77 [0.03]	0.28 [0.00]	0.42 [0.00]
Number of crimes/delinquencies (1980)
9+	1	1	1
3-8	0.81 [0.07]	0.93 [0.59]	0.57 [0.01]
1-2	0.75 [0.03]	0.81 [0.15]	0.63 [0.03]
0	0.50 [0.00]	0.47 [0.00]	0.28 [0.00]
Sex
Male	2.25 [0.00]	1.25 [0.04]	6.61 [0.00]
Female	1	1	1
Race/Ethnicity
White	1	1	1
Black	0.94 [0.59]	1.08 [0.54]	4.62 [0.00]
Hispanic	0.98 [0.84]	0.69 [0.02]	1.64 [0.04]
Educational attainment
High school dropout	1	1	1
High school graduate	0.85 [0.24]	0.91 [0.57]	0.47 [0.00]
Some college and up	0.69 [0.02]	0.72 [0.06]	0.24 [0.00]
Rosenberg self-esteem scale	0.96 [0.01]	1.00 [0.71]	0.92 [0.00]
Parent with alcohol problem
Yes	1.33 [0.01]	1.21 [0.09]	1.32 [0.00]
No	1	1	1
Family Type
Both biological parents	1	1	1
Single mother	0.71 [0.02]	0.96 [0.80]	0.97 [0.89]
Single father	0.87 [0.72]	0.85 [0.73]	1.48 [0.49]
Mothers and stepfather	0.99 [0.96]	1.23 [0.25]	1.08 [0.80]
Dads and stepmother	0.42 [0.04]	0.60 [0.25]	2.61 [0.03]
Others (relatives and non-relatives)	1.21 [0.52]	0.96 [0.91]	1.55 [0.35]
Mother's Education Attainment
High school dropout	1	1	1
High school graduate	0.97 [0.76]	0.89 [0.33]	0.99 [0.95]
Some college	0.81 [0.22]	1.51 [0.02]	1.54 [0.18]
College graduate	0.69 [0.07]	1.36 [0.15]	0.48 [0.21]
Father's Education Attainment
High school dropout	1	1	1
High school graduate	1.21 [0.08]	1.08 [0.51]	0.54 [0.00]
Some college	0.96 [0.83]	0.94 [0.75]	0.77 [0.44]
College graduate	1.33 [0.08]	0.94 [0.75]	0.44 [0.05]

B2. Estimates from Regressions of Adult Economic Outcomes B2
Independent Variables	Economic Outcomes
	Logistic Regressions Odds Ratio [P-value]		Negative Binomial Regressions Marginal Effects [P-value]		OLS Marginal Effects [P-value]	Weibull Regression Marginal Effects [P-Value]
	Ever Being in Poverty Ages 25-29	Ever Being on Welfare Ages 21-33	Years in Poverty Ages 25-29	Years on Welfare Ages 21-33	Percent time Employed	Age when Reached 2 Years of Tenure
Number of observations	5571	5571	5520	5571	5534	5096
Age of alcohol initiation
11-15	1	1	0	0	0	0
16-17	0.93 [0.48]	0.90 [0.28]	-0.031 [0.33]	-0.103 [0.09]	-0.003 [0.56]	-0.401 [0.12]
18-19	1.15 [0.18]	0.80 [0.05]	0.035 [0.32]	-0.126 [0.06]	-0.003 [0.68]	0.500 [0.07]
not by 19	1.27 [0.05]	1.06 [0.64]	0.051 [0.18]	0.002 [0.98]	-0.007 [0.35]	0.735 [0.02]
Age of marijuana initiation
11-15	1	1	0	0	0	0
16-17	0.92 [0.45]	0.98 [0.86]	-0.012 [0.73]	-0.006 [0.92]	0.003 [0.63]	-0.169 [0.54]
18-19	0.79 [0.07]	0.72 [0.01]	-0.083 [0.05]	-0.148 [0.07]	0.007 [0.35]	-0.596 [0.07]
Not by 19	0.90 [0.27]	0.80 [0.03]	-0.028 [0.39]	-0.091 [0.15]	0.001 [0.88]	-0.273 [0.31]
Age of cocaine initiation
11-15	1	1	0	0	0	0
16-17	0.86 [0.67]	0.72 [0.34]	-0.075 [0.49]	-0.204 [0.35]	0.029 [0.19]	-0.749 [0.46]
18-19	0.67 [0.22]	0.59 [0.11]	-0.140 [0.19]	-0.303 [0.15]	0.052 [0.01]	-1.063 [0.27]
Not by 19	0.59 [0.08]	0.58 [0.08]	-0.174 [0.08]	-0.317 [0.11]	0.048 [0.02]	-0.830 [0.37]
Age of sex initiation
11-15	1	1	0	0	0	0
16-17	0.86 [0.10]	0.89 [0.22]	-0.060 [0.04]	-0.048 [0.40]	0.011 [0.05]	-0.698 [0.01]
18-19	0.71 [0.00]	0.60 [0.00]	-0.109 [0.00]	-0.264 [0.00]	0.017 [0.01]	-1.583 [0.00]
Not by 19	0.67 [0.00]	0.34 [0.00]	-0.168 [0.00]	-0.586 [0.00]	0.024 [0.00]	-2.725 [0.00]
Number of crimes/delinquencies (1980)
9+	1	1	0	0	0	0
3-8	0.81 [0.07]	0.83 [0.10]	-0.082 [0.03]	-0.144 [0.05]	0.016 [0.02]	-0.391 [0.19]
1-2	0.96 [0.69]	0.80 [0.06]	-0.035 [0.35]	-0.133 [0.07]	0.014 [0.04]	-0.285 [0.36]
0	0.85 [0.17]	0.69 [0.00]	-0.075 [0.04]	-0.221 [0.00]	0.015 [0.02]	-0.606 [0.05]
Sex
Male	0.59 [0.00]	0.32 [0.00]	-0.231 [0.00]	-0.712 [0.00]	0.020 [0.00]	-1.003 [0.00]
Female	1	1	0	0	0	0
Race/Ethnicity
White	1	1	0	0	0	0
Black	2.78 [0.00]	2.86 [0.00]	0.399 [0.00]	0.756 [0.00]	-0.084 [0.00]	1.011 [0.00]
Hispanic	1.73 [0.00]	1.72 [0.00]	0.189 [0.00]	0.320 [0.00]	-0.006 [0.33]	0.425 [0.10]
Educational Attainment
High school dropout	1	1	0	0	0	0
High school graduate	0.38 [0.00]	0.46 [0.00]	-0.218 [0.00]	-0.386 [0.00]	0.072 [0.00]	-1.066 [0.00]
Some college and up	0.24 [0.00]	0.25 [0.00]	-0.385 [0.00]	-0.704 [0.00]	0.108 [0.00]	-0.777 [0.03]
Rosenberg self-esteem scale	0.95 [0.00]	0.96 [0.00]	-0.016 [0.00]	-0.027 [0.00]	0.003 [0.00]	-0.044 [0.05]
Parent with alcohol problem
Yes	1.22 [0.02]	1.41 [0.00]	0.040 [0.13]	0.234 [0.04]	0.002 [0.66]	0.805 [0.00]
No	1	1	0	0	0	0
Family Type
Both biological parents	1	1	0	0	0	0
Single mother	1.32 [0.01]	1.27 [0.02]	0.110 [0.00]	0.176 [0.01]	-0.019 [0.00]	0.533 [0.06]
Single father	2.73 [0.00]	2.34 [0.01]	0.275 [0.00]	0.243 [0.22]	-0.021 [0.01]	1.899 [0.04]
Mother and stepfather	1.47 [0.01]	1.49 [0.00]	0.149 [0.00]	0.217 [0.01]	-0.031 [0.00]	0.779 [0.05]
Father and stepmother	1.12 [0.68]	1.94 [0.01]	0.100 [0.24]	0.348 [0.03]	-0.014 [0.41]	1.897 [0.01]
Others (relatives and non-relatives)	1.79 [0.01]	1.48 [0.08]	0.154 [0.02]	0.260 [0.07]	-0.018 [0.21]	[0.05]
Mother's Education Attainment
High school dropout	1	1	0	0	0	0
High school graduate	0.71 [0.00]	0.88 [0.14]	-0.124 [0.00]	-0.147 [0.01]	0.019 [0.00]	-1.085 [0.00]
Some college	0.77 [0.06]	0.90 [0.49]	-0.078 [0.09]	-0.115 [0.19]	0.021 [0.01]	-0.566 [0.11]
College graduate	0.55 [0.00]	0.58 [0.01]	-0.270 [0.00]	-0.611 [0.00]	0.021 [0.02]	-0.663 [0.10]
Father's Education Attainment
High school dropout	1	1	0	0	0	0
High school graduate	0.85 [0.06]	0.72 [0.00]	-0.079 [0.00]	-0.219 [0.00]	0.015 [0.00]	-0.447 [0.05]
Some college	0.86 [0.29]	0.81 [0.12]	-0.106 [0.02]	-0.223 [0.01]	0.010 [0.18]	-0.088 [0.80]
College graduate	1.03 [0.81]	0.50 [0.00]	-0.062 [0.18]	-0.452 [0.00]	0.016 [0.03]	-0.504 [0.14]

B3. Relative Risk Ratios from Multinomial Logistic Regression of Adult Marital and Fertility Status at Age 33
Independent Variables	Family Formation Outcomes
	Marital and Fertility and Status at age 33 Multi-logistic Regressions Relative Risk Ratios [P-value] Reference Group: Married with Children
	Never married without children	Never married with children	Married without children	Married but divorced without children	Married but divorced with children
Number of observations=5431
Age of alcohol initiation
11-15	1	1	1	1	1
16-17	0.83 [0.13]	1.39 [0.05]	1.10 [0.48]	0.88 [0.47]	0.89 [0.31]
18-19	0.96 [0.76]	1.37 [0.09]	1.14 [0.39]	0.94 [0.77]	0.95 [0.72]
not by 19	0.86 [0.32]	1.12 [0.57]	1.03 [0.87]	0.68 [0.14]	0.88 [0.38]
Age of marijuana initiation
11-15	1	1	1	1	1
16-17	1.17 [0.25]	0.98 [0.92]	0.91 [0.53]	0.94 [0.75]	0.99 [0.97]
18-19	1.00 [0.99]	0.67 [0.07]	0.97 [0.86]	1.38 [0.15]	0.88 [0.40]
not by 19	0.96 [0.73]	0.90 [0.52]	0.90 [0.44]	0.92 [0.69]	0.92 [0.49]
Age of cocaine initiation
11-15	1	1	1	1	1
16-17	1.10 [0.83]	0.48 [0.21]	2.71 [0.14]	0.90 [0.86]	1.97 [0.14]
18-19	0.81 [0.63]	0.70 [0.50]	1.94 [0.31]	0.88 [0.84]	1.13 [0.79]
not by 19	0.41 [0.03]	0.48 [0.13]	1.26 [0.72]	0.53 [0.27]	1.08 [0.86]
Age of sex initiation
11-15	1	1	1	1	1
16-17	1.20 [0.17]	0.68 [0.01]	1.06 [0.69]	1.13 [0.53]	0.72 [0.00]
18-19	1.65 [0.00]	0.56 [0.00]	1.22 [0.23]	0.86 [0.51]	0.48 [0.00]
not by 19	3.23 [0.00]	0.33 [0.00]	1.76 [0.00]	0.73 [0.22]	0.23 [0.00]
Number of crimes/delinquencies (1980)
9+	1	1	1	1	1
3-8	0.76 [0.05]	0.62 [0.01]	0.96 [0.79]	1.16 [0.50]	0.99 [0.92]
1-2	0.83 [0.19]	0.54 [0.00]	1.08 [0.64]	0.92 [0.74]	0.94 [0.66]
0	0.67 [0.00]	0.45 [0.00]	0.96 [0.82]	0.78 [0.30]	0.83 [0.20]
Sex
Male	2.06 [0.00]	0.74 [0.02]	1.46 [0.00]	0.75 [0.05]	0.58 [0.00]
Female	1	1	1	1	1
Race/Ethnicity
White	1	1	1	1	1
Black	1.91 [0.00]	11.73 [0.00]	0.88 [0.34]	0.67 [0.06]	0.94 [0.56]
Hispanic	0.87 [0.27]	2.36 [0.00]	0.76 [0.05]	0.77 [0.20]	0.98 [0.85]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.17 [0.38]	0.63 [0.01]	1.11 [0.64]	1.21 [0.54]	1.13 [0.38]
Some college and up	1.64 [0.01]	0.52 [0.00]	2.12 [0.00]	1.72 [0.08]	0.86 [0.33]
Rosenberg self-esteem scale	0.96 [0.00]	0.96 [0.00]	1.00 [0.57]	1.00 [0.88]	1.00 [0.74]
Parent with alcohol problem
Yes	0.96 [0.75]	1.19 [0.19]	0.85 [0.22]	1.09 [0.64]	1.17 [0.11]
No	1	1	1	1	1
Family Type
Both biological parents	1	1	1	1	1
Single mother	1.50 [0.00]	1.48 [0.01]	1.04 [0.83]	1.00 [0.99]	1.02 [0.88]
Single father	1.58 [0.26]	2.00 [0.14]	0.69 [0.56]	0.51 [0.51]	1.01 [0.99]
Mother and stepfather	0.91 [0.67]	1.29 [0.26]	0.97 [0.91]	1.23 [0.48]	1.61 [0.00]
Father and stepmother	0.45 [0.08]	0.88 [0.76]	1.10 [0.79]	0.70 [0.56]	0.84 [0.60]
Others (relatives and non-relatives)	0.98 [0.96]	0.87 [0.71]	0.97 [0.94]	1.53 [0.38]	1.68 [0.05]
Mother's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	1.04 [0.74]	0.77 [0.06]	1.07 [0.56]	1.25 [0.22]	0.96 [0.67]
Some college	1.05 [0.76]	0.99 [0.95]	1.06 [0.76]	1.15 [0.60]	0.77 [0.14]
College graduate	1.26 [0.21]	0.55 [0.09]	1.44 [0.06]	1.12 [0.70]	0.91 [0.63]
Father's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	1.24 [0.06]	1.14 [0.34]	1.12 [0.36]	1.30 [0.14]	1.17 [0.11]
Some college	1.19 [0.30]	1.25 [0.34]	1.08 [0.66]	1.51 [0.09]	0.95 [0.74]
College graduate	1.21 [0.22]	0.80 [0.43]	1.13 [0.46]	1.02 [0.93]	0.85 [0.33]

Probabilities of Marital and Fertility Outcomes at age 33 by Age of Alcohol Initiation
Marriage and Fertility Outcome At age 33	Age of Alcohol Initiation
Marriage and Fertility Outcome At age 33	11-15	16-17	18-19	not by age 19
Never married without children	0.164	0.139	0.154	0.149
Never married with children	0.071	0.096	0.092	0.083
Married without children	0.104	0.117	0.116	0.112
Married with children	0.437	0.445	0.428	0.459
Married but divorced without children	0.051	0.046	0.048	0.037
Married but divorced with children	0.173	0.156	0.162	0.160

Probabilities of Marital and Fertility Outcomes at age 33 by Age of Marijuana Initiation
Marriage and Fertility Outcome At age 33	Age of Marijuana Initiation
Marriage and Fertility Outcome At age 33	11-15	16-17	18-19	not by age 19
Never married without children	0.145	0.167	0.150	0.145
Never married with children	0.093	0.091	0.069	0.089
Married without children	0.120	0.108	0.119	0.112
Married with children	0.431	0.428	0.446	0.450
Married but divorced without children	0.045	0.042	0.064	0.044
Married but divorced with children	0.166	0.164	0.152	0.160

Probabilities of Marital and Fertility Outcomes at age 33 by Age of Cocaine Initiation
Marriage and Fertility Outcome At age 33	Age of Cocaine Initiation
Marriage and Fertility Outcome At age 33	11-15	16-17	18-19	not by age 19
Never married without children	0.256	0.238	0.212	0.141
Never married with children	0.130	0.062	0.100	0.088
Married without children	0.068	0.151	0.131	0.111
Married with children	0.358	0.300	0.360	0.454
Married but divorced without children	0.067	0.049	0.059	0.045
Married but divorced with children	0.121	0.200	0.138	0.162

Probabilities of Marital and Fertility Outcomes at age 33 by Age of Sex Initiation
Marriage and Fertility Outcome At age 33	Age of Sex Initiation
Marriage and Fertility Outcome At age 33	11-15	16-17	18-19	not by age 19
Never married without children	0.093	0.119	0.166	0.284
Never married with children	0.112	0.087	0.075	0.042
Married without children	0.089	0.100	0.116	0.149
Married with children	0.416	0.448	0.465	0.429
Married but divorced without children	0.046	0.056	0.044	0.034
Married but divorced with children	0.242	0.191	0.134	0.062

Probabilities of Marital and Fertility Outcomes at age 33 by number of Crimes/delinquencies in 1980
Marriage and Fertility Outcome At age 33	Number of Crimes/Delinquencies in 1980
Marriage and Fertility Outcome At age 33	9+	3-8	1-2	0 >
Never married without children	0.173	0.145	0.158	0.139
Never married with children	0.126	0.093	0.082	0.075
Married without children	0.103	0.107	0.120	0.117
Married with children	0.396	0.430	0.432	0.469
Married but divorced without children	0.045	0.056	0.045	0.042
Married but divorced with children	0.158	0.169	0.163	0.158

Probabilities of Marital and Fertility Outcomes at age 33 by family structure
Marriage and Fertility Outcome At age 33	Family Type
Marriage and Fertility Outcome At age 33	Both Biological parents	Single Mother	Single Father	Mother and Stepfather	Father and Stepmother	Others (relatives and non-relatives)
Never married without children	0.147	0.194	0.209	0.124	0.077	0.133
Never married with children	0.083	0.106	0.134	0.093	0.084	0.066
Married without children	0.114	0.107	0.074	0.103	0.144	0.101
Married with children	0.450	0.405	0.414	0.406	0.508	0.404
Married but divorced without children	0.047	0.042	0.022	0.052	0.037	0.063
Married but divorced with children	0.158	0.146	0.146	0.223	0.150	0.232

Probabilities of Marital and Fertility and Outcomes at age 33 by Mother's Educational Attainment
Marriage and Fertility Outcome At age 33	Mother's Educational Attainment
Marriage and Fertility Outcome At age 33	High School Dropout	High School Graduates	Some College	College Graduate
Never married without children	0.144	0.150	0.153	0.173
Never married with children	0.097	0.079	0.099	0.058
Married without children	0.105	0.112	0.112	0.142
Married with children	0.444	0.445	0.453	0.430
Married but divorced without children	0.041	0.050	0.048	0.044
Married but divorced with children	0.169	0.164	0.135	0.153

Probabilities of Marital and Fertility Outcomes at age 33 by Father's Educational Attainment
Marriage and Fertility Outcome At age 33	Father's Educational Attainment
Marriage and Fertility Outcome At age 33	High School Dropout	High School Graduates	Some College	College Graduate
Never married without children	0.138	0.155	0.152	0.165
Never married with children	0.100	0.102	0.114	0.084
Married without children	0.106	0.108	0.107	0.117
Married with children	0.452	0.413	0.425	0.452
Married but divorced without children	0.040	0.047	0.057	0.041
Married but divorced with children	0.163	0.174	0.146	0.141

Appendix C: New Categories of Risky Behavior Variables for Regressions of Different Adult Outcomes Among Early Initiators

Table C1: Definitions of Early Initiators for each Outcome/Behavior Pairing
Adult Outcomes	Age of alcohol initiation	Age of marijuana initiation	Age of cocaine initiation	Age of sex initiation	Number of crimes/delinquencies
Alcohol abuse or dependence ~ 30	11-17	11-19	11-15	11-19	9 +
Past-month drug use ~ 30	11-19	11-19	11-19	11-17	9 +
Ever in Jail by 33	11-15	11-15	11-15	11-17	1 +
Ever in poverty ages 25-29	No Differences by Age	No Differences by Age	11-15	11-17	9 +
Years in poverty ages 25-29	No Differences by Age	11-17	11-17	11-17	9 +
Ever on welfare Ages 21- 33	No Differences by Age	11-17	11-15	11-17	9 +
Years on welfare Ages 21-33	No Differences by Age	11-17	11-17	11-17	9 +
Percent time employed from leaving school to age 33	11-19	11-19	11-17	11-15	9 +
Age when reached 2 years of tenure since leaving school	11-17	11-17	11-15	11-15	1 +
Marital and Fertility status at age 33	No differences by Age	No differences by Age	No differences by Age	11-17	9 +

Appendix D: Regressions of Adult Outcomes on Adolescent Risky Behaviors and Other Independent Variables Among Early Initiators: Effects of Family Structure

Note: Results from regressions with no differences by age are not shown.

