Cohort Profile: The Woodlawn Study

Abstract

The Woodlawn Study is an epidemiologically- defined community cohort study of 1242 Black Americans (51% female and 49% male), who were in first grade in 1966–67 in Woodlawn, a neighborhood of Chicago, Illinois. The study comprises extensive interview data over the life course including self-, mother-, and/or teacher-reported assessments at ages 6, 16, 32, 42, and 62 (in progress), administrative records (i.e., education, crime, and death records), and census data. These data cover a wide range of focal areas across the life course, including family environment, socioeconomic indicators, education, social integration (e.g., marriage, community engagement, religious involvement) and social support, employment, racial discrimination, substance use, crime/victimization, and mental and physical health, including mortality. Over the past 50 years, Woodlawn research has mapped cumulative disadvantage, substance use, and criminal offending and has identified key risk and protective factors of adversity, resilience, and success across the full life course. In turn, these findings have informed life course theory and policy for a population that experiences significant criminal and health disparities.

Why Was the Cohort Set Up?

The Woodlawn Study is a landmark longitudinal study of a community cohort of Black American youth (N = 1242) who attended first grade in 1966–67 in Woodlawn, a neighborhood in Southside Chicago, and have been assessed across multiple life stages over the past 50 years. The Woodlawn community is bordered by the University of Chicago to the north, Jackson Park and Lake Michigan to the east, and Interstate 90 to the west. Between 1950 and 1966, the Woodlawn community had undergone significant economic and social change (see Kellam et al., 1975). For instance, in 1950, Woodlawn was characterized as a diverse, middle-class neighborhood with 40% Black residents and 5% on public assistance. As a result of white flight, the great migration, and real estate practices, such as redlining, by 1966, Woodlawn was a predominantly Black American community (98%) where 21% of the residents received public assistance, a number that was three times higher than for the city of Chicago (7%, Kellam et al., 1975). By the 1960s, Woodlawn was under-resourced with few businesses in the community, dilapidated housing, and overcrowded schools (Chicago Historical Society, 2005). Despite these conditions, Woodlawn’s residents were economically heterogeneous with middle- and working-class families all living together in the community (e.g., homeownership within the 2.2 square miles of Woodlawn ranged from 40% in some blocks to 1% in others, Kellam et al., 1975).

The Woodlawn Study follows the first-grade cohort who participated in a classroom-based intervention program and evaluation study in 1966–67. The program sought to improve the psychological well-being, social adaptation, and long-term success of the community’s youth and was implemented during the 1964–65 through the 1967–68 school years in the Woodlawn first grade classrooms (see Schiff, 1972). The intervention program and evaluative assessments were implemented through the Woodlawn Mental Health Center (WMHC), which was established in 1963 by three white male psychiatrists, Drs. Sheppard Kellam, Sheldon Schiff, and Edward Futterman, with support from the Chicago Board of Health and the Illinois Department of Mental Health (Kellam et al., 1975). These men co-directed the WMHC and later held faculty positions in the Department of Psychiatry at the University of Chicago.¹

In response to challenges in garnering sustained community support during the planning phase of the WMHC and to confront general community distrust, the directors of the WMHC and community leaders assembled a community advisory board (CAB) for the center. This CAB served as a liaison between the WMHC and the community and as a sanctioning and policy-making entity (Kellam & Schiff, 1966). The CAB was integral in the planning and development of the WMHC and in setting its programming priorities (Schiff, 1970). The first mental health priority set forth by the CAB was to provide mental health services to Woodlawn’s young children. First graders were chosen as the target population of the intervention program because they represented a total population (i.e., attendance in first grade was compulsory), and this age group allowed for a whole systems approach that included parents, teachers, and clinicians (Kellam et al., 1972).

What is now known as the Woodlawn Study comprises the longitudinal data from the third (1966–67) of the four cohorts who attended first grade in the 12 elementary schools of Woodlawn and participated in the community-wide classroom-based intervention and evaluation. The first grade intervention program was fully developed by the 1966–67 school year, and the first-grade assessments included data from all of the children’s mothers or mother surrogates.² Although ethical review boards had yet to be established at this time, the directors of the WMHC and the CAB took care to ensure ethical practices in involving families, agreed that all data would be treated as medically confidential, and decided that all reports detailing the assessment findings would be reviewed by the CAB prior to dissemination (Schiff, 1970).