Summary Table
Family Type by Risky Behavior	Health Outcomes		Crime Outcome	Economic Outcomes
Family Type by Risky Behavior	Alcohol abuse or dependence Odds Ratios [P-value]	Past-month drug use Odds Ratios [P-value]	In jail by age 33 Odds Ratios [P-value]	Ever in poverty ages 25-29 Odds Ratios [P-value]	Years in poverty ages 25-29 Marginal Effects [P-value]	Ever on welfare ages 21-33 Odds Ratios [P-value]	Years on welfare ages 21-33 Marginal Effects [P-value]	Percent Time Employed Marginal Effects [P-value]	Age achieved job lasting 2 years Marginal Effects [P-value]
Early Initiators of Alcohol
Both biological parents	1	1	N/A	N/A	N/A	1	N/A	N/A	N/A
Single mother	0.75 [0.07]	0.91 [0.55]				0.92 [0.70]
Single father	0.94 [0.88]	0.85 [0.73]				2.31 [0.10]
Mother and stepfather	0.99 [0.97]	1.29 [0.17]				1.88 [0.03]
Father and stepmother	0.54 [0.15]	0.66 [0.36]				3.17 [0.04]
Others (relatives and non-relatives)	1.08 [0.82]	1.33 [0.38]				2.10 [0.09]
Early Initiators of Marijuana
Both biological parents	1	1	1	N/A	0	1	0	0	0
Single mother	0.79 [0.12]	0.63 [0.04]	0.95 [0.88]		0.108 [0.03]	1.11 [0.56]	0.165 [0.12]	-0.011 [0.15]	0.134 [0.75]
Single father	0.91 [0.83]	0.81 [0.20]	1.26 [0.78]		0.174 [0.22]	1.81 [0.29]	0.097 [0.75]	-0.038 [0.11]	1.114 [0.34]
Mother and stepfather	0.84 [0.38]	0.76 [0.59]	1.10 [0.82]		0.172 [0.22]	2.00 [0.00]	0.297 [0.03]	-0.022 [0.04]	0.867 [0.11]
Father and stepmother	0.54 [0.14]	1.12 [0.56]	1.05 [0.96]		0.034 [0.79]	1.54 [0.36]	0.498 [0.05]	-0.026 [0.16]	1.811 [0.08]
Others (relatives and non-relatives)	1.33 [0.38]	1.15 [0.70]	0.96 [0.96]		0.246 [0.03]	1.78 [0.14]	0.582 [0.01]	-0.035 [0.05]	1.526 [0.12]
Early Initiators of Cocaine
Both biological parents	N/A	1	1	1	0	1	0	0	0
Single mother		0.57 [0.06]	1.97 [0.19]	2.24 [0.33]	0.344 [0.03]	1.53 [0.24]	0.243 [0.48]	-0.045 [0.07]	-4.080 [0.05]
Single father		0.22 [0.16]	1.47 [0.74]	-	0.107 [0.79]	2.08 [0.43]	-0.153 [0.85]	0.018 [0.77]	-7.759 [0.01]
Mother and stepfather		0.93 [0.84]	3.20 [0.10]	4.16 [0.20]	0.339 [0.12]	2.48 [0.08]	0.550 [0.24]	-0.046 [0.20]	-
Father and stepmother		1.16 [0.81]	3.11 [0.35]	-	0.229 [0.57]	1.17 [0.87]	0.744 [0.38]	-0.064 [0.33]	-9.931 [0.12]
Others (relatives and non-relatives)		1.24 [0.74]	1.87 [0.68]	8.77 [0.25]	0.557 [0.13]	1.16 [0.88]	1.093 [0.18]	-0.198 [0.00]	-0.738 [0.91]
Early Initiators of Sex
Both biological parents	1	1	1	1	0	1	0	0	0
Single mother	0.73 [0.03]	0.94 [0.70]	1.12 [0.60]	1.03 [0.85]	0.123 [0.01]	1.23 [0.07]	0.160 [0.11]	-0.004 [0.77]	0.277 [0.44]
Single father	1.00 [0.99]	0.57 [0.32]	1.58 [0.42]	4.00 [0.01]	0.391 [0.01]	1.71 [0.14]	0.356 [0.01]	0.013 [0.74]	1.847 [0.09]
Mother and stepfather	1.04 [0.83]	1.24 [0.26]	1.27 [0.42]	1.86 [0.01]	0.192 [0.00]	1.62 [0.00]	0.355 [0.01]	-0.035 [0.07]	1.115 [0.02]
Father and stepmother	0.45 [0.05]	0.61 [0.32]	2.85 [0.02]	0.79 [0.59]	0.169 [0.16]	2.36 [0.01]	0.631 [0.02]	0.008 [0.80]	2.163 [0.02]
Others (relatives and non-relatives)	1.34 [0.31]	1.17 [0.66]	1.38 [0.52]	1.81 [0.08]	0.220 [0.02]	1.45 [0.17]	0.450 [0.05]	-0.025 [0.36]	1.300 [0.11]
Committed More Crimes /Delinquencies in 1980
Both biological parents	1	1	1	1	0	1	0	0	0
Single mother	0.66 [0.14]	1.05 [0.76]	0.89 [0.79]	0.91 [0.73]	0.089 [0.35]	0.96 [0.87]	0.102 [0.60]	-0.008 [0.64]	0.514 [0.14]
Single father	0.24 [0.18]	0.73 [0.53]	2.81 [0.20]	2.79 [0.12]	0.383 [0.10]	4.62 [0.04]	0.182 [0.72]	-0.062 [0.16]	2.191 [0.04]
Mother and stepfather	0.42 [0.02]	1.27 [0.22]	1.48 [0.38]	1.79 [0.05]	0.278 [0.01]	1.39 [0.29]	0.235 [0.33]	-0.050 [0.01]	1.100 [0.02]
Father and stepmother	0.27 [0.10]	0.78 [0.59]	1.13 [0.88]	0.66 [0.51]	0.057 [0.79]	3.73 [0.02]	0.876 [0.03]	-0.053 [0.14]	1.360 [0.14]
Others (relatives and non-relatives)	1.28 [0.71]	1.15 [0.70]	1.16 [0.89]	1.30 [0.70]	0.221 [0.40]	1.23 [0.89]	0.612 [0.24]	-0.043 [0.33]	2.098 [0.01]

Probabilities of Marital and Fertility Outcomes from Multi-logistic Regressions Among Early Sex Initiators, by Family Structure
Marriage and Fertility Outcome at 33	Family Type
Marriage and Fertility Outcome at 33	Both biological parents	Single mother	Single father	Mother and stepfather	Father and stepmother	Others (relatives and non-relatives)
Never married without children	0.109	0.163	0.136	0.090	0.033	0.082
Never married with children	0.092	0.118	0.176	0.112	0.103	0.086
Married without children	0.098	0.097	0.132	0.090	0.152	0.081
Married with children	0.442	0.366	0.330	0.390	0.512	0.387
Married but divorced without children	0.057	0.051	0.040	0.025	-	0.081
Married but divorced with children	0.202	0.204	0.186	0.294	0.200	0.283

Probabilities of Marital and Fertility Outcomes from Multi-logistic Regressions Among those who Committed 9 or more Crimes/Delinquencies in 1980, by Family Structure
Marriage and Fertility Outcome at 33	Family Type
Marriage and Fertility Outcome at 33	Both biological parents	Single mother	Single father	Mother and stepfather	Father and stepmother	Others (relatives and non-relatives)
Never married without children	0.149	0.224	0.223	0.133	0.076	0.149
Never married with children	0.090	0.102	0.114	0.110	0.111	0.072
Married without children	0.112	0.110	0.109	0.083	0.179	0.088
Married with children	0.433	0.379	0.377	0.398	0.430	0.407
Married but divorced without children	0.052	0.034	0.030	0.055	0.062	0.059
Married but divorced with children	0.163	0.152	0.147	0.221	0.141	0.225

D1. Odds Ratios from Logistic Regressions of Alcohol Abuse or Dependence Among Early Initiators
Independent Variables	Adult Alcohol Abuse or Dependence
Independent Variables	Early initiators of alcohol at ages 11-17	Early initiators of marijuana at ages 11-19	Early initiators of sex at age s 11-19	Committed 9 + crimes/delinquencies in 1980
Number of observations	3113	3104	4531	767
Age of alcohol initiation
Early Initiators	-	1	1	1
Else	-	0.55 [0.00]	0.50 [0.00]	0.49 [0.00]
Age of marijuana initiation
Early Initiators	1	-	1	1
Else	0.55 [0.00]	-	0.49 [0.00]	0.82 [0.36]
Age of cocaine initiation
Early Initiators	1	1	1	1
Else	0.86 [0.62]	0.82 [0.50]	0.91 [0.76]	0.59 [0.28]
Age of sex initiation
Early Initiators	1	1	-	1
Else	0.75 [0.11]	0.74 [0.11]	-	0.32 [0.02]
Number of crimes/delinquencies (1980)
Committed 9 or more	1	1	1	-
Else	0.65 [0.01]	0.71 [0.00]	0.65 [0.00]	-
Sex
Male	2.28 [0.00]	1.90 [0.00]	2.37 [0.00]	1.73 [0.02]
Female	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1
Black	0.85 [0.21]	0.88 [0.34]	0.88 [0.28]	0.74 [0.24]
Hispanic	0.91 [0.51]	0.95 [0.71]	0.98 [0.90]	1.28 [0.31]
Educational Attainment
High school dropout	1	1	1	1
High school graduate	0.81 [0.19]	0.83 [0.24]	0.83 [0.19]	0.78 [0.36]
Some college and up	0.65 [0.01]	0.63 [0.01]	0.68 [0.01]	0.62 [0.12]
Rosenberg self-esteem scale	0.97 [0.04]	0.97 [0.01]	0.96 [0.00]	0.99 [0.61]
Parent with alcohol problem
Yes	1.39 [0.00]	1.26 [0.04]	1.35 [0.00]	1.47 [0.06]
No	1	1	1	1
Family Type
Both biological parents	1	1	1	1
Single mother	0.75 [0.07]	0.79 [0.12]	0.73 [0.03]	0.66 [0.14]
Single father	0.94 [0.88]	0.91 [0.83]	1.00 [0.99]	0.24 [0.18]
Mother and stepfather	0.99 [0.97]	0.84 [0.38]	1.04 [0.83]	0.42 [0.02]
Father and stepmother	0.54 [0.15]	0.54 [0.14]	0.45 [0.05]	0.27 [0.10]
Others (relatives and non-relatives)	1.08 [0.82]	1.33 [0.38]	1.34 [0.31]	1.28 [0.71]
Mother's Education Attainment
High school dropout	1	1	1	1
High school graduate	0.98 [0.90]	0.96 [0.75]	1.01 [0.90]	1.05 [0.83]
Some college	0.84 [0.39]	0.85 [0.38]	0.85 [0.36]	0.77 [0.49]
College graduate	0.81 [0.36]	0.70 [0.12]	0.70 [0.11]	0.72 [0.43]
Father's Education Attainment
High school dropout	1	1	1	1
High school graduate	1.12 [0.36]	1.22 [0.11]	1.19 [0.11]	0.92 [0.71]
Some college	1.00 [0.98]	1.15 [0.46]	1.03 [0.89]	0.81 [0.57]
College graduate	1.18 [0.38]	1.40 [0.07]	1.34 [0.09]	0.80 [0.07]

D2. Odds Ratios from Logistic Regressions of Past-month Drug Use Among Early Initiators
Independent Variables	Adult Drug Use
Independent Variables	Early initiators of alcohol at ages 11-19	Early initiators of marijuana at ages 11-19	Early initiators of cocaine at ages 11-19	Early initiators of sex at ages 11-17	Committed 1+ crimes/delinquencies in 1980
Number of observations	4548	3148	667	3126	3460
Age of alcohol initiation
Early Initiators	-	1	1	1	1
Else	-	0.63 [0.04]	0.42 [0.18]	0.51 [0.01]	0.58 [0.01]
Age of marijuana initiation
Early Initiators	1	-	1	1	1
Else	0.32 [0.00]	-	0.53 [0.15]	0.35 [0.00]	0.38 [0.00]
Age of cocaine initiation
Early Initiators	1	1	-	1	1
Else	0.48 [0.00]	0.49 [0.00]	-	0.48 [0.00]	0.46 [0.00]
Age of sex initiation
Early Initiators	1	1	1	-	1
Else	0.55 [0.00]	0.60 [0.00]	0.54 [0.03]	-	0.59 [0.00]
Number of crimes/delinquencies (1980)
Committed 9 or more	1	1	1	1	-
Else	0.53 [0.00]	0.60 [0.00]	0.52 [0.02]	0.54 [0.00]	-
Sex
Male	1.31 [0.01]	1.29 [0.03]	1.08 [0.71]	1.26 [0.05]	1.22 [0.08]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	1.05 [0.72]	1.04 [0.76]	0.79 [0.39]	1.07 [0.62]	1.07 [0.61]
Hispanic	0.70 [0.02]	0.77 [0.10]	0.77 [0.34]	0.71 [0.04]	0.70 [0.03]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.89 [0.49]	0.93 [0.70]	1.01 [0.97]	0.91 [0.60]	0.93 [0.68]
Some college and up	0.69 [0.04]	0.77 [0.10]	0.86 [0.67]	0.67 [0.03]	0.74 [0.11]
Rosenberg self-esteem scale	0.99 [0.50]	1.00 [0.73]	0.96 [0.14]	1.00 [0.81]	0.99 [0.47]
Parent with alcohol problem
Yes	1.25 [0.05]	1.28 [0.04]	1.33 [0.16]	1.13 [0.32]	1.14 [0.30]
No	1	1	1	1	1
Family Type
Both biological parents	1	1	1	1	1
Single mother	0.91 [0.55]	0.63 [0.04]	0.57 [0.06]	0.94 [0.70]	1.05 [0.76]
Single father	0.85 [0.73]	0.81 [0.20]	0.22 [0.16]	0.57 [0.32]	0.73 [0.53]
Mother and stepfather	1.29 [0.17]	0.76 [0.59]	0.93 [0.84]	1.24 [0.26]	1.27 [0.22]
Father and stepmother	0.66 [0.36]	1.12 [0.56]	1.16 [0.81]	0.61 [0.32]	0.78 [0.59]
Others (relatives and non-relatives)	1.33 [0.38]	1.15 [0.70]	1.24 [0.74]	1.17 [0.66]	1.15 [0.70]
Mother's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	0.88 [0.30]	0.91 [0.48]	1.02 [0.94]	0.88 [0.33]	0.90 [0.42]
Some college	1.48 [0.04]	1.70 [0.48]	1.81 [0.09]	1.58 [0.02]	1.50 [0.04]
College graduate	1.40 [0.13]	1.32 [0.24]	1.64 [0.20]	1.26 [0.35]	1.36 [0.20]
Father's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	1.12 [0.37]	1.13 [0.34]	1.10 [0.68]	1.10 [0.47]	1.05 [0.73]
Some college	0.91 [0.62]	0.88 [0.54]	1.60 [0.17]	0.90 [0.63]	0.86 [0.46]
College graduate	0.88 [0.52]	0.90 [0.59]	0.52 [0.07]	1.03 [0.90]	0.80 [0.29]

D3. Odds Ratios from Logistic Regressions of Ever Being in Jail Among Early Initiators
Independent Variables	In Jail By Age 33
Independent Variables	Early initiators of marijuana at ages 11-15	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-19	Committed 9+ crimes/delinquencies in 1980
Number of Observations	1260	299	3191	767
Age of alcohol initiation
Early Initiators	1	1	1	1
Else	0.99 [0.97]	0.52 [0.16]	1.07 [0.71]	0.81 [0.49]
Age of marijuana initiation
Early Initiators	-	1	1	1
Else	-	0.53 [0.22]	0.58 [0.00]	0.61 [0.10]
Age of cocaine initiation
Early Initiators	1	-	1	1
Else	0.30 [0.00]	-	0.36 [0.00]	0.44 [0.03]
Age of sex initiation
Early Initiators	1	1	-	1
Else	0.67 [0.38]	2.08 [0.31]	-	0.68 [0.46]
Number of crimes/delinquencies (1980)
Committed 9 or more	1	1	1	-
Else	0.51 [0.01]	0.72 [0.47]	0.55 [0.00]	-
Sex
Male	7.62 [0.00]	6.50 [0.00]	7.77 [0.00]	7.43 [0.00]
Female	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1
Black	7.81 [0.00]	9.62 [0.00]	4.05 [0.00]	5.22 [0.00]
Hispanic	2.26 [0.02]	4.41 [0.01]	1.74 [0.02]	1.16 [0.73]
Educational Attainment
High school dropout	1	1	1	1
High school graduate	0.44 [0.01]	0.92 [0.88]	0.50 [0.00]	0.27 [0.00]
Some college and up	0.28 [0.02]	0.41 [0.20]	0.28 [0.00]	0.14 [0.00]
Rosenberg self-esteem scale	0.91 [0.01]	0.81 [0.00]	0.92 [0.00]	0.94 [0.09]
Parent with alcohol problem
Yes	1.20 [0.51]	0.58 [0.27]	1.25 [0.24]	1.45 [0.24]
No	1	1	1	1
Family Type
Both biological parents	1	1	1	1
Single mother	0.95 [0.88]	1.97 [0.19]	1.12 [0.60]	0.89 [0.79]
Single father	1.26 [0.78]	1.47 [0.74]	1.58 [0.42]	2.81 [0.20]
Mother and stepfather	1.10 [0.82]	3.20 [0.10]	1.27 [0.42]	1.48 [0.38]
Father and stepmother	1.05 [0.96]	3.11 [0.35]	2.85 [0.02]	1.13 [0.88]
Others (relatives and non-relatives)	0.96 [0.96]	1.87 [0.68]	1.38 [0.52]	1.16 [0.89]
Mother's Education Attainment
High school dropout	1	1	1	1
High school graduate	1.02 [0.95]	2.82 [0.04]	0.98 [0.92]	1.08 [0.81]
Some college	1.45 [0.45]	2.11 [0.47]	1.09 [0.82]	1.00 [0.99]
College graduate	0.33 [0.32]	2.09 [0.60]	0.74 [0.57]	0.18 [0.13]
Father's Education Attainment
High school dropout	1	1	1	1
High school graduate	0.61 [0.09]	0.37 [0.05]	0.54 [0.00]	0.56 [0.10]
Some college	1.10 [0.84]	0.22 [0.10]	0.88 [0.70]	0.94 [0.92]
College graduate	0.24 [0.08]	0.34 [0.34]	0.47 [0.08]	1.33 [0.61]

D4. Odds Ratios from Logistic Regressions of Ever Being In Poverty Among Early Initiators
Independent Variables	Ever in Poverty Between Ages 25-29
Independent Variables	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-15	Committed 9+ crimes/delinquencies in 1980 (9+)
Number of Observations	75	1292	767
Age of alcohol initiation
Early Initiators	1	1	1
Else	0.29 [0.21]	1.02 [0.92]	1.12 [0.59]
Age of marijuana initiation
Early Initiators	1	1	1
Else	1.18 [0.88]	0.74 [0.04]	0.90 [0.60]
Age of cocaine initiation
Early Initiators	-	1	1
Else	-	0.79 [0.50]	0.83 [0.72]
Age of sex initiation
Early Initiators	1	-	1
Else	0.49 [0.34]	-	0.80 [0.27]
Number of crimes/delinquencies (1980)
Committed 9 or more	1	1	-
Else	2.03 [0.28]	0.91 [0.57]	-
Sex
Male	1.09 [0.90]	0.43 [0.00]	7.28 [0.12]
Female	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1
Black	0.47 [0.37]	2.76 [0.00]	2.25 [0.00]
Hispanic	0.53 [0.51]	1.64 [0.01]	1.45 [0.13]
Educational Attainment
High school dropout	1	1	1
High school graduate	0.55 [0.52]	0.47 [0.00]	0.45 [0.00]
Some college and up	1.56 [0.68]	0.30 [0.00]	0.38 [0.00]
Rosenberg self-esteem scale	0.81 [0.02]	0.94 [0.00]	0.95 [0.05]
Parent with alcohol problem
Yes	1.20 [0.80]	1.31 [0.09]	1.56 [0.03]
No	1	1	1
Family Type
Both biological parents	1	1	1
Single mother	2.24 [0.33]	1.03 [0.85]	0.91 [0.73]
Single father	-	4.00 [0.01]	2.79 [0.12]
Mother and stepfather	4.16 [0.20]	1.86 [0.01]	1.79 [0.05]
Father and stepmother	-	0.79 [0.59]	0.66 [0.51]
Others (relatives and non-relatives)	8.77 [0.25]	1.81 [0.08]	1.30 [0.70]
Mother's Education Attainment
High school dropout	1	1	1
High school graduate	0.84 [0.81]	0.68 [0.01]	0.67 [0.05]
Some college	1.22 [0.88]	0.86 [0.60]	1.07 [0.85]
College graduate	14.98 [0.15]	0.70 [0.38]	0.46 [0.09]
Father's Education Attainment
High school dropout	1	1	1
High school graduate	0.34 [0.21]	0.87 [0.35]	0.87 [0.51]
Some college	0.61 [0.74]	0.83 [0.50]	0.63 [0.23]
College graduate	0.13 [0.21]	0.60 [0.11]	1.09 [0.81]

D5. Marginal Effects from Negative Binomial Regressions of Years Being in Poverty Among Early Initiators
Independent Variables	Years in Poverty Between Ages 25-29
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of Observations	2413	297	3157	754
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	0.051 [0.20]	0.134 [0.41]	0.026 [0.47]	-0.020 [0.81]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	-0.075 [0.74]	-0.038 [0.28]	-0.078 [0.29]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	-0.164 [0.00]	-	-0.161 [0.01]	-0.187 [0.06]
Age of sex initiation
Early Initiators	0	0	-	0
Else	-0.092 [0.04]	-0.095 [0.59]	-	0.074 [0.41]
Number of crimes/delinquencies (1980)
Committed 9 or more	0	0	0	-
Else	-0.085 [0.06]	-0.208 [0.15]	-0.070 [0.12]	-
Sex
Male	-0.235 [0.00]	-0.262 [0.07]	-0.404 [0.00]	-0.232 [0.01]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	0.531 [0.00]	0.436 [0.08]	0.472 [0.00]	0.520 [0.00]
Hispanic	0.120 [0.05]	0.116 [0.56]	0.149 [0.01]	0.070 [0.51]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.236 [0.00]	-0.383 [0.03]	-0.472 [0.00]	-0.209 [0.03]
Some college and up	-0.336 [0.00]	-0.306 [0.12]	-0.442 [0.00]	-0.331 [0.00]
Rosenberg self-esteem scale	-0.015 [0.00]	-0.023 [0.12]	-0.019 [0.00]	-0.213 [0.01]
Parent with alcohol problem
Yes	0.047 [0.23]	-0.005 [0.97]	0.059 [0.12]	0.034 [0.66]
No	0	0	0	0
Family Type
Both biological parents	0	0	0	0
Single mother	0.108 [0.03]	0.344 [0.03]	0.123 [0.01]	0.089 [0.35]
Single father	0.174 [0.22]	0.107 [0.79]	0.391 [0.01]	0.383 [0.10]
Mother and stepfather	0.172 [0.22]	0.339 [0.12]	0.192 [0.00]	0.278 [0.01]
Father and stepmother	0.034 [0.79]	0.229 [0.57]	0.169 [0.16]	0.057 [0.79]
Others (relatives and non-relatives)	0.246 [0.03]	0.557 [0.13]	0.220 [0.02]	0.221 [0.40]
Mother's Education Attainment
High school dropout	0	0	0	0
High school graduate	-0.111 [0.01]	-0.340 [0.02]	-0.088 [0.03]	-0.161 [0.04]
Some college	-0.213 [0.00]	-0.711 [0.01]	-0.094 [0.18]	-0.054 [0.71]
College graduate	-0.233 [0.01]	-0.203 [0.46]	-0.157 [0.04]	-0.412 [0.02]
Father's Education Attainment
High school dropout	0	0	0	0
High school graduate	-0.044 [0.31]	-0.180 [0.23]	-0.088 [0.03]	-0.111 [0.18]
Some college	-0.108 [0.13]	-0.039 [0.87]	-0.094 [0.18]	-0.250 [0.08]
College graduate	-0.096 [0.20]	-0.128 [0.62]	-0.157 [0.04]	-0.070 [0.61]

D6.Odds Ratios from Logistic Regressions of Ever Being on Welfare Among Early Initiators
Independent Variables	Ever on welfare Between Ages 21-33
Independent Variables	Early initiators of alcohol at ages11-15	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of Observations	1077	1260	299	3191	767
Age of alcohol initiation
Early Initiators	-	1	1	1	1
Else	-	1.02 [0.77]	1.04 [0.89]	0.82 [0.05]	1.05 [0.80]
Age of marijuana initiation
Early Initiators	1	-	1	1	1
Else	1.01 [0.94]	-	0.90 [0.74]	0.95 [0.59]	0.49 [0.00]
Age of cocaine initiation
Early Initiators	1	1	-	1	1
Else	0.62 [0.04]	0.67 [0.03]	-	0.70 [0.02]	0.86 [0.56]
Age of sex initiation
Early Initiators	1	1	1	-	1
Else	0.52 [0.00]	0.83 [0.29]	0.81 [0.62]	-	0.56 [0.03]
Number of crimes/delinquencies (1980)
Committed 9 or more	1	1	1	1	-
Else	0.94 [0.72]	0.55 [0.00]	0.90 [0.76]	0.75 [0.01]	-
Sex
Male	0.50 [0.00]	0.45 [0.00]	0.63 [0.13]	0.29 [0.00]	0.34 [0.00]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	2.65 [0.00]	3.36 [0.00]	3.13 [0.00]	2.62 [0.00]	2.92 [0.00]
Hispanic	1.37 [0.12]	1.59 [0.01]	0.87 [0.72]	1.31 [0.03]	1.81 [0.02]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.58 [0.01]	0.49 [0.00]	0.43 [0.04]	0.48 [0.00]	0.45 [0.00]
Some college and up	0.23 [0.00]	0.37 [0.00]	0.23 [0.00]	0.28 [0.00]	0.29 [0.00]
Rosenberg self-esteem scale	0.94 [0.00]	0.97 [0.14]	0.96 [0.28]	0.94 [0.00]	0.97 [0.26]
Parent with alcohol problem
Yes	1.22 [0.24]	1.64 [0.00]	2.38 [0.00]	1.41 [0.00]	1.43 [0.08]
No	1	1	1	1	1
Family Type
Both biological parents	1	1	1	1	1
Single mother	0.92 [0.70]	1.11 [0.56]	1.53 [0.24]	1.23 [0.07]	0.96 [0.87]
Single father	2.31 [0.10]	1.81 [0.29]	2.08 [0.43]	1.71 [0.14]	4.62 [0.04]
Mother and stepfather	1.88 [0.03]	2.00 [0.00]	2.48 [0.08]	1.62 [0.00]	1.39 [0.29]
Father and stepmother	3.17 [0.04]	1.54 [0.36]	1.17 [0.87]	2.36 [0.01]	3.73 [0.02]
Others (relatives and non-relatives)	2.10 [0.09]	1.78 [0.14]	1.16 [0.88]	1.45 [0.17]	1.23 [0.89]
Mother's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	0.93 [0.69]	0.99 [0.94]	0.64 [0.19]	0.93 [0.44]	0.84 [0.41]
Some college	0.84 [0.59]	0.56 [0.05]	0.42 [0.14]	0.85 [0.38]	0.57 [0.16]
College graduate	1.07 [0.85]	0.92 [0.80]	1.01 [0.99]	0.64 [0.05]	0.45 [0.09]
Father's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	0.83 [0.30]	0.74 [0.06]	0.67 [0.23]	0.73 [0.00]	0.60 [0.02]
Some college	0.74 [0.29]	0.69 [0.17]	0.86 [0.77]	0.83 [0.26]	0.96 [0.90]
College graduate	0.45 [0.01]	0.42 [0.00]	0.70 [0.52]	0.50 [0.00]	0.66 [0.27]

D7.Marginal Effects from Negative Binomial Regressions of Years Being on Welfare Among Early Initiators
Independent Variables	Years on Welfare Between Ages 21-33
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of Observations	2433	299	3192	767
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	-0.108 [0.19]	-0.002 [0.99]	-0.176 [0.05]	-0.132 [0.38]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	0.038 [0.93]	-0.115 [0.13]	0.066 [0.66]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	-0.268 [0.02]	-	-0.232 [0.09]	-0.317 [0.13]
Age of sex initiation
Early Initiators	0	0	-	0
Else	-0.357 [0.00]	-0.130 [0.72]	-	-0.226 [0.22]
Number of crimes/delinquencies (1980)
Committed 9 or more	0	0	0	-
Else	-0.084 [0.37]	-0.237 [0.45]	-0.230 [0.02]	-
Sex
Male	-0.869 [0.00]	-0.852 [0.02]	-1.566 [0.00]	-0.960 [0.00]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	1.054 [0.00]	1.452 [0.06]	0.782 [0.00]	0.782 [0.01]
Hispanic	0.300 [0.02]	0.355 [0.43]	0.193 [0.11]	0.323 [0.20]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.556 [0.00]	-0.593 [0.09]	-0.587 [0.00]	-0.530 [0.01]
Some college and up	-0.792 [0.00]	-1.142 [0.01]	-0.987 [0.00]	-0.840 [0.00]
Rosenberg self-esteem scale	-0.036 [0.00]	-0.070 [0.04]	-0.045 [0.00]	-0.040 [0.03]
Parent with alcohol problem
Yes	0.316 [0.00]	0.520 [0.06]	0.288 [0.00]	0.356 [0.02]
No	0	0	0	0
Family Type
Both biological parents	0	0	0	0
Single mother	0.165 [0.12]	0.243 [0.48]	0.160 [0.11]	0.102 [0.60]
Single father	0.097 [0.75]	-0.153 [0.85]	0.356 [0.01]	0.182 [0.72]
Mother and stepfather	0.297 [0.03]	0.550 [0.24]	0.355 [0.01]	0.235 [0.33]
Father and stepmother	0.498 [0.05]	0.744 [0.38]	0.631 [0.02]	0.876 [0.03]
Others (relatives and non-relatives)	0.582 [0.01]	1.093 [0.18]	0.450 [0.05]	0.612 [0.24]
Mother's Education Attainment
High school dropout	0	0	0	0
High school graduate	-0.207 [0.02]	-0.657 [0.03]	-0.180 [0.04]	-0.169 [0.29]
Some college	-0.468 [0.00]	-1.087 [0.03]	-0.140 [0.37]	-0.154 [0.61]
College graduate	-0.609 [0.00]	-0.255 [0.65]	-0.691 [0.00]	-0.703 [0.05]
Father's Education Attainment
High school dropout	0	0	0	0
High school graduate	-0.144 [0.10]	-0.481 [0.11]	-0.237 [0.01]	-0.338 [0.04]
Some college	-0.284 [0.04]	-0.110 [0.81]	-0.191 [0.20]	-0.619 [0.03]
College graduate	-0.476 [0.00]	-0.811 [0.11]	-0.697 [0.00]	-0.487 [0.08]

D8. Marginal Effects from OLS Regressions of Percent Time Employed Among Early Initiators
Independent Variables	Percent Time Employed
Independent Variables	Early initiators of marijuana at ages 11-19	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-15	Committed 9+ crimes/delinquencies in 1980
Number of Observations	3090	298	1280	765
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	0.008 [0.36]	0.000 [1.00]	0.003 [0.82]	-0.049 [0.02]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	0.020 [0.62]	0.003 [0.78]	0.018 [0.16]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	0.030 [0.00]	-	0.043 [0.01]	0.031 [0.13]
Age of sex initiation
Early Initiators	0	0	-	0
Else	0.015 [0.01]	0.032 [0.11]	-	0.014 [0.25]
Number of crimes/delinquencies (1980)
Committed 9 or more	0	0	0	-
Else	0.019 [0.00]	0.021 [0.31]	0.022 [0.08]	-
Sex
Male	0.010 [0.06]	0.024 [0.23]	0.041 [0.00]	-0.001 [0.93]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	-0.098 [0.00]	-0.085 [0.00]	-0.103 [0.00]	-0.088 [0.00]
Hispanic	-0.003 [0.71]	-0.003 [0.92]	-0.019 [0.22]	-0.004 [0.82]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	0.063 [0.00]	0.080 [0.01]	0.061 [0.00]	0.046 [0.01]
Some college and up	0.089 [0.00]	0.138 [0.00]	0.087 [0.00]	0.071 [0.00]
Rosenberg self-esteem scale	0.003 [0.00]	0.004 [0.07]	0.005 [0.00]	0.002 [0.12]
Parent with alcohol problem
Yes	-0.004 [0.50]	0.038 [0.06]	-0.001 [0.90]	-0.008 [0.52]
No	0	0	0	0
Family Type
Both biological parents	0	0	0	0
Single mother	-0.011 [0.15]	-0.045 [0.07]	-0.004 [0.77]	-0.008 [0.64]
Single father	-0.038 [0.11]	0.018 [0.77]	0.013 [0.74]	-0.062 [0.16]
Mother and stepfather	-0.022 [0.04]	-0.046 [0.20]	-0.035 [0.07]	-0.050 [0.01]
Father and stepmother	-0.026 [0.16]	-0.064 [0.33]	0.008 [0.80]	-0.053 [0.14]
Others (relatives and non-relatives)	-0.035 [0.05]	-0.198 [0.00]	-0.025 [0.36]	-0.043 [0.33]
Mother's Education Attainment
High school dropout	0	0	0	0
High school graduate	0.020 [0.00]	0.034 [0.14]	0.019 [0.12]	0.009 [0.49]
Some college	0.027 [0.00]	0.042 [0.26]	0.019 [0.35]	0.012 [0.57]
College graduate	0.028 [0.01]	0.027 [0.51]	-0.006 [0.82]	0.021 [0.38]
Father's Education Attainment
High school dropout	0	0	0	0
High school graduate	0.005 [0.44]	-0.025 [0.29]	0.001 [0.95]	0.011 [0.40]
Some college	0.013 [0.15]	-0.015 [0.67]	0.008 [0.70]	0.040 [0.06]
College graduate	0.699 [0.00]	-0.017 [0.64]	0.019 [0.38]

D9.Marginal Effects from Weibull Regressions of Age when Reached 2 Years of Tenure Among Early Initiators
Independent Variables	Age Reached two years of Job Tenure
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-17	Committed 1+ crimes/delinquencies in 1980
Number of Observations	2218	61	2858	3217
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	0.700 [0.02]	-1.149 [0.62]	0.932 [0.00]	0.542 [0.02]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	0.254 [0.92]	-0.504 [0.05]	-0.482 [0.04]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	-0.900 [0.28]	-	-1.177 [0.17]	-1.291 [0.15]
Age of sex initiation
Early Initiators	0	0	-	0
Else	-1.471 [0.00]	-6.792 [0.02]	-	-1.201 [0.00]
Number of crimes/delinquencies (1980)
Committed 9 or more	0	0	0	-
Else	-0.477 [0.13]	0.666 [0.73]	-0.158 [0.58]	-
Sex
Male	-0.484 [0.09]	-0.316 [0.88]	-1.033 [0.00]	-0.725 [0.00]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	1.660 [0.00]	-1.169 [0.56]	1.085 [0.00]	1.241 [0.00]
Hispanic	0.354 [0.36]	-0.737 [0.73]	0.379 [0.29]	0.637 [0.05]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.837 [0.09]	1.391 [0.54]	-0.695 [0.09]	-0.924 [0.02]
Some college and up	-0.376 [0.47]	2.325 [0.38]	-0.481 [0.27]	-0.660 [0.13]
Rosenberg self-esteem scale	-0.005 [0.88]	-0.313 [0.19]	-0.046 [0.14]	-0.021 [0.45]
Parent with alcohol problem
Yes	1.365 [0.00]	2.388 [0.23]	0.385 [0.20]	0.954 [0.00]
No	0	0	0	0
Family Type
Both biological parents	0	0	0	0
Single mother	0.134 [0.75]	-4.080 [0.05]	0.277 [0.44]	0.514 [0.14]
Single father	1.114 [0.34]	-7.759 [0.01]	1.847 [0.09]	2.191 [0.04]
Mother and stepfather	0.867 [0.11]	-	1.115 [0.02]	1.100 [0.02]
Father and stepmother	1.811 [0.08]	-9.931 [0.12]	2.163 [0.02]	1.360 [0.14]
Others (relatives and non-relatives)	1.526 [0.12]	-0.738 [0.91]	1.300 [0.11]	2.098 [0.01]
Mother's Education Attainment
High school dropout	0	0	0	0
High school graduate	-1.189 [0.00]	-0.674 [0.68]	-1.097 [0.00]	-1.367 [0.00]
Some college	-1.628 [0.00]	1.108 [0.76]	-0.745 [0.13]	-1.264 [0.00]
College graduate	-1.039 [0.07]	-5.206 [0.17]	-0.969 [0.09]	-1.472 [0.00]
Father's Education Attainment
High school dropout	0	0	0	0
High school graduate	-0.458 [0.18]	1.848 [0.34]	-0.395 [0.18]	-0.248 [0.37]
Some college	-0.164 [0.74]	2.451 [0.50]	0.262 [0.57]	0.149 [0.72]
College graduate	-0.415 [0.40]	1.325 [0.69]	-0.404 [0.40]	-0.136 [0.75]

D10. Relative Risk Ratios from Multi-logistic Regression of Adult Marital and Fertility Status at age 33 Among Early Initiators
Independent Variables	Adult Family Formation
	Early initiators of sex at ages 11-17
	Never married without children	Never married with children	Married without children	Married but divorced without children	Married but divorced with children
Number of observations = 3048
Age of alcohol initiation
11-15	1	1	1	1	1
16-17	0.90 [0.51]	1.54 [0.02]	1.51 [0.03]	1.00 [0.99]	0.87 [0.27]
18-19	1.06 [0.76]	1.76 [0.01]	1.57 [0.04]	0.98 [0.94]	0.93 [0.62]
not by 19	0.81 [0.37]	1.26 [0.33]	1.25 [0.41]	0.59 [0.16]	0.83 [0.29]
Age of marijuana initiation
11-15	1	1	1	1	1
16-17	0.83 [0.30]	0.89 [0.52]	0.95 [0.77]	0.70 [0.14]	0.96 [0.75]
18-19	0.92 [0.67]	0.69 [0.12]	0.92 [0.72]	1.12 [0.70]	0.86 [0.40]
not by 19	0.71 [0.07]	0.91 [0.58]	0.92 [0.64]	0.76 [0.26]	0.91 [0.49]
Age of cocaine initiation
11-15	1	1	1	1	1
16-17	0.99 [0.98]	0.52 [0.28]	6.84 [0.07]	0.60 [0.46]	1.77 [0.22]
18-19	0.66 [0.35]	0.82 [0.72]	4.31 [0.17]	0.76 [0.67]	1.08 [0.86]
not by 19	0.38 [0.02]	0.56 [0.25]	3.11 [0.28]	0.51 [0.25]	1.03 [0.95]
Age of sex initiation
11-15	-	-	-	-	-
16-17	-	-	-	-	-
18-19	-	-	-	-	-
not by 19	-	-	-	-	-
Number of crimes/delinquencies (1980)
9+	1	1	1	1	1
3-8	0.83 [0.30]	0.66 [0.03]	1.04 [0.84]	1.29 [0.34]	0.89 [0.44]
1-2	0.92 [0.67]	0.59 [0.01]	1.05 [0.83]	1.00 [0.99]	0.88 [0.41]
0	0.71 [0.07]	0.44 [0.00]	0.88 [0.57]	0.86 [0.61]	1.03 [0.95]
Sex
Male	2.29 [0.00]	0.94 [0.66]	1.45 [0.02]	0.62 [0.01]	0.57 [0.00]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	1.46 [0.01]	10.21 [0.00]	0.65 [0.02]	0.54 [0.01]	0.81 [0.10]
Hispanic	2.09 [0.00]	2.33 [0.00]	2.34 [0.00]	0.80 [0.40]	0.91 [0.50]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.26 [0.31]	0.65 [0.02]	1.03 [0.91]	0.54 [0.01]	1.06 [0.71]
Some college and up	2.09 [0.00]	0.57 [0.01]	2.34 [0.00]	0.80 [0.40]	0.90 [0.54]
Rosenberg self-esteem scale	0.98 [0.22]	0.97 [0.09]	1.00 [0.96]	1.02 [0.34]	1.02 [0.15]
Parent with alcohol problem
Yes	0.90 [0.50]	1.21 [0.22]	0.72 [0.08]	0.68 [0.10]	1.11 [0.38]
No	1	1	1	1	1
Family Type
Both biological parents	1	1	1	1	1
Single mother	1.87 [0.00]	1.64 [0.00]	1.20 [0.41]	1.09 [0.77]	1.21 [0.22]
Single father	1.73 [0.31]	2.96 [0.05]	1.80 [0.34]	0.93 [0.95]	1.22 [0.71]
Mother and stepfather	0.92 [0.76]	1.44 [0.14]	1.01 [0.98]	0.49 [0.14]	1.67 [0.01]
Father and stepmother	0.26 [0.07]	0.98 [0.97]	1.33 [0.52]	-	0.85 [0.66]
Others (relatives and non-relatives)	0.85 [0.72]	1.05 [0.90]	0.92 [0.89]	1.65 [0.37]	1.62 [0.13]
Mother's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	1.17 [0.29]	0.80 [0.15]	1.15 [0.45]	1.25 [0.31]	0.93 [0.52]
Some college	1.17 [0.52]	1.02 [0.94]	1.31 [0.31]	1.35 [0.37]	0.72 [0.15]
College graduate	1.05 [0.87]	0.51 [0.10]	1.76 [0.08]	0.78 [0.58]	0.84 [0.50]
Father's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	1.02 [0.88]	1.17 [0.31]	1.09 [0.63]	1.23 [0.36]	1.11 [0.39]
Some college	0.94 [0.79]	0.96 [0.89]	1.15 [0.59]	1.56 [0.14]	0.95 [0.81]
College graduate	1.24 [0.36]	0.92 [0.79]	1.21 [0.45]	0.90 [0.78]	0.93 [0.75]