In 1975, Dr. Sheppard Kellam led a follow-up study of the 1966–67 cohort, with the assistance of Dr. Kellam’s research assistant, Margaret Ensminger, that located the mothers and cohort members for an adolescent assessment. This data collection effort was funded by the National Institute on Drug Abuse (NIDA) to identify key risk and protective factors of adolescent substance use to inform prevention and intervention efforts (see Kellam et al., 1982b). NIDA also funded the second and third follow-ups of the cohort, at ages 32 (1992–93) and 42 (2002–03), which were led by Dr. Margaret Ensminger while a faculty member at the Johns Hopkins Bloomberg School of Public Health in Baltimore, Maryland. These follow-ups focused on the long-term influence of childhood and adolescent risk and protective factors on substance use and crime throughout adulthood. Data collection of the cohort’s criminal history records that extend to age 52 was funded in 2012 by the Harry Frank Guggenheim Foundation and led by Dr. Elaine Doherty who joined the Woodlawn research team in 2005. The age 62 data collection effort that is currently underway, which focuses on health and cognition outcomes, is funded by the National Institute on Aging (NIA) and is led by Dr. Kerry Green at the University of Maryland School of Public Health. Dr. Green was a research assistant to Dr. Margaret Ensminger from 2000 to 2004 and worked on the midlife (age 42) data collection effort.³

In addition to these funding sources for each of the five assessments and administrative record retrieval, several analytic projects from the Woodlawn Study have been funded by NIH Institutes, including NIDA, the National Institute on Alcohol Abuse and Alcoholism (NIAAA), the National Institute on Mental Health (NIMH), and the National Institute on Child Health and Human Development (NICHD), as well as by internal and external sources, such as the University of Maryland Population Research Center and the Columbia Center for Homelessness Prevention Studies.

Who Is in the Cohort?

The Woodlawn Study comprises prospectively-collected data from the 1242 children who were first graders in the nine public and three parochial schools of Woodlawn during the 1966–67 school year (51% female [n = 636] and 49% male [n = 606]). Given the strong community support of the WMHC and its investment in the mental health needs of the community’s children, this cohort represents almost the entire population of first graders in the Woodlawn community at that time (99%, 13 families declined to participate). In 1966, 42% of the mothers reported having 12 or more years of education, close to half of the families were living above the poverty level (46%), and most of the children lived with two adults (56%; 42% in mother-father households and 14% in mother-other adult households).

How Often Have They Been Followed Up?

The original 1242 first graders (modal age 6) were followed up as adolescents (1975–76; modal age 16), young adults (1992–93; modal age 32), in midlife (2002–03; modal age 42), and currently as older adults (2021–23; modal age 62; see Fig. 1). During the initial assessment, teachers reported on each child’s classroom behavior; WMHC clinicians recorded observations of the children in standardized play situations; and mothers (or mother surrogates) were interviewed about their first-grade child’s social adaptation, well-being, family characteristics (e.g., income, family structure, residential mobility, social integration, mother’s education), and family child-rearing practices (e.g., discipline, affection).

Fig. 1.

Data sources and timeline

Despite the lack of formal ethical boards at the time, several ethical protocols were in place for the mother interviews in 1966. First, a letter from their child’s school principal was sent to each child’s address in advance of an interviewer contacting the mother at her home. Second, before beginning the interview, the respondent was given the opportunity to ask any questions regarding the interview, and a refusal was noted with no further action. Third, all interviews were conducted in private, away from the child, and respondents were assured that their answers would be kept confidential.

Mothers (or mother surrogates) were again interviewed in 1975–76 about the family and the study child during adolescence (n = 939) using a structured questionnaire. Mothers provided written consent for their own interview and written consent for their child’s assessment. The teenagers who provided written assent (n = 705) were assessed in groups about their psychological well-being, alcohol and drug use, delinquency, family relationships, and participation in school, church, and other activities using questionnaires presented on slides and audio tape to control for reading differences.