Independent Variables	Adult Family Formation
	Committed 9+ crimes/delinquencies in 1980
	Never married without children	Never married with children	Married without children	Married but divorced without children	Married but divorced with children
Number of observations = 3384
Age of alcohol initiation
11-15	1	1	1	1	1
16-17	0.86 [0.27]	1.43 [0.06]	1.21 [0.26]	0.80 [0.29]	0.84 [0.20]
18-19	0.95 [0.73]	1.58 [0.04]	1.19 [0.36]	0.98 [0.92]	0.89 [0.48]
not by 19	0.88 [0.61]	1.16 [0.55]	1.13 [0.59]	0.63 [0.18]	0.86 [0.42]
Age of marijuana initiation
11-15	1	1	1	1	1
16-17	1.11 [0.50]	0.90 [0.60]	0.90 [0.56]	0.96 [0.85]	1.03 [0.83]
18-19	1.07 [0.70]	0.54 [0.02]	0.88 [0.55]	1.19 [0.54]	0.95 [0.78]
not by 19	0.96 [0.80]	0.79 [0.19]	0.80 [0.20]	0.90 [0.67]	0.96 [0.80]
Age of cocaine initiation
11-15	1	1	1	1	1
16-17	1.74 [0.28]	0.58 [0.38]	3.64 [0.11]	1.02 [0.97]	2.62 [0.06]
18-19	1.11 [0.84]	0.73 [0.58]	2.43 [0.26]	0.84 [0.78]	1.16 [0.76]
not by 19	0.63 [0.33]	0.49 [0.17]	1.75 [0.47]	0.49 [0.23]	1.30 [0.60]
Age of sex initiation
11-15	1	1	1	1	1
16-17	1.08 [0.60]	0.68 [0.02]	1.04 [0.81]	1.03 [0.91]	0.75 [0.03]
18-19	1.48 [0.02]	0.48 [0.00]	1.09 [0.67]	0.76 [0.31]	0.45 [0.00]
not by 19	2.48 [0.00]	0.31 [0.00]	1.56 [0.04]	0.79 [0.46]	0.24 [0.00]
Number of crimes/delinquencies (1980)
9+	-	-	-	-	-
3-8	-	-	-	-	-
1-2	-	-	-	-	-
0	-	-]	-	-	-
Sex
Male	1.69 [0.00]	0.80 [0.13]	1.36 [0.02]	0.62 [0.01]	0.53 [0.00]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	1.69 [0.00]	9.76 [0.00]	1.17 [0.58]	0.68 [0.12]	0.78 [0.08]
Hispanic	0.88 [0.41]	2.14 [0.00]	2.40 [0.00]	0.82 [0.42]	0.96 [0.77]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.04 [0.92]	0.64 [0.02]	1.17 [0.58]	1.06 [0.86]	0.99 [0.94]
Some college and up	1.45 [0.09]	0.47 [0.00]	2.40 [0.00]	1.60 [0.18]	0.77 [0.17]
Rosenberg self-esteem scale	0.95 [0.00]	0.94 [0.00]	1.00 [0.88]	0.99 [0.54]	1.01 [0.58]
Parent with alcohol problem
Yes	0.99 [0.95]	1.15 [0.38]	0.67 [0.02]	0.91 [0.65]	1.17 [0.19]
No	1	1	1	1	1
Family Type
Both biological parents	1	1	1	1	1
Single mother	1.76 [0.00]	1.33 [0.12]	1.15 [0.52]	0.74 [0.36]	1.05 [0.79]
Single father	1.76 [0.23]	1.53 [0.47]	1.13 [0.85]	0.66 [0.69]	1.02 [0.96]
Mother and stepfather	0.95 [0.82]	1.42 [0.16]	0.78 [0.43]	1.15 [0.68]	1.53 [0.03]
Father and stepmother	0.51 [0.23]	1.26 [0.63]	1.63 [0.26]	1.21 [0.76]	0.87 [0.74]
Others (relatives and non-relatives)	1.04 [0.92]	0.83 [0.67]	0.82 [0.73]	1.22 [0.76]	1.50 [0.24]
Mother's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	1.07 [0.63]	0.69 [0.02]	0.94 [0.72]	1.19 [0.42]	0.87 [0.28]
Some college	1.31 [0.18]	1.03 [0.91]	1.25 [0.34]	1.40 [0.29]	0.75 [0.20]
College graduate	1.25 [0.34]	0.61 [0.22]	1.42 [0.16]	1.16 [0.69]	0.91 [0.70]
Father's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	1.21 [0.16]	1.16 [0.37]	1.20 [0.28]	1.06 [0.78]	1.18 [0.20]
Some college	1.03 [0.88]	0.93 [0.81]	1.21 [0.40]	1.17 [0.59]	0.84 [0.39]
College graduate	1.11 [0.59]	0.72 [0.32]	1.27 [0.27]	0.73 [0.32]	0.77 [0.21]

Appendix E: Regressions of Adult Outcomes on Adolescent Risky Behaviors and Other Independent Variables Among Early Initiators Living With Both Biological Parents at Age 14: Effects of Parents' Education

Summary Table

Note: Results from regressions with no differences by age are not shown.

Parents' Education by Risky Behavior	Health Outcomes		Crime	Economic Outcomes
Parents' Education by Risky Behavior	Alcohol abuse or dependence Odds Ratios [P-value]	Past-month drug use Odds Ratios [P-value]	In jail by age 33 Odds Ratios [P-value]	Ever in poverty ages 25-29 Odds Ratios [P-value]	Years in poverty ages 25-29 Marginal Effects [P-value]	Ever on welfare ages 21-33 Odds Ratios [P-value]	Years on welfare ages 21-33 Marginal Effects [P-value]	Percent Time Employed Marginal Effects [P-value]	Age achieved job lasting 2 years Marginal Effects [P-value]
Early Initiators of Alcohol			N/A	N/A	N/A		N/A	N/A	N/A
Mother's Educational Attainment
High school dropout	1	1				1
High school graduate	1.05 [0.74]	0.88 [0.40]				0.76 [0.19]
Some college	0.89 [0.60]	1.28 [0.27]				0.71 [0.39]
College graduate	0.82 [0.44]	1.29 [0.32]				0.69 [0.41]
Father's Educational Attainment
High school dropouts	1	1				1
High school graduate	1.11 [0.48]	1.07 [0.67]				0.87 [0.52]
Some college	1.04 [0.85]	0.83 [0.41]				0.87 [0.66]
College graduate	1.22 [0.34]	0.88 [0.55]				0.48 [0.07]
Early Initiators of Marijuana				N/A
Mother's Educational Attainment
High school dropout	1	1	1		0	1	0	0	0
High school graduate	0.98 [0.91]	0.90 [0.48]	1.01 [0.98]		-0.114 [0.01]	0.99 [0.97]	-0.199 [0.02]	0.023 [0.00]	-1.032 [0.01]
Some college	0.91 [0.65]	1.50 [0.50]	1.83 [0.34]		-0.182 [0.02]	0.61 [0.19]	-0.420 [0.00]	0.035 [0.00]	-1.338 [0.02]
College graduate	0.68 [0.14]	1.10 [0.54]	-		-0.187 [0.04]	0.82 [0.65]	-0.548 [0.00]	0.029 [0.01]	-0.729 [0.27]
Father's Educational Attainment
High school dropout	1	1	1		0	1	0	0	0
High school graduate	1.19 [0.22]	1.10 [0.54]	0.69 [0.33]		-0.066 [0.15]	0.57 [0.01]	-0.152 [0.08]	-0.0004 [0.96]	-0.403 [0.31]
Some college	1.17 [0.43]	0.86 [0.50]	1.37 [0.61]		-0.083 [0.22]	0.56 [0.07]	-0.208 [0.11]	-0.0008 [0.93]	0.029 [0.96]
College graduate	1.42 [0.08]	0.95 [0.82]	0.48 [0.39]		-0.078 [0.29]	0.30 [0.00]	-0.346 [0.01]	0.012 [0.19]	-0.366 [0.51]
Early Initiators of Cocaine	N/A
Mother's Educational Attainment
High school dropout		1	1	1	0	1	0	0	0
High school graduate		0.86 [0.60]	1.94 [0.33]	0.89 [0.91]	-0.296 [0.04]	0.50 [0.11]	-0.755 [0.02]	0.040 [0.09]	-2.071 [0.39]
Some college		1.50 [0.33]	1.85 [0.69]	0.73 [0.85]	-0.572 [0.02]	0.73 [0.65]	-0.664 [0.15]	0.054 [0.16]	0.444 [0.92]
College graduate		1.59 [0.31]	-	-	-0.324 [0.23]	0.85 [0.84]	-0.205 [0.71]	0.036 [0.39]	-2.869 [0.55]
Father's Educational Attainment
High school dropout		1	1	1	0	1	0	0	0
High school graduate		0.99 [0.97]	0.46 [0.26]	0.36 [0.31]	-0.365 [0.01]	0.52 [0.15]	-0.516 [0.09]	-0.026 [0.28]	3.251 [0.17]
Some college		0.60 [0.22]	0.20 [0.21]	1.08 [0.96]	0.023 [0.91]	1.13 [0.84]	0.070 [0.87]	-0.024 [0.48]	5.263 [0.27]
College graduate		0.48 [0.09]	0.65 [0.76]	-	-0.017 [0.94]	0.54 [0.39]	-0.549 [0.26]	-0.020 [0.58]	-1.169 [0.80]
Early Initiators of Sex
Mother's Educational Attainment
High school dropout	1	1	1	1	0	1	0	0	0
High school graduate	1.07 [0.58]	0.91 [0.55]	0.99 [0.96]	0.59 [0.01]	-0.162 [0.00]	0.87 [0.25]	-0.244 [0.01]	0.027 [0.05]	-1.133 [0.00]
Some college	0.87 [0.51]	1.45 [0.13]	0.63 [0.39]	0.86 [0.68]	-0.121 [0.10]	0.85 [0.48]	-0.109 [0.51]	0.032 [0.19]	-0.678 [0.24]
College graduate	0.68 [0.13]	1.12 [0.71]	0.19 [0.12]	0.68 [0.47]	-0.356 [0.00]	0.64 [0.12]	-0.586 [0.01]	-0.007 [0.83]	-1.249 [0.06]
Father's Educational Attainment
High school dropout	1	1	1	1	0	1	0	0	0
High school graduate	1.18 [0.20]	1.02 [0.93]	0.54 [0.02]	0.77 [0.20]	-0.109 [0.01]	0.68 [0.00]	-0.219 [0.02]	-0.004 [0.78]	-0.550 [0.12]
Some college	1.07 [0.72]	0.94 [0.79]	1.11 [0.79]	0.82 [0.56]	-0.088 [0.19]	0.87 [0.46]	-0.187 [0.21]	-0.004 [0.78]	0.345 [0.51]
College graduate	1.41 [0.07]	1.04 [0.87]	0.63 [0.39]	0.43 [0.05]	-0.132 [0.09]	0.48 [0.00]	-0.672 [0.00]	0.039 [0.11]	-0.404 [0.46]
Committed 9+ Crimes /Delinquencies in 1980
Mother's Educational Attainment
High school dropout	1	1	1	1	0	1	0	0	0
High school graduate	1.11 [0.68]	0.92 [0.59]	1.00 [1.00]	0.51 [0.01]	-0.172 [0.03]	0.83 [0.49]	-0.175 [0.27]	0.012 [0.40]	-1.376 [0.00]
Some college	1.04 [0.93]	1.38 [0.18]	1.19 [0.81]	0.87 [0.75]	-0.095 [0.49]	0.72 [0.47]	-0.126 [0.66]	0.022 [0.33]	-1.067 [0.03]
College graduate	0.90 [0.81]	1.26 [0.41]	-	0.34 [0.05]	-0.411 [0.02]	0.65 [0.43]	-0.554 [0.12]	0.010 [0.70]	-1.318 [0.02]
Father's Educational Attainment
High school dropout	1	1	1	1	0	1	0	0	0
High school graduate	1.02 [0.95]	0.92 [0.60]	0.66 [0.36]	0.90 [0.70]	-0.090 [0.27]	0.61 [0.06]	-0.278 [0.10]	0.002 [0.89]	-0.144 [0.65]
Some college	0.90 [0.80]	0.78 [0.30]	0.99 [0.99]	0.77 [0.57]	-0.167 [0.23]	0.86 [0.72]	-0.424 [0.11]	0.029 [0.18]	0.239 [0.60]
College graduate	1.92 [0.07]	0.70 [0.15]	1.28 [0.70]	1.25 [0.58]	0.012 [0.92]	0.77 [0.53]	-0.291 [0.27]	0.022 [0.28]	-0.250 [0.59]

Probabilities Of Marital and Fertility Outcomes From Multi-Logistic Regressions Among Early Sex Initiators Living With Both Biological Parents At Age 14 by Parental Educational Attainment
Marital and Fertility Outcome at age 33	Mother's Educational Attainment				Father's Educational Attainment
Marital and Fertility Outcome at age 33	High school dropout	High school graduate	Some college	College and up	High school dropout	High school graduate	Some college	College and up
Never married without children	0.106	0.117	0.114	0.109	0.104	0.113	0.103	0.132
Never married with children	0.098	0.090	0.096	0.048	0.110	0.095	0.088	0.084
Married without children	0.089	0.098	0.099	0.137	0.093	0.089	0.102	0.111
Married with children	0.442	0.433	0.483	0.488	0.439	0.453	0.426	0.445
Married but divorced without children	0.049	0.063	0.069	0.040	0.049	0.056	0.078	0.049
Married but divorced with children	0.217	0.199	0.139	0.179	0.206	0.193	0.203	0.179

Probabilities Of Marital and Fertility Outcomes From Multi-Logistic Regressions Among Those Who Committed 9 or More Crimes/Delinquencies In 1980 Living With Both Biological Parents At Age 14 by Parental Educational Attainment
Marital and Fertility Outcome at age 33	Mother's Educational Attainment				Father's Educational Attainment
Marital and Fertility Outcome at age 33	High school dropout	High school graduate	Some college	College and up	High school dropout	High school graduate	Some college	College and up
Never married without children	0.138	0.156	0.166	0.176	0.139	0.165	0.139	0.160
Never married with children	0.104	0.079	0.096	0.060	0.106	0.105	0.096	0.064
Married without children	0.106	0.106	0.116	0.139	0.098	0.100	0.116	0.133
Married with children	0.430	0.445	0.427	0.425	0.430	0.414	0.430	0.478
Married but divorced without children	0.045	0.056	0.059	0.054	0.052	0.051	0.067	0.039
Married but divorced with children	0.177	0.157	0.135	0.146	0.173	0.164	0.152	0.125

E1. Odds Ratios from Logistic Regressions of Alcohol Abuse or Dependence Among Early Initiators Living With Both Biological Parents At Age 14
Independent Variables	Alcohol Abuse or Dependence
Independent Variables	Early initiators of alcohol at ages 11-17	Early initiators of marijuana at ages 11-19	Early initiators of sex at age s 11-19	Committed 9 + crimes/delinquencies in 1980
Number of observations	2379	2370	3435	565
Age of alcohol initiation
Early Initiators	-	1	1	1
Else	-	0.55 [0.00]	0.51 [0.00]	0.59 [0.00]
Age of marijuana initiation
Early Initiators	1	-	1	1
Else	0.55 [0.00]	-	0.49 [0.00]	0.81 [0.38]
Age of cocaine initiation
Early Initiators	1	1	1	1
Else	0.83 [0.61]	0.88 [0.71]	0.98 [0.95]	0.49 [0.21]
Age of sex initiation
Early Initiators	1	1	-	1
Else	0.80 [0.23]	0.77 [0.18]	-	0.32 [0.03]
Number of crimes/delinquencies (1980)
Committed 9 or more	1	1	1	-
Else	0.59 [0.00]	0.62 [0.00]	0.56 [0.00]	-
Sex
Male	2.22 [0.00]	1.91 [0.00]	2.35 [0.00]	2.03 [0.01]
Female	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1
Black	0.81 [0.20]	0.89 [0.48]	0.94 [0.64]	0.86 [0.61]
Hispanic	0.88 [0.44]	0.93 [0.67]	0.98 [0.87]	1.13 [0.67]
Educational Attainment
High school dropout	1	1	1	1
High school graduate	0.89 [0.56]	0.91 [0.65]	0.90 [0.54]	0.75 [0.39]
Some college and up	0.70 [0.08]	0.68 [0.06]	0.98 [0.87]	0.53 [0.08]
Rosenberg self-esteem scale	0.97 [0.02]	0.96 [0.01]	0.96 [0.00]	0.99 [0.64]
Parent with alcohol problem
Yes	1.43 [0.01]	1.33 [0.03]	1.42 [0.00]	1.58 [0.05]
No	1	1	1	1
Mother's Education Attainment
High school dropout	1	1	1	1
High school graduate	1.05 [0.74]	0.98 [0.91]	1.07 [0.58]	1.11 [0.68]
Some college	0.89 [0.60]	0.91 [0.65]	0.87 [0.51]	1.04 [0.93]
College graduate	0.82 [0.44]	0.68 [0.14]	0.68 [0.13]	0.90 [0.81]
Father's Education Attainment
High school dropout	1	1	1	1
High school graduate	1.11 [0.48]	1.19 [0.22]	1.18 [0.20]	1.02 [0.95]
Some college	1.04 [0.85]	1.17 [0.43]	1.07 [0.72]	0.90 [0.80]
College graduate	1.22 [0.34]	1.42 [0.08]	1.41 [0.07]	1.92 [0.07]

E2. Odds Ratios from Logistic Regressions of Past-month Drug Use Among Early Initiators Living With Both Biological Parents At Age 14
Independent Variables	Adult Drug Use
Independent Variables	Early initiators of alcohol at ages 11-19	Early initiators of marijuana at ages 11-19	Early initiators of cocaine at ages 11-19	Early initiators of sex at ages 11-17	Committed 1+ crimes/delinquencies in 1980
Number of observations	3565	2325	479	2264	2651
Age of alcohol initiation
Early Initiators	-	1	1	1	1
Else	-	0.71 [0.20]	0.71 [0.62]	0.57 [0.05]	0.65 [0.10]
Age of marijuana initiation
Early Initiators	1	-	1	1	1
Else	0.27 [0.00]	-	0.54 [0.23]	0.33 [0.00]	0.32 [0.00]
Age of cocaine initiation
Early Initiators	1	1	-	1	1
Else	0.44 [0.00]	0.45 [0.00]	-	0.41 [0.00]	0.41 [0.00]
Age of sex initiation
Early Initiators	1	1	1	-	1
Else	0.54 [0.00]	0.61 [0.00]	0.43 [0.01]	-	0.59 [0.00]
Number of crimes/delinquencies (1980)
Committed 9+	1	1	1	1	-
Else	0.58 [0.00]	0.66 [0.01]	0.59 [0.08]	0.61 [0.00]	-
Sex
Male	1.36 [0.01]	1.34 [0.02]	1.12 [0.64]	1.35 [0.04]	1.27 [0.07]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	1.00 [0.98]	0.93 [0.68]	0.88 [0.70]	1.01 [0.93]	1.04 [0.79]
Hispanic	0.61 [0.01]	0.66 [0.03]	0.47 [0.04]	0.58 [0.01]	0.56 [0.01]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.07 [0.77]	1.21 [0.41]	1.11 [0.80]	1.13 [0.60]	1.20 [0.42]
Some college and up	0.89 [0.60]	0.94 [0.79]	0.92 [0.84]	0.87 [0.57]	0.99 [0.95]
Rosenberg self-esteem scale	0.98 [0.30]	0.98 [0.30]	0.96 [0.15]	0.98 [0.33]	0.98 [0.27]
Parent with alcohol problem
Yes	1.23 [0.14]	1.27 [0.09]	1.26 [0.35]	1.07 [0.70]	1.07 [0.66]
No	1	1	1	1	1
Mother's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	0.88 [0.40]	0.90 [0.48]	0.86 [0.60]	0.91 [0.55]	0.92 [0.59]
Some college	1.28 [0.27]	1.50 [0.50]	1.50 [0.33]	1.45 [0.13]	1.38 [0.18]
College graduate	1.29 [0.32]	1.10 [0.54]	1.59 [0.31]	1.12 [0.71]	1.26 [0.41]
Father's Education Attainment
High school dropout	1	1	1	1	1
High school graduate	1.07 [0.67]	1.10 [0.54]	0.99 [0.97]	1.02 [0.93]	0.92 [0.60]
Some college	0.83 [0.41]	0.86 [0.50]	0.60 [0.22]	0.94 [0.79]	0.78 [0.30]
College graduate	0.88 [0.55]	0.95 [0.82]	0.48 [0.09]	1.04 [0.87]	0.70 [0.15]

E3. Odds Ratios from Logistic Regressions of Ever Being in Jail Among Early Initiators Living With Both Biological Parents At Age 14
Independent Variables	Ever in Jail by Age 33
Independent Variables	Early initiators of marijuana at ages11-15	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-19	Committed 9+ crimes/delinquencies in 1980
Number of observations	804	181	2313	516
Age of alcohol initiation
Early Initiators	1	1	1	1
Else	0.66 [0.21]	0.80 [0.74]	1.00 [1.00]	0.81 [0.58]
Age of marijuana initiation
Early Initiators	-	1	1	1
Else	-	0.61 [0.49]	0.50 [0.00]	0.57 [0.13]
Age of cocaine initiation
Early Initiators	1	-	1	1
Else	0.45 [0.03]	-	0.49 [0.03]	0.57 [0.25]
Age of sex initiation
Early Initiators	1	1	-	1
Else	0.63 [0.42]	4.11 [0.18]	-	0.80 [0.67]
Number of crimes/delinquencies (1980)
Committed 9+	1	1	1	-
Else	0.45 [0.02]	0.66 [0.51]	0.52 [0.00]	-
Sex
Male	5.28 [0.00]	8.17 [0.02]	5.56 [0.00]	6.61 [0.00]
Female	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1
Black	10.08 [0.00]	10.93 [0.00]	4.53 [0.00]	5.91 [0.00]
Hispanic	2.20 [0.08]	3.87 [0.08]	1.58 [0.14]	1.28 [0.64]
Educational attainment
High school dropout	1	1	1	1
High school graduate	0.54 [0.13]	1.10 [0.90]	0.56 [0.02]	0.31 [0.01]
Some college and up	0.23 [0.01]	0.27 [0.18]	0.32 [0.00]	0.17 [0.00]
Rosenberg self-esteem scale	0.91 [0.03]	0.82 [0.02]	0.92 [0.00]	0.94 [0.17]
Parent with alcohol problem
Yes	1.62 [0.17]	0.74 [0.66]	1.40 [0.17]	2.11 [0.06]
No	1	1	1	1
Mother's Education Attainment
High school dropout	1	1	1	1
High school graduate	1.01 [0.98]	1.94 [0.33]	0.99 [0.96]	1.00 [1.00]
Some college	1.83 [0.34]	1.85 [0.69]	0.63 [0.39]	1.19 [0.81]
College graduate	-	-	0.19 [0.12]	-
Father's Education Attainment
High school dropout	1	1	1	1
High school graduate	0.69 [0.33]	0.46 [0.26]	0.54 [0.02]	0.66 [0.36]
Some college	1.37 [0.61]	0.20 [0.21]	1.11 [0.79]	0.99 [0.99]
College graduate	0.48 [0.39]	0.65 [0.76]	0.63 [0.39]	1.28 [0.70]

E4. Odds Ratios from Logistic Regressions of Ever Being in Poverty Among Early Initiators Living With Both Biological Parents At Age 14
Independent Variables	Ever in Poverty between ages 25-29
Independent Variables	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-15	Committed 9+ crimes/delinquencies in 1980
Number of observations	53	849	565
Age of alcohol initiation
Early Initiators	1	1	1
Else	0.35 [0.42]	0.90 [0.59]	1.17 [0.55]
Age of marijuana initiation
Early Initiators	1	1	1
Else	0.99 [1.00]	0.69 [0.04]	0.89 [0.62]
Age of cocaine initiation
Early Initiators	-	1	1
Else	-	1.14 [0.76]	1.09 [0.92]
Age of sex initiation
Early Initiators	1	-	1
Else	1.49 [0.68]	-	0.97 [0.27]
Number of crimes/delinquencies (1980)
Committed 9+	1	1	-
Else	2.26 [0.41]	0.91 [0.65]	-
Sex
Male	1.17 [0.86]	0.44 [0.00]	0.81 [0.40]
Female	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1
Black	1.25 [0.82]	3.06 [0.00]	2.86 [0.00]
Hispanic	1.47 [0.75]	1.81 [0.02]	1.28 [0.42]
Educational Attainment
High school dropout	1	1	1
High school graduate	0.40 [0.42]	0.48 [0.00]	0.40 [0.01]
Some college and up	1.36 [0.84]	0.28 [0.00]	0.31 [0.00]
Rosenberg self-esteem scale	0.87 [0.19]	3.06 [0.00]	0.95 [0.07]
Parent with alcohol problem
Yes	1.50 [0.66]	1.38 [0.12]	1.74 [0.03]
No	1	1	1
Mother's Education Attainment
High school dropout	1	1	1
High school graduate	0.89 [0.91]	0.59 [0.01]	0.51 [0.01]
Some college	0.73 [0.85]	0.86 [0.68]	0.87 [0.75]
College graduate	-	0.68 [0.47]	0.34 [0.05]
Father's Education Attainment
High school dropout	1	1	1
High school graduate	0.36 [0.31]	0.77 [0.20]	0.90 [0.70]
Some college	1.08 [0.96]	0.82 [0.56]	0.77 [0.57]
College graduate	-	0.43 [0.05]	1.25 [0.58]

E5. Marginal Effects from Negative Binomial Regressions of Years Being in Poverty Among Early Initiators and Living With Both Biological Parents At Age 14
Independent Variables	Years in Poverty between ages 25-29
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of observations	1801	201	2289	556
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	0.064 [0.12]	0.163 [0.35]	0.030 [0.41]	-0.023 [0.77]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	0.009 [0.97]	-0.053 [0.15]	-0.060 [0.40]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	-0.097 [0.09]	-	-0.109 [0.09]	-0.029 [0.79]
Age of sex initiation
Early Initiators	0	0	-	0
Else	-0.078 [0.07]	-0.168 [0.33]	-	0.120 [0.13]
Number of crimes/delinquencies (1980)
Committed 9+	0	0	0	-
Else	-0.072 [0.12]	-0.009 [0.96]	-0.044 [0.36]	-
Sex
Male	-0.199 [0.00]	-0.195 [0.22]	-0.327 [0.00]	-0.122 [0.16]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	0.503 [0.00]	0.120 [0.58]	0.456 [0.00]	0.544 [0.00]
Hispanic	0.101 [0.10]	0.291 [0.28]	0.121 [0.05]	0.037 [0.72]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.180 [0.00]	-0.356 [0.04]	-0.231 [0.00]	-0.225 [0.02]
Some college and up	-0.302 [0.00]	-0.291 [0.28]	-0.380 [0.00]	-0.342 [0.00]
Rosenberg self-esteem scale	-0.015 [0.00]	-0.023 [0.14]	-0.017 [0.00]	-0.022 [0.01]
Parent with alcohol problem
Yes	0.071 [0.09]	0.078 [0.57]	0.060 [0.14]	0.081 [0.29]
No	0	0	0	0
Mother's Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.114 [0.01]	-0.296 [0.04]	-0.162 [0.00]	-0.172 [0.03]
Some college	-0.182 [0.02]	-0.572 [0.02]	-0.121 [0.10]	-0.095 [0.49]
College graduate	-0.187 [0.04]	-0.324 [0.23]	-0.356 [0.00]	-0.411 [0.02]
Father's Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.066 [0.15]	-0.365 [0.01]	-0.109 [0.01]	-0.090 [0.27]
Some college	-0.083 [0.22]	0.023 [0.91]	-0.088 [0.19]	-0.167 [0.23]
College graduate	-0.078 [0.29]	-0.017 [0.94]	-0.132 [0.09]	0.012 [0.92]

E6. Odds Ratios from Logistic Regressions of Ever Being on Welfare Among Early Initiators Living With Both Biological Parents At Age 14
Independent Variables	Ever on Welfare Between Ages 21-33
Independent Variables	Early initiators of alcohol at ages 11-15	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of observations	803	897	202	2313	565
Age of alcohol initiation
Early Initiators	-	1	1	1	1
Else	-	0.99 [0.97]	1.04 [0.92]	0.82 [0.12]	0.98 [0.92]
Age of marijuana initiation
Early Initiators	1	-	1	1	1
Else	1.05 [0.82]	-	0.97 [0.93]	0.95 [0.65]	0.64 [0.06]
Age of cocaine initiation
Early Initiators	1	1	-	1	1
Else	0.65 [0.13]	0.70 [0.11]	-	0.72 [0.09]	0.89 [0.74]
Age of sex initiation
Early Initiators	1	1	1	-	1
Else	0.55 [0.01]	0.73 [0.15]	0.44 [0.14]	-	0.47 [0.01]
Number of crimes/delinquencies (1980)
Committed 9+	1	1	1	1	-
Else	0.84 [0.41]	0.59 [0.01]	0.78 [0.56]	0.69 [0.01]	-
Sex
Male	0.47 [0.00]	0.48 [0.00]	0.63 [0.13]	0.31 [0.00]	0.38 [0.00]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	2.12 [0.00]	3.14 [0.00]	2.07 [0.16]	2.56 [0.00]	3.18 [0.00]
Hispanic	1.41 [0.15]	1.44 [0.11]	0.89 [0.83]	1.29 [0.09]	2.04 [0.02]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.63 [0.07]	0.52 [0.01]	0.31 [0.03]	0.54 [0.00]	0.38 [0.00]
Some college and up	0.19 [0.00]	0.37 [0.00]	0.14 [0.00]	0.27 [0.00]	0.22 [0.00]
Rosenberg self-esteem scale	0.95 [0.05]	1.00 [0.90]	0.98 [0.67]	0.98 [0.08]	0.98 [0.58]
Parent with alcohol problem
Yes	1.36 [0.15]	1.81 [0.00]	3.02 [0.00]	1.50 [0.00]	2.03 [0.01]
No	1	1	1	1	1
Mother's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.76 [0.19]	0.99 [0.97]	0.50 [0.11]	0.87 [0.25]	0.83 [0.49]
Some college	0.71 [0.39]	0.61 [0.19]	0.73 [0.65]	0.85 [0.48]	0.72 [0.47]
College graduate	0.69 [0.41]	0.82 [0.65]	0.85 [0.84]	0.64 [0.12]	0.65 [0.43]
Father's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.87 [0.52]	0.57 [0.01]	0.52 [0.15]	0.68 [0.00]	0.61 [0.06]
Some college	0.87 [0.66]	0.56 [0.07]	1.13 [0.84]	0.87 [0.46]	0.86 [0.72]
College graduate	0.48 [0.07]	0.30 [0.00]	0.54 [0.39]	0.48 [0.00]	0.77 [0.53]

E7. Marginal Effects from Negative Binomial Regressions of Years Being on Welfare Among Early Initiators Living With Both Biological parents At Age 14
Independent Variables	Years on Welfare Between Ages 21-33
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of observations	1815	202	2313	565
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	-0.096 [0.24]	-0.076 [0.79]	-0.168 [0.08]	-0.159 [0.29]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	0.067 [0.88]	-0.060 [0.45]	-0.290 [0.05]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	-0.174 [0.13]	-	-0.215 [0.15]	-0.181 [0.44]
Age of sex initiation
Early Initiators	0	0	-	0
Else	-0.313 [0.00]	-0.652 [0.08]	-	-0.192 [0.24]
Number of crimes/delinquencies (1980)
Committed 9+	0	0	0	-
Else	-0.002 [0.98]	-0.152 [0.66]	-0.194 [0.06]	-
Sex
Male	-0.730 [0.00]	-0.714 [0.04]	-1.303 [0.00]	-0.510 [0.03]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	0.941 [0.00]	0.636 [0.29]	0.723 [0.00]	1.022 [0.01]
Hispanic	0.257 [0.05]	0.550 [0.32]	0.144 [0.26]	0.414 [0.16]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.523 [0.00]	-0.670 [0.07]	-0.522 [0.00]	-0.530 [0.02]
Some college and up	-0.757 [0.00]	-1.141 [0.01]	-0.904 [0.00]	-0.769 [0.00]
Rosenberg self-esteem scale	-0.026 [0.01]	-0.031 [0.37]	-0.027 [0.01]	-0.032 [0.08]
Parent with alcohol problem
Yes	0.332 [0.00]	0.495 [0.09]	0.286 [0.00]	0.439 [0.01]
No	0	0	0	0
Mother's Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.199 [0.02]	-0.755 [0.02]	-0.244 [0.01]	-0.175 [0.27]
Some college	-0.420 [0.00]	-0.664 [0.15]	-0.109 [0.51]	-0.126 [0.66]
College graduate	-0.548 [0.00]	-0.205 [0.71]	-0.586 [0.01]	-0.554 [0.12]
Father's Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.152 [0.08]	-0.516 [0.09]	-0.219 [0.02]	-0.278 [0.10]
Some college	-0.208 [0.11]	0.070 [0.87]	-0.187 [0.21]	-0.424 [0.11]
College graduate	-0.346 [0.01]	-0.549 [0.26]	-0.672 [0.00]	-0.291 [0.27]

E8. Marginal effects from OLS Regressions of Percent Time Employed Among Early Initiators Living With Both Biological parents At Age 14
Independent Variables	Percent Time Employed Since Leaving School
Independent Variables	Early initiators of marijuana at ages 11-19	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-15	Committed 9+ crimes/delinquencies in 1980
Number of observations	2360	201	842	565
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	-0.005 [0.62]	0.031 [0.50]	-0.019 [0.28]	-0.063[0.01]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	-0.025 [0.54]	0.011 [0.38]	0.010 [0.43]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	0.015 [0.10]	-	0.023 [0.25]	-0.002 [0.90]
Age of sex initiation
Early Initiators	0	0	-	0
Else	0.019 [0.00]	0.023 [0.28]	-	0.013 [0.31]
Number of crimes/delinquencies (1980)
Committed 9+	0	0	0	-
Else	0.014 [0.04]	-0.013 [0.54]	0.016 [0.28]	-
Sex
Male	0.011 [0.04]	0.019 [0.36]	0.041 [0.00]	-0.000 [1.00]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	-0.095 [0.00]	-0.062 [0.04]	-0.098 [0.00]	-0.100 [0.00]
Hispanic	-0.005 [0.54]	-0.032 [0.27]	-0.019 [0.28]	-0.019 [0.26]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	0.061 [0.00]	0.042 [0.18]	0.056 [0.00]	0.026 [0.19]
Some college and up	0.085 [0.00]	0.112 [0.00]	0.078 [0.00]	0.046 [0.03]
Rosenberg self-esteem scale	0.002 [0.02]	0.0002 [0.93]	0.004 [0.02]	0.001 [0.69]
Parent with alcohol problem
Yes	-0.010 [0.12]	0.001 [0.96]	-0.001 [0.93]	-0.015 [0.26]
No	0	0	0	0
Mother's Educational Attainment
High shool dropout	0	0	0	0
High school graduate	0.023 [0.00]	0.040 [0.09]	0.027 [0.05]	0.012 [0.40]
Some college	0.035 [0.00]	0.054 [0.16]	0.032 [0.19]	0.022 [0.33]
College graduate	0.029 [0.01]	0.036 [0.39]	-0.007 [0.83]	0.010 [0.70]
Father's Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.0004 [0.96]	-0.026 [0.28]	-0.004 [0.78]	0.002 [0.89]
Some college	-0.0008 [0.93]	-0.024 [0.48]	-0.004 [0.78]	0.029 [0.18]
College graduate	0.012 [0.19]	-0.020 [0.58]	0.039 [0.11]	0.022 [0.28]

E9. Marginal Effects from Weibull Regressions of Age when Reached 2 Years of Tenure Among Early Initiators Living With Both Biological parents At Age 14
Independent Variables	Age Reached Two Years Job Tenure
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-17	Committed 1+ crimes/delinquencies in 1980
Number of observations	1687	43	2110	2514
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	0.689 [0.04]	-1.538 [0.62]	1.006 [0.00]	0.496 [0.06]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	-0.541 [0.88]	-0.655 [0.03]	-0.629 [0.02]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	-2.158 [0.03]	-	-2.501 [0.02]	-2.015 [0.04]
Age of sex initiation
Early Initiators	0	0	-	0
Else	-1.658 [0.00]	-10.321 [0.01]	-	-1.255 [0.00]
Number of crimes/delinquencies (1980)
Committed more	0	0	0	-
Else	-0.608 [0.09]	0.440 [0.88]	-0.008 [0.98]	-
Sex
Male	-0.677 [0.04]	-1.325 [0.65]	-0.938 [0.00]	-0.712 [0.01]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	2.105 [0.00]	-1.548 [0.52]	1.175 [0.00]	1.506 [0.00]
Hispanic	0.772 [0.08]	-0.707 [0.80]	0.350 [0.41]	0.681 [0.07]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.754 [0.21]	0.611 [0.84]	-0.492 [0.33]	-1.030 [0.03]
Some college and up	-0.441 [0.49]	0.922 [0.79]	-0.152 [0.78]	-0.603 [0.02]
Rosenberg self-esteem scale	0.024 [0.54]	-0.189 [0.56]	-0.065 [0.08]	-0.022 [0.49]
Parent with alcohol problem
Yes	1.434 [0.00]	1.598 [0.61]	0.146 [0.69]	0.916 [0.00]
No	0	0	0	0
Mother's Educational Attainment
High school dropout	0	0	0	0
High school graduate	-1.032 [0.01]	-2.071 [0.39]	-1.133 [0.00]	-1.376 [0.00]
Some college	-1.338 [0.02]	0.444 [0.92]	-0.678 [0.24]	-1.067 [0.03]
College graduate	-0.729 [0.27]	-2.869 [0.55]	-1.249 [0.06]	-1.318 [0.02]
Father's Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.403 [0.31]	3.251 [0.17]	-0.550 [0.12]	-0.144 [0.65]
Some college	0.029 [0.96]	5.263 [0.27]	0.345 [0.51]	0.239 [0.60]
College graduate	-0.366 [0.51]	-1.169 [0.80]	-0.404 [0.46]	-0.250 [0.59]

E10. Relative Risk Ratios from Multinomial logistic Regression of Adult Marital and Fertility Status at age 33 Among Early Initiators Living With Both Biological parents At Age 14
Independent Variables	Adult Family Formation
	Early initiators of sex at ages 11-17
	Never married without children	Never married with children	Married without children	Married but divorced without children	Married but divorced with children
Number of observations = 2222
Age of alcohol initiation
11-15	1	1	1	1	1
16-17	0.92 [0.67]	1.34 [0.24]	1.50 [0.06]	1.04 [0.85]	0.79 [0.12]
18-19	1.18 [0.45]	1.87 [0.02]	1.65 [0.05]	0.96 [0.88]	0.96 [0.83]
not by 19	0.96 [0.87]	1.05 [0.88]	1.39 [0.29]	0.68 [0.34]	0.68 [0.08]
Age of marijuana initiation
11-15	1	1	1	1	1
16-17	1.22 [0.30]	0.76 [0.26]	1.07 [0.76]	0.66 [0.14]	0.93 [0.67]
18-19	1.02 [0.95]	0.66 [0.18]	0.93 [0.78]	1.18 [0.61]	0.84 [0.42]
not by 19	1.06 [0.77]	1.03 [0.91]	0.91 [0.68]	0.87 [0.60]	0.80 [0.18]
Age of cocaine initiation
11-15	1	1	1	1	1
16-17	1.68 [0.40]	0.61 [0.53]	6.77 [0.08]	0.81 [0.79]	2.32 [0.16]
18-19	1.46 [0.52]	1.00 [1.00]	3.86 [0.21]	0.94 [0.93]	1.51 [0.47]
not by 19	0.64 [0.43]	0.51 [0.27]	2.21 [0.46]	0.50 [0.30]	1.27 [0.67]
Age of sex initiation
11-15	-	-	-	-	-
16-17	-	-	-	-	-
18-19	-	-	-	-	-
not by 19	-	-	-	-	-
Number of crimes/delinquencies (1980)
9+	1	1	1	1	1
3-8	0.80 [0.30]	0.54 [0.01]	1.07 [0.76]	1.34 [0.33]	1.01 [0.96]
1-2	0.92 [0.71]	0.50 [0.01]	0.93 [0.80]	1.27 [0.45]	0.95 [0.78]
0	0.76 [0.22]	1.03 [0.91]	0.91 [0.68]	0.91 [0.78]	0.83 [0.33]
Sex
Male	2.01 [0.00]	0.87 [0.42]	1.26 [0.18]	0.65 [0.04]	0.54 [0.00]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	1.65 [0.00]	12.11 [0.00]	0.61 [0.03]	0.62 [0.09]	0.87 [0.38]
Hispanic	0.75 [0.22]	2.03 [0.02]	0.75 [0.24]	0.84 [0.55]	0.81 [0.22]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.38 [0.28]	0.82 [0.44]	1.11 [0.77]	1.28 [0.54]	1.02 [0.94]
Some college and up	2.59 [0.00]	0.72 [0.25]	3.04 [0.00]	1.58 [0.28]	0.83 [0.39]
Rosenberg self-esteem scale	0.97 [0.14]	0.98 [0.29]	1.00 [0.87]	1.04 [0.28]	1.04 [0.02]
Parent with alcohol problem
Yes	1.05 [0.81]	1.30 [0.19]	0.72 [0.13]	0.71 [0.20]	1.02 [0.87]
No	1	1	1	1	1
Mother's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.14 [0.46]	0.92 [0.66]	1.14 [0.53]	1.32 [0.27]	0.93 [0.61]
Some college	1.00 [1.00]	0.89 [0.76]	1.03 [0.92]	1.30 [0.49]	0.58 [0.05]
College graduate	0.93 [0.83]	0.40 [0.09]	1.44 [0.26]	0.75 [0.56]	0.74 [0.31]
Father's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.06 [0.74]	0.82 [0.32]	0.94 [0.75]	1.13 [0.64]	0.91 [0.50]
Some college	1.02 [0.95]	0.78 [0.50]	1.16 [0.60]	1.68 [0.11]	1.02 [0.93]
College graduate	1.27 [0.37]	0.72 [0.41]	1.21 [0.49]	1.00 [1.00]	0.85 [0.50]