In 1992–93, the adult children were re-interviewed using a structured questionnaire at ages 32–33 (n = 952). They were assessed on living arrangements, family relationships, education and employment histories, their economic situation, health, social support, participation in church and other associations, psychiatric symptoms, alcohol and drug abuse/dependence, and criminal activities. In 2002–03 the research team re-interviewed the cohort at ages 42–43 (n = 833) for the midlife interview, using an assessment very similar to the one used in 1992–93. Interviews were conducted in prisons and jails with willing incarcerated participants at both adult assessments. Data collection efforts are currently underway to interview cohort members at age 62. This older adult assessment includes domains previously assessed, as well as domains most relevant to this age group, such as successful aging and cognitive health. Over the years, data have been collected from the schools, the Chicago Board of Education, the National Death Index, the Chicago Police Department, the Illinois Criminal Justice Authority, and the Federal Bureau of Investigations (FBI). In 1998, the mothers of the cohort members were re-interviewed when the cohort children were 38 years old (n = 680 mothers), with the focus on the mother’s own well-being.

It is important to note that prior to each assessment, including the data collection of criminal justice records in 2012, the principal investigators convened the community advisory board (CAB) to provide guidance on the research approach, data collection procedures, and interview content. Whereas the advisory board has changed its membership over the years, it has always comprised Black Americans who were roughly the same age as the cohort members and have typically included prior participants, past residents of Woodlawn, and current community leaders in either Woodlawn or Chicago.

Selection of Adult Characteristics

Whereas the full cohort lived in Woodlawn at age 6, by young adulthood only 9% remained in Woodlawn, another 65% lived within the Chicago city limits, and 26% lived outside of Chicago. By age 42, a larger percentage of the cohort had moved outside of Chicago (46%), leaving 7% in Woodlawn and another 47% within the Chicago city limits. The cohort improved their financial situation over the years. Whereas just over half of the families were living below the federal poverty level in 1966 (54%), this percentage had been cut in half by age 42 with closer to one-quarter living in poverty (26%).

Table 1 provides a snapshot of the lives of the Woodlawn cohort in adulthood with respect to social integration and community involvement. By age 42, over 80% had earned a high school diploma or GED, 69% had ever been married by age 42 with 35% currently married and another 7% living with a partner, 74% were employed in the past week (full-time, part-time, or self-employed), 46% had children living in the household, and close to half attended church or religious services on a weekly basis (41%). The prevalence of social role engagement among the Woodlawn cohort is largely comparable with those found in national estimates of Black Americans in both young adulthood and midlife with the exception of marriage, which is substantially lower than national estimates of Black Americans (28% vs. 50% married in young adulthood; 35% vs. 45% married in midlife, respectively; Ensminger et al., 2016).⁴

Table 1.

Social integration indicators in adulthood

	Young adult (n = 952)	Midlife (n = 833)
High school graduate or GED	79.5%	81.6%
Currently married	27.9%	34.5%
Currently employed	62.6%	73.7%
Children in household (< 18)	55.4%	46.0%
Weekly religious service attendance	28.6%	40.9%

Cohort Retention

Like all prospective studies, the Woodlawn Study has experienced attrition over the years due to death,⁵ incapacitation, an inability to locate participants, and refusals (see Table 2). The lowest response rate occurred during the adolescent data collection effort, in part due to funding limitations that restricted recruitment to those living in the Chicago area. Also, although 939 of the mothers or mother surrogates were re-interviewed in 1975–76 (representing 77% of the living adolescents), only 705 adolescent cohort members were available and assented to the in-person group-based assessment (57%). In 1992 and 2002, researchers began to recruit all of the original cohort members who were still living, including cohort members who were in jail or prison (36 in 1992–93 and 18 in 2002–03), resulting in relatively high response rates (80% and 72%, respectively). In full, 85% of the original cohort were interviewed during at least one of the age 32 or age 42 interviews.

Table 2.

Interview disposition status

	First grade 1966–67	Adolescence 1975–76	Young adulthood 1992–93	Mid-adulthood 2002–03	n (%)
All four assessments completed	✓	✓	✓	✓	474 (38.2%)
Three assessments completed	✓	✓	✓		125 (10.1%)
	✓	✓		✓	38 (3.1%)
	✓		✓	✓	257 (20.7%)
Two assessments completed	✓	✓			68 (5.5%)
	✓		✓		96 (7.7%)
	✓			✓	64 (5.2%)
One assessment completed	✓				120 (9.7%)
Assessment N	1242	705	952	833	1242 (100.0%)
Response ratea	n/a	57.0%	80.1%	72.4%
Unlocatable/unavailable/refused	n/a	531	237	318
Incapacitated/deceased	n/a	6	53	91