Independent Variables	Adult Family Formation
	Committed 9+ crimes/delinquencies in 1980
	Never married without children	Never married with children	Married without children	Married but divorced without children	Married but divorced with children
Number of observations = 2608
Age of alcohol initiation
11-15	1	1	1	1	1
16-17	0.87 [0.39]	1.37 [0.20]	1.19 [0.33]	0.80 [0.34]	0.78 [0.11]
18-19	1.07 [0.71]	1.68 [0.06]	1.16 [0.47]	0.93 [0.79]	0.88 [0.48]
not by 19	1.10 [0.66]	1.05 [0.88]	1.13 [0.65]	0.70 [0.34]	0.81 [0.35]
Age of marijuana initiation
11-15	1	1	1	1	1
16-17	0.98 [0.92]	0.81 [0.38]	1.06 [0.77]	0.93 [0.78]	0.97 [0.88]
18-19	0.90 [0.63]	0.51 [0.04]	0.97 [0.89]	1.35 [0.32]	0.88 [0.56]
not by 19	0.78 [0.15]	0.71 [0.16]	0.86 [0.46]	0.90 [0.70]	0.82 [0.23]
Age of cocaine initiation
11-15	1	1	1	1	1
16-17	2.31 [0.20]	0.79 [0.75]	3.32 [0.15]	1.08 [0.92]	2.64 [0.10]
18-19	1.99 [0.27]	0.84 [0.79]	2.18 [0.34]	0.89 [0.87]	1.28 [0.67]
not by 19	0.92 [0.89]	0.45 [0.18]	1.35 [0.70]	0.46 [0.25]	1.44 [0.50]
Age of sex initiation
11-15	1	1	1	1	1
16-17	1.12 [0.52]	0.63 [0.02]	1.00 [0.98]	1.28 [0.35]	0.69 [0.02]
18-19	1.58 [0.02]	0.51 [0.01]	1.13 [0.59]	0.82 [0.52]	0.48 [0.00]
not by 19	2.59 [0.00]	0.31 [0.00]	1.50 [0.10]	0.89 [0.75]	0.26 [0.00]
Sex
Male	1.58 [0.00]	0.79 [0.21]	1.26 [0.11]	0.63 [0.02]	0.53 [0.00]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	2.16 [0.00]	11.74 [0.00]	0.98 [0.91]	0.76 [0.33]	0.78 [0.14]
Hispanic	0.87 [0.47]	1.85 [0.02]	0.84 [0.40]	0.79 [0.39]	0.81 [0.20]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.12 [0.67]	0.63 [0.07]	1.44 [0.34]	0.81 [0.57]	0.93 [0.74]
Some college and up	1.66 [0.06]	0.54 [0.03]	3.17 [0.00]	1.06 [0.87]	0.72 [0.14]
Rosenberg self-esteem scale	0.94 [0.00]	0.93 [0.00]	1.00 [0.83]	1.01 [0.74]	1.02 [0.23]
Parent with alcohol problem
Yes	1.14 [0.40]	1.29 [0.21]	0.67 [0.04]	0.83 [0.45]	1.10 [0.23]
No	1	1	1	1	1
Mother's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.10 [0.54]	0.69 [0.06]	0.98 [0.90]	1.21 [0.42]	0.84 [0.24]
Some college	1.24 [0.38]	0.91 [0.80]	1.12 [0.66]	1.32 [0.44]	0.75 [0.27]
College graduate	1.31 [0.31]	0.52 [0.21]	1.37 [0.26]	1.22 [0.64]	0.81 [0.47]
Father's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.25 [0.17]	1.03 [0.90]	1.06 [0.76]	1.01 [0.97]	0.98 [0.90]
Some college	1.00 [0.99]	0.87 [0.69]	1.20 [0.46]	1.30 [0.39]	0.87 [0.54]
College graduate	1.05 [0.84]	0.49 [0.10]	1.26 [0.33]	0.68 [0.27]	0.63 [0.06]

Appendix F: Regressions of Adult Outcomes on Adolescent Risky Behaviors and Other Independent Variables Among Early Initiators Living with Single Mothers at Age 14: Effects of Mothers' Education

Summary Table
Mothers' Education By Risky Behavior	Health Outcomes		Crime	Economic Outcomes
Mothers' Education By Risky Behavior	Alcohol abuse or dependence Odds Ratios [P-value]	Past-month drug use Odds Ratios [P-value]	In jail by age 33 Odds Ratios [P-value]	Ever in poverty ages 25-29 Odds Ratios [P-value]	Years in poverty ages 25-29 Marginal Effects [P-value]	Ever on welfare ages 21-33 Odds Ratios [P-value]	Years on welfare ages 21-33 Marginal Effects [P-value]	Percent Time Employed Marginal Effects [P-value]	Age achieved job lasting 2 years Marginal Effects [P-value]
Early Initiators of Alcohol
Mother's Education Attainment			N/A	N/A	N/A		N/A	N/A	N/A
High school dropout	1	1				1
High school graduate	0.74 [0.29]	0.90 [0.72]				1.14 [0.77]
Some college	0.42 [0.11]	1.78 [0.13]				2.49 [0.20]
College graduate	0.86 [0.77]	1.28 [0.63]				2.78 [0.18]
Early Initiators of Marijuana
Mother's Educational Attainment				N/A
High school dropout	1	1	1		0	1	0	0	0
High school graduate	0.81 [0.44]	0.99 [0.96]	1.40 [0.50]		0.050 [0.72]	1.17 [0.63]	-0.285 [0.37]	0.030 [0.11]	-1.686 [0.03]
Some college	0.33 [0.05]	1.95 [0.10]	1.32 [0.75]		-0.227 [0.35]	0.97 [0.95]	-0.633 [0.23]	0.028 [0.31]	-3.013 [0.01]
College graduate	0.69 [0.46]	1.09 [0.88]	0.99 [1.15]		-0.427 [0.14]	2.15 [0.18]	-0.707 [0.20]	0.021 [0.49]	-1.655 [0.19]
Early Initiators of Cocaine
Mother's Educational Attainment	N/A			N/A					N/A
High school dropout		1	1	-	0	1	0	0
High school graduate		2.85 [0.09]	2.46 [0.31]	-	0.306 [0.89]	0.47 [0.33]	0.120 [0.82]	-0.019 [0.73]	-
Some college		2.40 [0.31]	0.85 [0.93]	-	0.217 [0.76]	0.81 [0.88]	-0.569 [0.63]	-0.006 [0.95]	-
College graduate		2.26 [0.43]	-	-	0.384 [0.51]	1.82 [0.62]	0.090 [0.94]	0.013 [0.88]	-
Early Initiators of Sex
Mother's Educational Attainment
High school dropout	1	1	1	1	0	1	0	0	0
High school graduate	0.68 [0.12]	0.98 [0.93]	1.04 [0.90]	0.84 [0.51]	-0.138 [0.22]	0.85 [0.40]	-0.182 [0.44]	0.044 [0.06]	-0.782 [0.17]
Some college	0.40 [0.05]	2.08 [0.07]	1.46 [0.49]	0.74 [0.51]	-0.110 [0.59]	0.93 [0.83]	-0.392 [0.34]	0.075 [0.05]	-2.597 [0.01]
College graduate	0.36 [0.08]	1.53 [0.42]	1.33 [0.68]	1.17 [0.78]	-0.347 [0.22]	1.47 [0.35]	-0.255 [0.62]	-0.055 [0.23]	-1.035 [0.40]
Committed 9+ Crimes/Delinquencies in 1980
Mother's Educational Attainment
High school dropout	1	1	1	1	0	1	0	0	0
High school graduate	0.60 [0.32]	0.85 [0.56]	0.71 [0.58]	1.32 [0.50]	-0.083 [0.71]	0.43 [0.06]	-0.585 [0.11]	0.020 [0.56]	-0.896 [0.16]
Some college	-	1.78 [0.13]	-	0.54 [0.51]	-0.292 [0.61]	0.75 [0.75]	-1.399 [0.14]	0.119 [0.10]	-2.183 [0.03]
College graduate	-	1.28 [0.61]	-	0.82 [0.87]	-0.388 [0.57]	1.26 [0.85]	-0.440 [0.69]	0.063 [0.43]	-2.256 [0.06]

Probabilities Of Marital and Fertility Outcomes From Multi-Logistic Regressions Among Early Sex Initiators Living With Single mothers at Age 14 by mother's Educational Attainment
Marital and Fertility Outcome at Age 33	Mother's Educational Attainment
Marital and Fertility Outcome at Age 33	High school dropout	High school graduate	Some college	College and up
Never married without children	0.140	0.153	0.177	0.241
Never married with children	0.152	0.143	0.151	0.115
Married without children	0.053	0.096	0.115	0.131
Married with children	0.384	0.344	0.298	0.281
Married but divorced without children	0.082	0.058	0.083	-
Married but divorced with children	0.189	0.206	0.175	0.233

Probabilities Of Marital and Fertility Outcomes From Multi-Logistic Regressions Among Those Who Committed 9 or More Crimes/Delinquencies In 1980 Living With single mothers At Age 14 by Mother's Educational Attainment
Marital and Fertility Outcome at Age 33	Mother's Educational Attainment
Marital and Fertility Outcome at Age 33	High school dropout	High school graduate	Some college	College and up
Never married without children	0.220	0.199	0.283	0.276
Never married with children	0.145	0.126	0.125	0.096
Married without children	0.080	0.127	0.155	0.125
Married with children	0.368	0.362	0.286	0.219
Married but divorced without children	0.044	0.024	0.066	-
Married but divorced with children	0.142	0.161	0.085	0.285

F1. Odds Ratios from Logistic Regressions of Alcohol Abuse or Dependence Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Adult Alcohol Abuse or Dependence
Independent Variables	Early initiators of alcohol at ages 11-17	Early initiators of marijuana at ages 11-19	Early initiators of sex at ages 11-19	Committed 9+ crimes/delinquencies in 1980
Number of observations	559	586	915	128
Age of alcohol initiation
Early Initiators	-	1	1	1
Else	-	0.52 [0.02]	0.49 [0.00]	0.39 [0.18]
Age of marijuana initiation
Early Initiators	1	-	1	1
Else	0.48 [0.02]	-	0.45 [0.00]	0.33 [0.12]
Age of cocaine initiation
Early Initiators	1	1	1	1
Else	0.53 [0.29]	0.43 [0.13]	0.48 [0.18]	1.60 [0.71]
Age of sex initiation
Early Initiators	1	1	-	-
Else	0.79 [0.71]	0.89 [0.84]	-	-
Number of crimes/delinquencies (1980)
Committed more	1	1	1	-
Else	0.84 [0.54]	0.86 [0.58]	0.91 [0.72]	-
Sex
Male	2.61 [0.00]	1.84 [0.01]	2.46 [0.00]	2.03 [0.01]
Female	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1
Black	1.01 [0.98]	0.78 [0.39]	0.76 [0.29]	0.63 [0.47]
Hispanic	0.99 [0.98]	1.06 [0.85]	0.87 [0.66]	2.07 [0.23]
Educational Attainment
High school dropout	1	1	1	1
High school graduate	0.57 [0.07]	0.70 [0.24]	0.69 [0.17]	0.78 [0.70]
Some college and up	0.51 [0.06]	0.64 [0.20]	0.65 [0.16]	1.54 [0.58]
Rosenberg self-esteem scale	0.99 [0.98]	0.98 [0.46]	0.97 [0.33]	1.00 [1.00]
Parent with alcohol problem
Yes	1.65 [0.07]	0.98 [0.93]	1.33 [0.23]	1.99 [0.17]
No	1	1	1	1
Mother's Educational Attainment
High school dropout)	1	1	1	1
High school graduate)	0.74 [0.29]	0.81 [0.44]	0.68 [0.12]	0.60 [0.32]
Some college)	0.42 [0.11]	0.33 [0.05]	0.40 [0.05]	-
College graduate)	0.86 [0.77]	0.69 [0.46]	0.36 [0.08]	-

F2. Odds Ratios from Logistic Regressions of Past-month Drug Use Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Adult Drug Use
Independent Variables	Early initiators of alcohol ages 11-19	Early initiators of marijuana at ages 11-19	Early initiators of cocaine at ages 11-19	Early initiators of sex at ages 11-17	Committed 1+ crimes/delinquencies in 1980
Number of observations	789	572	117	725	655
Age of alcohol initiation
Early Initiators	-	1	-	1	1
Else	-	0.56 [0.17]	-	0.54 [0.12]	0.66 [0.25]
Age of marijuana initiation
Early Initiators	1	-	1	1	1
Else	0.46 [0.01]	-	2.41 [0.31]	0.36 [0.00]	0.47 [0.01]
Age of cocaine initiation
Early Initiators	1	1	-	1	1
Else	0.60 [0.08]	0.72 [0.28]	-	0.79 [0.45]	0.68 [0.20]
Age of sex initiation
Early Initiators	1	1	1	-	1
Else	0.68 [0.26]	0.55 [0.14]	1.35 [0.69]	-	0.84 [0.60]
Number of crimes/delinquencies (1980)
Committed 9+	1	1	1	1	-
Else	0.28 [0.00]	0.27 [0.00]	0.14 [0.02]	0.32 [0.00]	-
Sex
Male	1.17 [0.53]	1.06 [0.82]	0.44 [0.19]	1.07 [0.79]	1.23 [0.41]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	1.05 [0.87]	1.56 [0.16]	1.53 [0.53]	1.08 [0.79]	1.07 [0.83]
Hispanic	0.98 [0.95]	1.28 [0.52]	3.13 [0.09]	0.96 [0.92]	1.10 [0.79]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.51 [0.03]	0.48 [0.02]	0.32 [0.14]	0.43 [0.01]	0.49 [0.02]
Some college and up	0.41 [0.02]	0.39 [0.01]	0.28 [0.13]	0.35 [0.00]	0.45 [0.03]
Rosenberg self-esteem scale	1.04 [0.24]	1.04 [0.23]	1.08 [0.28]	1.06 [0.01]	1.04 [0.24]
Parent with alcohol problem
Yes	1.02 [0.94]	1.06 [0.84]	0.71 [0.56]	1.03 [0.93]	1.11 [0.70]
No	1	1	1	1	1
Mother's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.90 [0.72]	0.99 [0.96]	2.85 [0.09]	0.98 [0.93]	0.85 [0.56]
Some college	1.78 [0.13]	1.95 [0.10]	2.40 [0.31]	2.08 [0.07]	1.78 [0.13]
College graduate	1.28 [0.63]	1.09 [0.88]	2.26 [0.43]	1.53 [0.42]	1.28 [0.61]

F3. Odds Ratios from Logistic Regressions of Ever Being in Jail Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	In Jail by Age 33
Independent Variables	Early initiators of marijuana at ages 11-15	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-19	Committed 9+ crimes/delinquencies in 1980
Number of observations	270	62	743	116
Age of alcohol initiation
Early Initiators	1	1	1	1
Else	2.16 [0.13]	0.33 [0.19]	0.97 [0.93]	0.54 [0.34]
Age of marijuana initiation
Early Initiators	-	1	1	1
Else	-	1.01 [1.00]	0.73 [0.29]	0.65 [0.49]
Age of cocaine initiation
Early Initiators	1	-	1	1
Else	0.21 [0.00]	-	0.26 [0.00]	0.29 [0.14]
Age of sex initiation
Early Initiators	1	1	-	-
Else	0.65 [0.61]	0.26 [0.36]	-	-
Number of crimes/delinquencies (1980)
Committed 9+	1	1	1	-
Else	0.76 [0.61]	0.97 [0.97]	0.82 [0.54]	-
Sex
Male	6.36 [0.00]	6.74 [0.10]	10.29 [0.00]	8.88 [0.05]
Female	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1
Black	3.09 [0.06]	1.41 [0.73]	2.94 [0.01]	1.59 [0.52]
Hispanic	2.06 [0.31]	4.00 [0.19]	2.16 [0.12]	0.78 [0.78]
Educational Attainment
High school dropout	1	1	1	1
High school graduate	0.28 [0.02]	2.00 [0.54]	0.54 [0.06]	0.77 [0.73]
Some college and up	0.24 [0.03]	3.27 [0.40]	0.21 [0.00]	0.25 [0.27]
Rosenberg self-esteem scale	0.95 [0.03]	0.79 [0.04]	0.96 [0.30]	0.91 [0.25]
Parent with alcohol problem
Yes	0.58 [0.30]	0.43 [0.35]	1.08 [0.81]	0.40 [0.28]
No	1	1	1	1
Mother's Educational Attainment
High school dropout	1	1	1	1
High school graduate	1.40 [0.50]	2.46 [0.31]	1.04 [0.90]	0.71 [0.58]
Some college	1.32 [0.75]	0.85 [0.93]	1.46 [0.49]	-
College graduate	0.99 [1.15]	-	1.33 [0.68]	-

F4.Odds Ratios from Logistic Regressions of Ever Being In Poverty Ages 25-29 Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Ever in Poverty Between Ages 25-29
Independent Variables	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-15	Committed 9+ crimes/delinquencies in 1980
Number of observations	16	397	146
Age of alcohol initiation
Early Initiators	-	1	1
Else	-	1.20 [0.46]	1.29 [0.58]
Age of marijuana initiation
Early Initiators	-	1	1
Else	-	1.02 [0.92]	1.83 [0.17]
Age of cocaine initiation
Early Initiators	-	1	1
Else	-	0.42 [0.36]	0.35 [0.38]
Age of sex initiation
Early Initiators		-	1
Else	-	-	0.53 [0.17]
Number of crimes/delinquencies (1980)
Committed 9+	-	1	-
Else	-	0.87 [0.61]	-
Sex
Male	-	0.36 [0.00]	0.62 [0.28]
Female	-	1	1
Race/Ethnicity
Non-black non-Hispanic	-	1	1
Black	-	2.15 [0.02]	3.36 [0.02]
Hispanic	-	0.77 [0.54]	1.31 [0.64]
Educational Attainment
High school dropout	-	1	1
High school graduate	-	0.29 [0.00]	0.42 [0.09]
Some college and up	-	0.15 [0.00]	0.38 [0.14]
Rosenberg self-esteem scale	-	0.94 [0.03]	0.89 [0.03]
Parent with alcohol problem
Yes	-	1.12 [0.69]	1.01 [1.00]
No	-	1	1
Mother's Educational Attainment
High school dropout	-	1	1
High school graduate	-	0.84 [0.51]	1.32 [0.50]
Some college	-	0.74 [0.51]	0.54 [0.51]
College graduate	-	1.17 [0.78]	0.82 [0.87]

F5. Marginal Effects from Negative Binomial Regressions of Years Being in Poverty ages 25-29 Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Years in Poverty Between Ages 25-29
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of observations	465	65	733	142
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	0.027 [0.83]	0.688 [0.07]	0.100 [0.31]	0.061 [0.80]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	-1.803 [0.00]	-0.023 [0.82]	0.304 [0.18]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	-0.488 [0.01]	-	-0.317 [0.09]	-0.893 [0.00]
Age of sex initiation
Early Initiators	0	0	-	0
Else	-0.032 [0.85]	-0.154 [0.75]	-	-0.365 [0.35]
Number of crimes/delinquencies (1980)
Committed more	0	0	0	-
Else	0.186 [0.26]	-0.476 [0.26]	-0.070 [0.60]	-
Sex
Male	-0.438 [0.00]	-0.378 [0.32]	-0.769 [0.00]	-0.521 [0.07]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	0.695 [0.00]	1.032 [0.10]	0.564 [0.00]	0.854 [0.01]
Hispanic	0.346 [0.12]	0.354 [0.54]	0.263 [0.20]	0.384 [0.40]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.282 [0.07]	-0.538 [0.21]	-0.408 [0.00]	-0.462 [0.08]
Some college and up	-0.521 [0.00]	-0.514 [0.27]	-0.817 [0.00]	-0.391 [0.27]
Rosenberg self-esteem scale	-0.022 [0.16]	-0.028 [0.42]	-0.036 [0.00]	-0.053 [0.07]
Parent with alcohol problem
Yes	-0.050 [0.71]	-0.288 [0.38]	-0.054 [0.63]	-0.250 [0.32]
No	0	0	0	0
Mother's Educational Attainment
High school dropout	0	0	0	0
High school graduate	0.050 [0.72]	0.306 [0.89]	-0.138 [0.22]	-0.083 [0.71]
Some college	-0.227 [0.35]	0.217 [0.76]	-0.110 [0.59]	-0.292 [0.61]
College graduate	-0.427 [0.14]	0.384 [0.51]	-0.347 [0.22]	-0.388 [0.57]