^aResponse rate = number of interviewed participants / (total population – (deceased + incapacitated))

Several Woodlawn publications report attrition analyses across the multiple time points (e.g., Crum et al., 2006; Green et al., 2010b; Ensminger et al., 2007; Doherty et al., 2008a). Comparisons of first-grade characteristics between those with and without an adolescent assessment show no differences in key demographics, such as sex, mother’s education, poverty status, welfare receipt, family income, or family type. Those missing in adolescence were less likely to have an adult interview, yet there were no differences in having an official criminal record or adult substance use or abuse/dependence for those with and without an adolescent interview. Comparisons between those with and without at least one adult interview reveal no significant differences on key variables, such as sex, early socioeconomic status (e.g., mother’s education, welfare participation during childhood), adolescent drug use (e.g., alcohol, cigarettes, cannabis, cocaine), early childhood social adaptation (e.g., aggression and shyness), adolescent problem behavior (e.g., self-reported suspension, fighting with parents, carrying a weapon), adolescent self-reported delinquency (e.g., violent crime, property), or criminal career characteristics (e.g., age of first arrest). Interestingly, cohort members with a criminal record were more likely to have an adult interview than not, perhaps due to the availability of administrative records for locating purposes. Those interviewed in adulthood were also more likely to have graduated from high school and less likely to be in poverty in first grade or adolescence. Overall, the interviewed respondents in each assessment are largely comparable to the original cohort on a range of key domains, although those followed into adulthood may represent a more educated and resourced sample.

What Has Been Measured?

Since its inception, the Woodlawn Study has been guided by a life-course framework to formulate research questions, develop measures, and contextualize findings. Over time, some combination of mothers, teachers, and/or the study participants themselves have described family, school (first grade and adolescence), work (ages 32 and 42), economic circumstances, stressful events, community characteristics, psychological and physical well-being, and risk behaviors as displayed in Table 3 (see Appendix Table 6 for detailed descriptions).

Table 3.

Key domains measured at each interview across time

	Childhood (first grade)	Adolescence (age 16)	Young adult hood (age 32)	Midlife (age 42)
Childhood and adolescence
School adaptation	✓	✓
Childhood through midlife
Socioeconomic resources	✓	✓	✓	✓
Family relationships	✓	✓	✓	✓
Religiosity	✓	✓	✓	✓
Neighborhood	✓	✓	✓	✓
Physical health	✓	✓	✓	✓
Mental health	✓	✓	✓	✓
Adolescence through midlife
Friendships		✓	✓	✓
Social support		✓	✓	✓
Community engagement		✓	✓	✓
Substance use			✓	✓
Criminal offending		✓	✓	✓
Young adulthood and midlife
Victimization experiences			✓	✓
Racial identity and discrimination			✓	✓

Selection of measures for the young adult and midlife follow-ups was guided by several principles beyond the conceptual framework of the study. First, throughout the study, Woodlawn instruments have included similar measures in each assessment to maintain continuity as much as possible, as well as incorporated newer, improved measures developed since the last data collection effort. Second, Woodlawn instruments have included standard measures used in other studies, whenever possible, to aid in the comparison of findings with other populations. Third, during instrument development for each assessment, researchers solicited feedback from the CAB regarding measures that were relevant to Black Americans at that time (e.g., civil rights activism, religiosity). Measures of racial identity and racial discrimination were added for the young adult assessment but were not assessed extensively until the midlife assessment. Finally, as much as possible, researchers have selected measures that have been validated for Black American populations in all assessments, such as well-being in childhood and adolescence (Conners, 1970) and racial discrimination (McNeilly et al., 1996) and racial identity (Sellers et al., 1998) in young and mid-adulthood. The instruments for each of the data collection efforts are available on the Woodlawn website (www.woodlawn.umd.edu).

The interview information is supplemented by several sources of administrative data (see Fig. 1). Census data are available for 1970, 1980, 1990, and 2000, which roughly coincided with the first grade, adolescent, young adult, and midlife interviews, respectively. These data provide key indicators of structural racism, such as racial segregation and inequities in education, employment, and home ownership. Board of Education data regarding school testing and high school degree information are available through 1982. Annualized official criminal involvement information from the Chicago Police Department and FBI is available for ages 17 to 32 and from the Illinois Criminal Justice Authority for ages 17 to 52 for 98% of the cohort (n = 1217). These data include the date, type of crime (up to 3 unique charges), type and method of disposition, and sentence for each arrest. Finally, mortality information from the National Death Index extends to 2021 (modal age 61) and includes the date and cause of death.

What Has It Found? Key Findings and Publications

Since the 1960s, Woodlawn Study researchers have produced several publications contributing to science. Before detailing these findings, we would like to take this opportunity to situate these publications in their historical and social context. First, whereas the Woodlawn Study was started by three white men, the research team over the years has included faculty and students representing a range of racial and ethnic backgrounds, including white, Black, Asian, Hispanic, and Latinx Americans. However, the principal investigators throughout the study’s tenure have been white men and women in senior faculty positions. Second, a major strength of the Woodlawn Study is its ability to provide rigorous longitudinal research that can document the heterogeneity in Black American lives. Research stemming from data provided by the Woodlawn cohort has dispelled some racial stereotypes of family composition, substance use, and criminal activity among Black Americans. However, over the past 30 years, the NIDA-funded focus on the prevalence and interrelationships of substance use and criminal involvement and the moderating influence of marriage may have also perpetuated some stereotypes, contributed to a continued focus in the research community on problem behaviors in Black communities, and unintentionally reinforced the importance of heteronormative institutions. Moreover, many of the Woodlawn publications on social institutions that are rooted in systemic racism, such as education and the criminal legal system, do not directly confront this historical and social context. Third, many of the Woodlawn publications do not adhere to current best practices with respect to adopting an anti-racist lens or using person-first language, which is important to bear in mind when assessing each research study’s content and presentation of its findings. Current and future research efforts using the Woodlawn data have made and will continue to make a concerted effort to focus on the strengths and successes of the cohort, expand the domains of potential buffers of risk beyond marriage and education, and draw on best practices for conducting and presenting the research. For instance, the focus of the current data collection effort at age 62 is on identifying the key features of successful aging, broadly defined, among the cohort and on predictors of those successes, including non-marital ties, which may be more relevant for the cohort, and perceptions of well-being rather than objective indicators.

Against the backdrop of these caveats, the Woodlawn Study has made significant contributions to science. We focus our discussion here on the findings most relevant to life course criminology, including patterns and interrelationships among substance use, crime, and criminal legal system involvement, early life risk and protective factors, and pathways to successful adult outcomes. Taken together, research using the Woodlawn data has produced an important body of work that highlights how life transitions (1) produce key turning points stemming from social integration and educational attainment, (2) interrupt substance use continuity, and (3) reduce criminal legal system interactions. In turn, these findings inform policy and intervention work for Black Americans. We refer the reader to the Woodlawn website www.woodlawn.umd.edu for a comprehensive list of Woodlawn publications.

Prevalence, Patterns, and Interrelationships of Substance Use, Crime, and Criminal Legal System Involvement over the Life Course

In this section, we describe the key findings regarding the prevalence, patterns, and interrelationships of criminal offending and substance use throughout the life course.

Prevalence and Patterns of Criminal Offending

Doherty and Ensminger (2014) and Doherty et al. (2016) provide detailed results regarding the criminal career prevalence and patterns of criminal offending. These studies show that approximately half of the Woodlawn cohort had been arrested by age 52 (54%) and 21% self-reported incarceration by age 42, with sex-specific rates showing that men were more likely to be arrested at least once by age 52 than women (65% and 29%, respectively), to self-report incarceration by age 32 (29% and 7%, respectively), and to self-report incarceration between the ages of 32 and 42 (21% and 6%, respectively). Yet, despite these relatively high prevalence rates and contrary to media portrayals of Black American men as life-course persistent offenders or “career” criminals, group-based trajectory analyses of annualized official arrests from ages 17 to 52 reveal that the majority of the cohort is categorized as either non-offenders or low-rate desisters (76% of men and 96% of the women; Doherty & Ensminger, 2014). Table 4 shows a clear decrease in the prevalence of self-reported violent and property offending for the Woodlawn cohort, with self-reported violence dropping from approximately 68% in adolescence to 46% in young adulthood and 15% in midlife. The corresponding property offending prevalence estimates for adolescence, young adulthood, and midlife are 69%, 38%, and 15%. As reported in Doherty and Green (2022), we compared the official arrest data with the self-reported information on 27 different violent and non-violent offenses from young adulthood (ages 17 to 32) to understand their interrelationship. The correlation between the number of arrests and number of different self-reported offenses is 0.359 (p < 0.001). Comparisons of identified offenders based on official arrest data and self-report data show agreement for 60% of the sample (i.e., self-reported no offenses and not arrested (30%); self-reported offending and were arrested (30%)). Potential criminal legal system bias or under-reporting occurred for 13% of the sample (i.e., arrested in young adulthood but no self-reported offending) and 27% represent those who were not arrested but self-reported offending in young adulthood (i.e., the dark figure of crime)

Table 4.