F6. Odds Ratios from Logistic Regressions of Ever Being on Welfare ages 21-33 Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Ever on Welfare Between Ages 21-33
Independent Variables	Early initiators of alcohol at ages 11-15	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of observations	193	270	66	743	145
Age of alcohol initiation
Early Initiators	-	1	1	1	1
Else	-	1.60 [0.12]	0.95 [0.94]	1.07 [0.74]	2.51 [0.07]
Age of marijuana initiation
Early Initiators	1	-	1	1	1
Else	2.06 [0.06]	-	0.87 [0.85]	1.08 [0.70]	0.75 [0.54]
Age of cocaine initiation
Early Initiators	1	1	-	1	1
Else	0.24 [0.01]	0.45[0.05]	-	0.60 [0.12]	0.20 [0.03]
Age of sex initiation
Early Initiators	1	1	1	-	1
Else	0.94 [0.91]	0.84 [0.64]	3.94 [0.19]	-	1.35 [0.69]
Number of crimes/delinquencies (1980)
Committed 9+	1	1	1	1	-
Else	2.35 [0.08]	1.04 [0.92]	1.46 [0.64]	1.04 [0.85]	-
Sex
Male	0.36 [0.01]	0.37 [0.00]	0.66 [0.59]	0.19 [0.00]	0.23 [0.00]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	6.70 [0.00]	3.47 [0.00]	5.69 [0.04]	3.73 [0.00]	5.43 [0.00]
Hispanic	4.01 [0.15]	2.41 [0.03]	1.45 [0.68]	2.58 [0.00]	2.58 [0.12]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.51 [0.17]	0.62 [0.20]	0.15 [0.65]	0.43 [0.00]	0.15 [0.00]
Some college and up	0.25 [0.02]	0.37 [0.02]	0.08 [0.06]	0.22 [0.00]	0.24 [0.06]
Rosenberg self-esteem scale	0.80 [0.00]	0.95 [0.18]	0.84 [0.05]	0.95 [0.03]	0.85 [0.00]
Parent with alcohol problem
Yes	1.54 [0.29]	1.46 [0.22]	1.65 [0.51]	1.26 [0.26]	0.50 [0.18]
No	1	1	1	1	1
Mother's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.14 [0.77]	1.17 [0.63]	0.47 [0.33]	0.85 [0.40]	0.43 [0.06]
Some college	2.49 [0.20]	0.97 [0.95]	0.81 [0.88]	0.93 [0.83]	0.75 [0.75]
College graduate	2.78 [0.18]	2.15 [0.18]	1.82 [0.62]	1.47 [0.35]	1.26 [0.85]

F7. Marginal Effects from Negative Binomial Regressions of Years Being on Welfare ages 21-33 Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Years on Welfare Between Ages 21-33
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-17	Committed 9+ crimes/delinquencies in 1980
Number of observations	471	66	743	145
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	0.727 [0.02]	0.754 [0.09]	0.310 [0.23]	1.384 [0.00]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	-0.916 [0.21]	-0.396 [0.06]	-0.579 [0.15]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	-0.549 [0.19]	-	-0.056 [0.89]	-1.309 [0.02]
Age of sex initiation
Early Initiators	0	0	-	0
Else	-0.368 [0.31]	1.145 [0.15]	-	-0.090 [0.89]
Number of crimes/delinquencies (1980)
Committed 9+	0	0	0	-
Else	0.020 [0.95]	-0.049 [0.94]	-0.192 [0.48]	-
Sex
Male	-1.84 [0.00]	-0.982 [0.17]	-3.102 [0.00]	-2.607 [0.00]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	1.410 [0.00]	1.655 [0.10]	1.003 [0.00]	0.827 [0.13]
Hispanic	0.692 [0.18]	0.408 [0.63]	0.679 [0.11]	0.715 [0.34]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.622 [0.08]	-0.975 [0.11]	-0.885 [0.00]	-2.020 [0.00]
Some college and up	-1.205 [0.01]	-2.083 [0.01]	-1.744 [0.00]	-2.250 [0.00]
Rosenberg self-esteem scale	-0.104 [0.00]	-0.171 [0.01]	-0.089 [0.00]	-0.119 [0.02]
Parent with alcohol problem
Yes	0.261 [0.38]	-0.233 [0.68]	0.238 [0.31]	-0.742 [0.07]
No	0	0	0	0
Mother's Educational Attainment
High school dropout	0	0	0	0
High school graduate	-0.285 [0.37]	0.120 [0.82]	-0.182 [0.44]	-0.585 [0.11]
Some college	-0.633 [0.23]	-0.569 [0.63]	-0.392 [0.34]	-1.399 [0.14]
College graduate	-0.707 [0.20]	0.090 [0.94]	-0.255 [0.62]	-0.440 [0.69]

F8. Marginal effects from OLS Regressions of Percent Time Employed Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Percent Time Employed Since Leaving School
Independent Variables	Early initiators of marijuana at ages 11-19	Early initiators of cocaine at ages 11-17	Early initiators of sex at ages 11-15	Committed 9+ crimes/delinquencies in 1980
Number of observations	577	66	390	143
Age of alcohol initiation
Early Initiators	0	0	0	0
Else	0.024 [0.26]	-0.085 [0.38]	0.045 [0.08]	0.020 [0.68]
Age of marijuana initiation
Early Initiators	-	0	0	0
Else	-	0.079 [0.38]	0.012 [0.58]	0.046 [0.23]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	0.031 [0.28]	-	0.002 [0.96]	0.031 [0.54]
Age of sex initiation
Early Initiators	0	0	-	0
Else	0.012 [0.46]	-0.032 [0.52]	-	0.031 [0.37]
Number of crimes/delinquencies (1980)
Committed 9+	0	0	0	-
Else	0.014 [0.48]	0.049 [0.40]	0.020 [0.42]	-
Sex
Male	0.051 [0.00]	0.003 [0.95]	0.090 [0.00]	0.048 [0.18]
Female	0	0	0	0
Race/Ethnicity
Non-black non-Hispanic	0	0	0	0
Black	-0.121 [0.00]	-0.145 [0.02]	-0.105 [0.00]	-0.169 [0.00]
Hispanic	-0.005 [0.84]	-0.076 [0.26]	-0.020 [0.58]	-0.057 [0.20]
Educational Attainment
High school dropout	0	0	0	0
High school graduate	0.069 [0.00]	0.165 [0.02]	0.061 [0.03]	0.121 [0.01]
Some college and up	0.115 [0.00]	0.176 [0.04]	0.115 [0.00]	0.104 [0.05]
Rosenberg self-esteem scale	0.008 [0.00]	0.010 [0.95]	0.011 [0.00]	0.013 [0.00]
Parent with alcohol problem
Yes	-0.017 [0.33]	0.071 [0.19]	0.016 [0.51]	0.017 [0.65]
No	0	0	0	0
Mother's Educational Attainment
High school dropout	0	0	0	0
High school graduate	0.030 [0.11]	-0.019 [0.73]	0.044 [0.06]	0.020 [0.56]
Some college	0.028 [0.31]	-0.006 [0.95]	0.075 [0.05]	0.119 [0.10]
College graduate	0.021 [0.49]	0.013 [0.88]	-0.055 [0.23]	0.063 [0.43]

F9. Marginal Effects from Weibull Regressions of Age when Reached 2 Years of Tenure Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Age Reached Two Years Job Tenure
Independent Variables	Early initiators of marijuana at ages 11-17	Early initiators of cocaine at ages 11-15	Early initiators of sex at ages 11-17	Committed 1+ crimes/delinquencies in 1980
Number of observations	391	14	610	553
Age of alcohol initiation
Early Initiators	0	-	0	0
Else	0.915 [0.20]	-	0.711 [0.18]	0.567 [0.32]
Age of marijuana initiation
Early Initiators	-	-	0	0
Else	-	-	0.311 [0.55]	0.691 [0.23]
Age of cocaine initiation
Early Initiators	0	-	0	0
Else	0.674 [0.71]	-	1.958 [0.28]	-0.878 [0.72]
Age of sex initiation
Early Initiators	0	-	-	0
Else	0.414 [0.62]	-	-	0.182 [0.79]
Number of crimes/delinquencies (1980)
Committed 9+	0	-	0	-
Else	0.488 [0.51]	-	-0.369 [0.51]	-
Sex
Male	0.159 [0.81]	-	-1.201 [0.03]	-0.070 [0.91]
Female	0	-	0	0
Race/Ethnicity
Non-black non-Hispanic	0	-	0	0
Black	-0.096 [0.90]	-	-0.193 [0.77]	-0.318 [0.64]
Hispanic	-0.786 [0.38]	-	-0.686 [0.39]	-0.588 [0.48]
Educational Attainment
High school dropout	0		0	0
High school graduate	0.712 [0.44]	-	0.278 [0.70]	-0.526 [0.53]
Some college and up	1.051 [0.30]	-	-0.070 [0.93]	-0.461 [0.61]
Rosenberg self-esteem scale	-0.049 [0.58]	-	0.010 [0.88]	-0.008 [0.91]
Parent with alcohol problem
Yes	1.500 [0.04]	-	1.095 [0.07]	0.936 [0.15]
No	0	-	0	0
Mother's Educational Attainment
High school dropout	0	-	0	0
High school graduate	-1.686 [0.03]	-	-0.782 [0.17]	-0.896 [0.16]
Some college	-3.013 [0.01]	-	-2.597 [0.01]	-2.183 [0.03]
College graduate	-1.655 [0.19]	-	-1.035 [0.40]	-2.256 [0.06]

F10. Relative Risk Ratios from Multinomial logistic Regression of Adult Marital and Fertility Status at age 33 Among Early Initiators Living With Single Mothers At Age 14
Independent Variables	Adult Family Formation
	Early initiators of sex at ages 11-17
	Never married without children	Never married with children	Married without children	Married but divorced without children	Married but divorced with children
Number of observations=705
Age of alcohol initiation
11-15	1	1	1	1	1
16-17	1.03 [0.91]	1.61 [0.13]	2.12 [0.12]	0.25 [0.04]	1.40 [0.29]
18-19	0.84 [0.65]	1.51 [0.25]	1.61 [0.41]	0.74 [0.64]	0.89 [0.76]
not by 19	0.78 [0.56]	1.16 [0.70]	1.30 [0.67]	0.09 [0.04]	1.55 [0.24]
Age of marijuana initiation
11-15	1	1	1	1	1
16-17	1.02 [0.96]	0.75 [0.36]	0.68 [0.45]	0.94 [0.92]	0.65 [0.19]
18-19	1.38 [0.46]	0.82 [0.63]	1.05 [0.94]	0.67 [0.65]	0.53 [0.18]
not by 19	1.07 [0.84]	0.70 [0.22]	1.42 [0.43]	0.54 [0.36]	0.93 [0.81]
Age of cocaine initiation
11-15	1	1	1	1	1
16-17	0.48 [0.43]	2.53 [0.50]	-	-	1.02 [0.98]
18-19	0.09 [0.01]	2.04 [0.60]	0.33 [0.42]	0.37 [0.51]	0.34 [0.31]
not by 19	0.23 [0.08]	1.95 [0.61]	0.46 [0.55]	0.92 [0.95]	0.91 [0.92]
Number of crimes/delinquencies (1980)
9+	1	1	1	1	1
3-8	0.74 [0.42]	0.61 [0.15]	0.94 [0.89]	1.04 [0.96]	0.91 [0.80]
1-2	1.03 [0.93]	0.73 [0.36]	0.73 [0.55]	0.39 [0.28]	1.09 [0.82]
0	0.47 [0.06]	0.55 [0.09]	0.29 [0.03]	0.96 [0.96]	0.67 [0.30]
Sex
Male	3.46 [0.00]	0.66 [0.07]	2.24 [0.04]	0.54 [0.24]	0.78 [0.32]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	1.01 [0.96]	6.15 [0.00]	0.70 [0.38]	0.27 [0.02]	0.73 [0.29]
Hispanic	0.64 [0.27]	1.67 [0.26]	0.40 [0.12]	0.27 [0.07]	0.79 [0.52]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.97 [0.94]	0.62 [0.09]	1.47 [0.53]	1.15 [0.86]	1.10 [0.77]
Some college and up	1.39 [0.40]	0.41 [0.01]	2.68 [0.11]	4.71 [0.05]	1.01 [0.97]
Rosenberg self-esteem scale	0.97 [0.43]	0.94 [0.04]	0.97 [0.47]	0.93 [0.29]	0.95 [0.09]
Parent with alcohol problem
Yes	0.59 [0.09]	0.73 [0.24]	0.90 [0.79]	0.83 [0.73]	1.17 [0.56]
No	1	1	1	1	1
Mother's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	1.27 [0.41]	1.05 [0.86]	2.11 [0.06]	0.74 [0.60]	1.23 [0.45]
Some college	1.74 [0.23]	1.27 [0.59]	3.02 [0.06]	1.31 [0.74]	1.19 [0.70]
College graduate	2.68 [0.06]	1.03 [0.97]	3.81 [0.06]	-	1.73 [0.33]

Independent Variables	Adult Family Formation
	Committed 9+ crimes/delinquencies in 1980
	Never married without children	Never married with children	Married without children	Married but divorced without children	Married but divorced with children
Number of observations = 639
Age of alcohol initiation
11-15	1	1	1	1	1
16-17	0.83 [0.55]	1.50 [0.25]	2.50 [0.07]	0.26 [0.07]	1.21 [0.58]
18-19	0.52 [0.07]	1.09 [0.83]	1.56 [0.43]	0.77 [0.69]	0.74 [0.46]
not by 19	0.37 [0.02]	1.02 [0.96]	1.65 [0.41]	-	1.05 [0.91]
Age of marijuana initiation
11-15	1	1	1	1	1
16-17	1.09 [0.80]	0.68 [0.27]	0.59 [0.28]	1.23 [0.77]	0.63 [0.21]
18-19	1.27 [0.56]	0.64 [0.31]	0.80 [0.70]	0.88 [0.89]	0.81 [0.64]
not by 19	1.25 [0.51]	1.04 [0.90]	1.13 [0.77]	0.52 [0.45]	1.17 [0.65]
Age of cocaine initiation
11-15	1	-	1	1	-
16-17	1.21 [0.87]	-	-	-	-
18-19	0.22 [0.18]	-	0.27 [0.37]	-	-
not by 19	0.49 [0.50]	-	0.31 [0.41]	0.12 [0.13]	-
Age of sex initiation
11-15	1	1	1	1	1
16-17	1.23 [0.49]	0.63 [0.10]	1.52 [0.29]	1.08 [0.90]	0.91 [0.74]
18-19	1.06 [0.88]	0.44 [0.03]	1.12 [0.82]	0.48 [0.44]	0.40 [0.03]
not by 19	2.42 [0.07]	0.33 [0.07]	1.80 [0.38]	0.48 [0.55]	0.41 [0.13]
Sex
Male	1.93 [0.02]	0.43 [0.00]	2.04 [0.06]	0.47 [0.20]	0.66 [0.14]
Female	1	1	1	1	1
Race/Ethnicity
Non-black non-Hispanic	1	1	1	1	1
Black	0.94 [0.85]	6.10 [0.00]	1.05 [0.90]	0.38 [0.15]	0.93 [0.82]
Hispanic	0.87 [0.71]	2.29 [0.09]	0.55 [0.29]	0.36 [0.19]	0.74 [0.47]
Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.85 [0.66]	0.73 [0.34]	1.73 [0.42]	4.10 [0.25]	1.04 [0.91]
Some college and up	1.19 [0.66]	0.32 [0.00]	3.97 [0.04]	8.40 [0.09]	0.74 [0.47]
Rosenberg self-esteem scale	0.98 [0.52]	0.94 [0.04]	0.98 [0.05]	0.90 [0.16]	0.95 [0.12]
Parent with alcohol problem
Yes	0.55 [0.04]	0.62 [0.11]	0.52 [0.10]	0.83 [0.76]	0.97 [0.91]
No	1	1	1	1	1
Mother's Educational Attainment
High school dropout	1	1	1	1	1
High school graduate	0.92 [0.78]	0.87 [0.62]	1.68 [0.18]	0.51 [0.33]	1.15 [0.64]
Some college	1.75 [0.21]	1.02 [0.97]	2.72 [0.07]	2.06 [0.37]	0.73 [0.61]
College graduate	2.21 [0.13]	1.09 [0.90]	2.78 [0.13]	-	3.51 [0.02]

References

Amato, P. R. (1999). Children of divorced parents as young adults. In E. M. Hetherington (Ed.), Coping With Divorce, Single Parenting, and Remarriage (pp. 147-164). Mahwah, NJ: Lawrence Erlbaum Associates.

Amato, P. R. (1996). Explaining the intergenerational transmission of divorce. Journal of Marriage and the Family, 58, 628-640.

Amato, P. R., and Keith, B. (1991a). Parental divorce and the well-being of children: A meta-analysis. Psychological Bulletin, 110, 26-46.

Amato, P. R., and Keith, B. (1991b). Parental divorce and adult well-being: A meta-analysis. Journal of Marriage and the Family, 53, 43-58.

Amato, P. R., Loomis, L. S., and Booth, A. (1995). Parental divorce, marital conflict, and offspring well-being during early adulthood. Social Forces, 73, 895-915.

American Psychiatric Association. (1987). Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised. Washington, DC: Author.

Anderson, K. H., Mitchell, J. M., and Butler, J. S. (1993). The effect of deviance during adolescence on the choice of jobs. Southern Economic Journal, 60, 341-357.

Avenevoli, S., Sessa, F. M., and Steinberg, L. (1999). Family structure, parenting practices, and adolescent adjustment: An ecological examination. In E. M. Hetherington (Ed.) Coping With Divorce, Single Parenting, and Remarriage (65-92). Mahwah, NJ: Lawrence Erlbaum Associates.

Bardone, A. M., Moffitt, T. E., Caspi, A., Dickson, N., Stanton, W. S., and Silva, P. A. (1998). Adult physical health outcomes of adolescent girls with conduct disorder, depression, and anxiety. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 594-601.

Bergmark, K. H., and Andersson, T. (1999). The development of advanced drinking habits in adolescence--a longitudinal study. Substance Use and Misuse, 34, 171-194.

Boggess, S. (1998). Family structure, economic status, and educational attainment. The Journal of Population Economics, 11, 205-222.

Booth, A., and Amato, P. R. (1994). Parental marital quality, parental divorce and relations with parents. Journal of Marriage and the Family, 56, 21-34.

Brook, J. S., Balka, E. B., and Whiteman, M. (1999). The risks for late adolescence of early adolescent marijuana use. American Journal of Public Health, 89, 1549-1554.

Center on Addiction and Substance Abuse, CASA (1999). Dangerous liaisons: Substance abuse and sex [On-line]. Available: http://www.casacolumbia.org/usr_doc/21598.PDF.

Chase-Lansdale, P. L., Cherlin, A. J., and Kiernan, K. E. (1995). The long-term effects of parental divorce on the mental health of young adults: A developmental perspective. Child Development, 66, 1614-1634.

Chassin, L., Pillow, F., Curran, P. Molina, B. S., and Barrera, M. (1993). Relation of parental alcoholism to early adolescent substance use: A test of three mediating mechanisms. Journal of Abnormal Psychology, 102, 3-19.

Chassin, L., Pitts, S. C., DeLucia, C., and Todd, M. (1999). A longitudinal study of children of alcoholics: Predicting young adult substance use disorders, anxiety, and depression. Journal of Abnormal Psychology, 108, 106-119.

Chou, S. P., and Pickering, R. P. (1992). Early onset of drinking as a risk factor for lifetime alcohol-related problems. British Journal of Addiction, 87, 1199-1204.

Cooper, M. L., Peirce, R. S., and Huselid, R. F. (1994). Substance use and sexual risk taking among black adolescents and white adolescents. Health Psychology, 13, 251-262.

Cox, David R., and Snell, E. J (1989). The Analysis of Binary Data, Second Edition, London, Chapman and Hall.

Dawson, D. A., Grant, B. F., and Harford, T. C. (1992). Parental history of alcoholism and probability of marriage. Journal of Substance Abuse, 4, 117-129.

Dawson, D. A., Harford, T. C., and Grant, B. F. (1992). Parental history of alcoholism and probability of marriage. Journal of Substance Abuse, 4, 117-129.

Donovan, J. E., Jessor, R., and Costa, F. (1988). The syndrome of problem behavior in adolescence: A replication. Journal of Consulting and Clinical Psychology, 56, 762-765.

Elder, G. H., Jr. (1994). Time, agency, and change: Perspectives on the life course. Social Psychology Quarterly, 57, 5-15.

Farrington, D. (1991). Childhood aggression and adult violence: Early precursors and later life outcomes. In D. J. Pepler and K. H. Rubin (Eds.), The Development and Treatment of Childhood Aggression (pp. 189-197). Hillsdale, NJ: Lawrence Erlbaum Associates.

Fergusson, D. M., and Horwood, L. J. (1997). Early onset cannabis use and psychosocial adjustment in young adults. Addiction, 92, 279-296.

Fergusson, D. M., Lynskey, M. T., and Horwood, L. J. (1994). Childhood exposure to alcohol and adolescent drinking patterns. Addiction, 89, 1007-1016.

Fergusson, D. M., Lynskey, M. T., and Horwood, L. J. (1997). The short term consequences of early cannabis use. Journal of Abnormal Child Psychology, 24, 499-512.

Freeman, Richard B. (1996), Why Do So Many Young American Men Commit Crimes and What Might We Do About It?, The Journal of Economic Perspectives,10:1, 25-42.

Furstenberg, F. F., Jr., and Cherlin, A. J. (1991). Divided Families. Cambridge, MA: Harvard University Press.

Furstenberg, F. F., Jr., Hughes, M. E., and Brooks-Gunn, J. (1992). The next generation: The children of teenage mothers grow up. In M. K. Rosenheim and M. F. Testa (Eds.), Early Parenthood and Coming of Age in the 1990s (pp. 113-135). New Brunswick, NJ: Rutgers University Press.

Ginther, D. K., and Pollack, R. A. (April, 2001). Does family structure affect children's educational outcomes? Paper presented at Society of Labor Economists' annual conference.

Giordano, P. C., Millhollin, T. J., Cernkovich, S. A., Pugh, M. D., and Rudolph, J. L. (1999). Delinquency, identity, and women's involvement in relationship violence. Criminology, 37, 17-37.

Glueck, S., and Glueck, D. (1950). Unraveling Juvenile Delinquency. New York: Commonwealth Fund.

Goodwin, D. W., Schulsinger, F., Knop, J., Mednick, S., and Guze, S. B. (1977). Alcoholism and depression in adopted-out daughters of alcoholics. Archives of General Psychiatry, 34, 751-755.

Gotham, H. J., and Sher, K. J. (1996). Children of alcoholics. In J. Kinney (Ed.), Clinical Manual of Substance Abuse (2^nd ed., pp. 272-287). New York: Mosby.

Grant, B. F. (1998). The impact of a family history of alcoholism on the relationship between age at onset of alcohol use and DSM-IV alcohol dependence. Alcohol Health and Research World, 22, 144-147.

Grant, B. F. and Dawson, D. A. (1997). Age at onset of alcohol use and its association with DSM-IV alcohol abuse and dependence: Results from the National Longitudinal Alcohol Epidemiologic Survey. Journal of Substance Abuse, 9, 103-110.

Greene, William H. (2000). Econometric Analysis, Third Edition, New York, Macmillan Publishing Company.

Greenburg, David F. (1985). Age, Crime, and Social Explanation. American Journal of Sociology, 91:1, 1-21.

Gruber, E., DiClemente, R. J., Anderson, M. M., and Lodico, M. (1996). Early drinking onset and its association with alcohol use and problem behavior in late adolescence. Preventive Medicine, 25, 293-300.

Guy, S. M., Smith, G. M., and Bentler, P. M. (1994). Consequences of adolescent drug use and personality factors on adult drug use. Journal of Drug Education, 24, 109-132.

Hagan, J. (1997). Defiance and despair: Subcultural and structural linkages between delinquency and despair in the life course. Social Forces, 76, 119-134.

Hamburg, B. A., and Dixon, S. L. (1992). Adolescent pregnancy and parenthood. In M. Rosenheim and Testa (Ed.), Early Parenthood and Coming of Age in the 1990s. New Brunswick, NJ: Rutgers University Press.

Hansen, W. B., Graham, J. W., Sobel, J. L., Shelton, D. R., Flay, B. R., and Johnson, C. A. (1987). The consistency of peer and parent influences on tobacco, alcohol, and marijuana use among young adolescents. Journal of Behavioral Medicine, 10, 559-579.

Hanson, T. L, McLanahan, S. S., and Thomson, E. (1996). Double jeopordy: Parental conflict and stepfamily outcomes for children. Journal of Marriage and the Family, 58, 141-154.

Harford, T. C., and Grant, B. F. (1994). Prevalence and population validity of DSM-III-R alcohol abuse and dependence: The 1989 National Longitudinal Survey on Youth. Journal of Substance Abuse, 6, 37-44.

Hausman, J., B. H. Hall, and Z. Griliches (1984). "Econometric Models for Count Data with an Application to the Patents-R&D Relationship." Econometrica, 52:4, 909-938.

Hawkins, J. D., Catalano, R. F., and Miller, J. Y. (1992). Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance abuse prevention. Psychological Bulletin, 112, 64-105.

Hawkins, J. D., Graham, J. W., Maguin, E., Abbott, R., Hill, K. G., and Catalano, R. F. (1997). Exploring the effects of age of alcohol use initiation and psychosocial risk factors on subsequent alcohol misuse. Journal of Studies on Alcohol, 58, 280-290.

Hayes, C. (1987). Risking the future: Adolescent sexuality, pregnancy and childbearing (Vol. 10). Washington, DC: National Academy Press.

Hetherington, E. M. (1989). Coping with family transitions: winners, losers, and survivors. Child Development, 60, 1-14.

Hetherington, E. M. (1993). An overview of the Virginia longitudinal study of divorce and remarriage with a focus on early adolescence. Journal of Family Psychology, 7, 39-56.

Hetherington, E. M. (1999). Social capital and the development of youth from nondivorced, divorced and remarried families. In W. A. Collins and B. Laursen (Eds.), Relationships as Developmental Contexts (pp. 177-209). Mahwah, NJ: Lawrence Erlbaum Associates.

Hetherington, E. M., and Clingempeel, G. (1992). Coping with marital transitions: A family systems perspective. Monographs of the Society for Research in Child Development, 57, (2-3, Serial No. 227).

Hill, M. S. and Duncan, G. (1987). Parental family income and the socioeconomic attainment of children. Social Science Research, 16, 39-73.

Hill, S. Y., and Yuan, H. (1999). Familial density of alcoholism and onset of adolescent drinking. Journal of Studies on Alcohol, 60, 7-17.

Hussong, A. M., and Chassin, L. (1997). Substance use initiation among adolescent children of alcoholics: Testing protective factors. Journal of Studies on Alcohol, 58, 272-279.

Hussong, A. M., Curran, P. J., and Chassin, L. (1998). Pathways of risk for accelerated heavy alcohol use among adolescent children of alcoholic parents. Journal of Abnormal Child Psychology, 26, 453-466.

Jessor, R., Donovan, J. E., and Costa, F. M. (1991). Beyond Adolescence: Problem Behavior and Young Adult Development. Cambridge: Cambridge University Press.

Jessor, R., and Jessor, S. (1977). Problem Behavior and Psychosocial Development: A Longitudinal Study of Youth. New York: Academic Press.

Johnson, E. O., Arria, A. M., Borges, G., Ialongo, N., and Anthony, J. C. (1995). The growth of conduct problem behaviors from middle childhood to early adolescence: Sex differences and the suspected influence of early alcohol use. Journal of Studies on Alcohol, 56, 661-671.

Johnson, R. A., Su, S. S., Gerstein, D. R., Shin, H., and Hoffmann, J. P. (1995). Parental influences on deviant behavior in early adolescence: A logistic response analysis of age- and sex- differentiated effects. Journal of Quantitative Criminology, 11, 167-193.

Judge, George G., Griffiths, William E., Hill, R. Carter, and Lee, Tsoung-chao (1985). The Theory and Practice of Econometrics, Second Edition, Wiley Series in Probabilities and Mathematical Statistics.

Kandel, D. B. (1996). The parental and peer contexts of adolescent deviance: An algebra of interpersonal influences. Journal of Drug Issues, 26, 289-315.

Keith, V. M., and Finlay, B. (1988). The impact of parental divorce on children's educational attainment, marital timing, and likelihood of divorce. Journal of Marriage and the Family, 50, 797-809.

Kilpatrick, D. G., Acierno, R., Saunders, B., Resnick, H. S., Best, C. L., and Schnurr, P. P. (2000). Risk factors for adolescent substance abuse and dependence: Data from a national sample. Journal of Consulting and Clinical Psychology, 68, 19-30.

Kosterman, R., Hawkins, J. D., Guo, J., Catalano, R. F., and Abbott, R. D. (2000). The dynamics of alcohol and marijuana initiation: Patterns and predictors of first use in adolescence. American Journal of Public Health, 90, 360-366.

Kratzer, L., and Hodgins, S. (1997). Adult outcomes of child conduct problems: A cohort study. Journal of Abnormal Child Psychology, 25, 65-81.

Krein, S. F. (1986). Growing up in a single parent family: The effect on education and earnings of young men. Family Relations, 35, 161-168.

Krohn, M. D., Lizotte, A. J., and Perez, C. M. (1997). The interrelationship between substance use and precocious transitions to adult statuses. Journal of Health and Social Behavior, 38, 87-103.

Lang, K., and Zagorsky, J. L. (2001). Does growing up with a parent absent really hurt? The Journal of Human Resources, 36, 253-273.

Levitt, S. D., and Lochner, L. (2000). The determinants of juvenile crime. In J. Gruber (Ed.), Risky Behavior Among Youth: An Economic Analysis. Chicago: University of Chicago Press.

Lindberg, L. D., Boggess, S., and Williams, S. (1999). Multiple Threats: The co-occurrence of teen health risk behaviors. Reported presented under contract from the Office of the Assistant Secretary for Planning and Evaluation (ASPE).

Liska, A. E., and Reed, M. D. (1985). Ties to conventional institutions and delinquency: Estimating reciprocal effects. American Sociological Review, 50, 547-560.

Loeber, R., and Dishion, T. (1983). Early predictors of male delinquency: A review. Psychological Bulletin, 94, 68-99.

Loeber, R., and Stouthamer-Loeber, M. (1987). Prediction. In H. C. Quay (Ed.), Handbook of Juvenile Delinquency (pp. 325-382). New York: Wiley.

Lynskey, M. T., and Fergusson, D. M. (1995). Childhood conduct problems, attention deficit behaviors, and adolescent alcohol, tobacco, and illicit drug use. Journal of Abnormal Child Psychology, 23, 281-302.

McGrath, C. E., Watson, A. L, and Chassin, L. (1999). Academic achievement in adolescent children of alcoholics. Journal of Studies on Alcohol, 60, 18-26.

McLanahan, S. (1999). Father absence and the welfare of children. In E. M. Hetherington (Ed.), Coping With Divorce, Single Parenting, and Remarriage (pp. 117-146). Mahwah, NJ: Lawrence Erlbaum Associates.

McLanahan, S., and Sandefur, G. (1994). Growing Up With A Single Parent: What Hurts, What Helps. Cambridge, MA: Harvard University Press.

McLeod, J. D. (1991). Childhood parental loss and adult depression. Journal of Health and Social Behavior, 32, 205-220.

Mensch, Barbara S., and Kandel, Denise B. (1988), Underreporting of Substance Use in a National Longitudinal Youth Cohort: Individual and Interviewer Effects, Public Opinion Quarterly 52:1, 100-124

Monk-Turner, E. (1989). Effects of high school delinquency on educational attainment and adult occupational status. Sociological Perspectives, 32, 413-418.

Moore, K. A., Miller, B. C., Sugland, B. W., Morrison, D. R., Glei, D. A., and Blumenthal, C. (1995). Beginning too soon: Adolescent sexual behavior, pregnancy and parenthood. A review of research and interventions [On-line]. Available: http://aspe.hhs.gov/hsp/cyp/xsteesex.htm.

Newcomb, M. D., and Bentler, P. M. (1988). Consequences of Adolescent Drug Use. Newbury Park, CA: Sage Publications.

Newcomb, M. D., and Bentler, P. M. (1993). Antecedents and consequences of cocaine use: An eight-year study from early adolescence to young adulthood. In L.N. Robins and M. Rutter (Eds.), Straight and Devious Pathways from Childhood to Adulthood (pp. 158-181). New York, NY: Cambridge University Press.

Nye, Ivan F. (1958). Family Relationships and Delinquent Behaviors. New York: John Wiley.

Nye, Ivan F. and Short, James F., Jr. (1957). Scaling Delinquent Behavior. American Sociological Review, 22:3, 326-331.

Office of Applied Studies, SAMHSA. (2000). Patterns of Alcohol Use Among Adolescents and Associations with Emotional and Behavioral Problems [On-line]. Working paper available: http://www.samhsa.gov/oas/oas.html.

Office of Applied Studies, SAMHSA. (2001). Risk and Protective Factors for Adolescent Drug Use: Findings from the 1997 National Household Survey on Drug Abuse. Rockville, MD: SAMHSA.

Offord, D. R., and Bennett, K. J. (1994). Conduct disorder long-term outcomes and intervention effectiveness. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 1069-1078.

Parker, D. A., and Harford, T. C. (1988). Alcohol-related problems, marital disruption and depressive symptoms among adult children of alcohol abusers in the United States. Journal of Studies on Alcohol, 49, 306-313.

Patterson, G. R. (1993). Orderly change in a stable world: The antisocial trait as a chimera. Journal of Consulting and Clinical Psychology, 61, 911-919.

Pedersen, W., and Skrondal, A. (1998). Alcohol consumption debut: Predictors and consequences. Journal of Studies on Alcohol, 59, 32-42.

Pergamit, Michael R. (1995). "Assessing School-to-Work Transitions in the United States," Statistical Journal of the United Nations Economic Commission for Europe, 12: 3and4, pp. 271-288.

Prescott, C. A., and Kendler, K. S. (1999). Age at first drink and risk for alcoholism: A noncausal association. Alcoholism: Clinical and Experimental Research, 23, 101-107.

Rankin, J. H. (1983). The family context of delinquency. Social Problems, 30, 466-479.

Resnick, M. D., Bearman, P. S., Blum, R. W., Bauman, K. E., Harris, K. M., Jones, J., Tabor, J., Beuhring, T., Sieving, R. E., Shew, M., Ireland, M., Bearinger, L. H., and Udry, J. R. (1997). Protecting adolescents from harm. Findings from the National Longitudinal Study on Adolescent Health. Journal of the American Medical Association, 278, 823-832.

Robins, L. N. (1978). Sturdy, childhood predictors of adult antisocial behavior: Replications from longitudinal studies. Psychological Medicine, 8, 611-622.

Rosenberg, Morris (1965). Society and the Adolescent Self-Image. Princeton, Princeton University Press.

Rosenheim, M. K. (1992). Teenage parenthood: Policies and perspectives. In M. K. Rosenheim and M. F. Testa (Eds.), Early Parenthood and Coming of Age in the 1990s (pp. 200-226). New Brunswick, NJ: Rutgers University Press.

Russell, M. (1990). Prevalence of alcoholism among children of alcoholics. In M. Windle and J. S. Searles (Eds.), Children of Alcoholics: Critical Perspectives (pp. 9-38). New York: Guilford.

Sampson, R. J., and Laub, J. H. (1990). Crime and deviance over the life course: The salience of adult social bonds. American Sociological Review, 55, 609-627.

Sampson, R. J., and Laub, J. H. (1993). Crime in the Making: Pathways and Turning Points Through Life. Cambridge, MA: Harvard University Press.

Scheier, L. M., Botvin, G. J., and Baker, E. (1997). Risk and protective factors as predictors of adolescent alcohol involvement and transitions in alcohol use: A prospective analysis. Journal of Studies on Alcohol, 58, 652-667.

Scheier, L. M., and Newcomb, M. D. (1991). Differentiation of early adolescent predictors of drug use versus abuse: A developmental risk-factor model. Journal of Substance Abuse, 3, 277-299.

Schulenberg, J., O'Malley, P. M., Bachman, J. G., Wadsworth, K. N., and Johnston, L. D. (1996). Getting drunk and growing up: Trajectories of frequent binge drinking during the transition to young adulthood. Journal of Studies on Alcohol, 57, 289-304.

Schulenberg, J., Wadsworth, K. N., O'Malley, P. M., Bachman, J. G., and Johnston, L. D. (1996). Adolescent risk factors for binge drinking during the transition to young adulthood: Variable- and pattern-centered approaches to change. Developmental Psychology, 32, 659-674.

Sewell, W. H. and Shah, V. P. (1967). Socioeconomic status, intelligence and the attainment of higher education. Sociology of Education, 40, 1-23.

STATA Reference Manual, Release 7 (2000). College Station, Texas, Stata Press.

Su, S. S., Hoffmann, J. P., Gerstein, D. R., and Johnson, R. A. (1997). The effect of home environment on adolescent substance use and depressive symptoms. Journal of Drug Issues, 4, 851-877.

Twaite, J. A. (1998). Children of Divorce. Northvale, NJ: Jason Aronson Inc.

Vitelli, R. (1997). Comparison of early and late start models of delinquency in adult offenders. International Journal of Offender Therapy and Comparative Criminology, 41, 351-357.

Wadsworth, M. E. J., and Maclean, M. (1986). Parents' divorce and children's life chances. Children and Youth Services Review, 8, 145-159.

Wallerstein, J. S. (1988). Children after divorce: wounds that don't heal. Perspectives in Psychiatric Care, 23, 107-113.

Wells, L. E., and Rankin, J. H. (1991). Families and delinquency: A meta-analysis of the impact of broken homes. Social Problems, 38, 71-93.

Weber, M. D., Graham, J. W., Hansen, W. B., Flay, B. R., and Johnson, C. A. (1989). Evidence for two paths of alcohol use onset in adolescents. Addictive Behavior, 14, 399-408.

Wilkinson, K. (1980). The broken home and delinquent behavior: An alternative interpretation of contradictory findings. In T. Hirschi and M. Gottfredson (Eds.), Understanding Crime: Current Theory and Research. Beverly Hills, CA: Sage.

Windle, M. (1990). A longitudinal study of antisocial behaviors in early adolescence as predictors of late adolescent substance use: Sex and ethnic group differences. Journal of Abnormal Psychology, 99, 86-91.

Wojtkiewicz, R. A. (1993). Simplicity and Complexity in the Effects of Parental Structure on High School Graduation. Demography, 30, 701-717.

Yamaguchi, K., and Kandel, D. B. (1984). Patterns of drug use from adolescence to young adulthood: II. Sequences of progression. American Journal of Public Health, 74, 668-672.

Zucker, R. A. and Harford, T. C. (1983). A national study of adolescent drinking practices in 1980: Demography. Journal of Studies on Alcohol, 44, 974-985

Topics

Prevention | Family Well-Being

Populations

Youth

The Long Term Impact of Adolescent Risky Behaviors and Family Environment

Executive Summary

Statement of the Problem

Data

Main Findings

Acknowledgements

Chapter I: Statement of the Problem

Chapter II: Literature Review

A. Predicting Adolescent Risky Behaviors

B. Adolescent Consequences of Risky Behaviors

C. Adult Consequences of Adolescent Risky Behaviors

D. Family Structure

E. Socioeconomic Status

F. Parental Alcoholism

Endnotes

Chapter III: Data and Methods

A. Data

B. Methods

1. Variable Creation

a. Adult Outcomes

b. Adolescent Risky Behaviors

c. Family Environment

2. Estimation

Endnotes

Chapter IV: The Relationship Between Adolescent Risky Behaviors and Family Environment

Chapter V: The Relationship of Adolescent Risky Behaviors and Family Environment to Adult Outcomes

A. Distributions

B. Regression Analysis

1. Introduction

2. The Relationship of Adolescent Risky Behaviors to Adult Outcomes

3. The Relationship of Adolescent Family Structure to Adult Outcomes

4. The Relationship of Parents' Education (SES) to Adult Outcomes

5. The Relationship of Parental Alcohol Problems to Adult Outcomes

Endnotes

Chapter VI: Exploring Pathways of Early Initiators

A. Introduction

B. Does Adolescent Family Structure Affect the Pathways of Early Initiators?

C. Does Parents' Education Affect the Pathways of Early Initiators?

Endnotes

Chapter VII: Discussion and Future Directions

Appendices

Appendix A: Variables and Estimation Methods

Description of Variables

Adult Outcome Variables

Youth Risky Behavior Variables

Family Environment Variables

Other Explanatory Variables

Description of Multivariate Regression Estimates

OLS Regression

Logistic Regression

Multinomial Logistic Regression

Negative Binomial Regression

Weibull Regression

Statistical Significance of Estimates

Endnote

Appendix B: Regressions of Adult Outcomes on Adolescent Risky Behaviors, Family Environment, and Other Independent Variables

Appendix C: New Categories of Risky Behavior Variables for Regressions of Different Adult Outcomes Among Early Initiators

Appendix D: Regressions of Adult Outcomes on Adolescent Risky Behaviors and Other Independent Variables Among Early Initiators: Effects of Family Structure

Appendix E: Regressions of Adult Outcomes on Adolescent Risky Behaviors and Other Independent Variables Among Early Initiators Living With Both Biological Parents at Age 14: Effects of Parents' Education

Appendix F: Regressions of Adult Outcomes on Adolescent Risky Behaviors and Other Independent Variables Among Early Initiators Living with Single Mothers at Age 14: Effects of Mothers' Education

References

Connect with Us