Prevalence of criminal offending, self-reported and official records

	Adolescence (13 to 16)	Young adulthood (17 to 32)		Midlife (33 to 42/52)
	Self-report (n = 705)	Self-report (n = 952)	Official arrest (n = 1217)	Self-report (33 to 42, n = 833)	Official arrest (33 to 52, n = 1176)
Violent offender	67.7%	45.7%	23.4%	15.0%	15.7%
Property offender	68.5%	37.9%	29.1%	14.7%	16.9%
Incarcerated	n/a	17.6%	n/a	12.2%	n/a

Prevalence and Patterns of Substance Use

As shown in Table 5, while cannabis was a commonly used drug during adolescence and young adulthood, this prevalence dropped to close to a quarter by midlife. While about one quarter used cocaine at least once by young adulthood, heroin use was less common with about 5% of cohort members using at various adult life stages.

Table 5.

Prevalence of drug use and disorders, self-reported

	Adolescence lifetime use by age 16 (n = 703)	Young adulthood lifetime use/ disorders by age 32 (n = 952)	Midlife past 10 year use/ disorders, ages 33–42 (n = 833)
Cannabisa	61.1%	57.5%	27.2%
Cocaine	7.3%	25.1%	17.2%
Heroin	< 1%	5.1%	5.0%
Drug use disorder	n/a	13.0%	12.5%

^aSee Ensminger et al. (2007) for a discussion of the inconsistencies in cannabis use reporting between adolescence and young adulthood and models examining “recanting” over time

Despite the relatively high prevalence of lifetime cannabis use, for many, the frequency of use is low. Trajectory analyses of use over the life course from ages 7 to 32 reveal that abstinence or low-level use is the norm with 49% of men and 65% of women identified as lifetime abstainers from/low-level users of cannabis (Juon et al., 2011) and analyses of use from ages 7 to 42 reveal that 71% of men and 80% of women are identified as never or low-level cocaine users (Ensminger et al., 2016). Findings of cocaine, in particular, align with the tenets of a crossover effect with a somewhat later onset of substance use but a longer course of use and worse consequences from use in Black American populations compared to their White counterparts from national samples (Doherty et al., 2008b; Ensminger et al., 1997). The prevalence of drug use disorders was 13% at both young adulthood and midlife.

Interrelationships

Research from the Woodlawn Study has demonstrated the strong interrelationship between substance use, crime, and criminal legal system involvement among the cohort. Across multiple studies, we find evidence of a long-term effect of serious adolescent delinquency on cannabis and cocaine initiation (Doherty et al., 2008a), of frequent adolescent drinking on adult violence (Green et al., 2011), of heavy adolescent cannabis use on self-reported offending, arrests, and incarceration (Green et al., 2010b; Holder et al., 2022), and of drug use duration on criminal arrest and incarceration (Green et al., 2020). Doherty et al. (2016) found a criminogenic effect of criminal arrest, with an arrest in young adulthood predicting close to a three-fold increase in midlife drug use and drug use disorders and an alarming 8- to 17-fold increase in future arrest, controlling for self-reported crime, indicating a secondary sanctioning effect. This study also found a protective effect of incarceration on substance use, potentially through increased opportunities for treatment or fewer opportunities to use.

Across the full life course, we find that reciprocal effects between drug use and criminal involvement are evident for men beginning in adolescence and for women beginning in young adulthood with early behavioral problems setting both boys and girls on a long-term trajectory of drug use and crime (Green et al., 2019). Studies also identify the interrelationships between violence, substance use, and mental health (McCord & Ensminger, 1997; Green et al. 2012b, c) and the role of victimization in substance use continuity into midlife (Doherty et al., 2012, 2018). Taken together, findings confirm the substance use and crime interrelationship found in previous studies and suggest that the criminal legal system may be implicated in perpetuating substance use and criminal legal system involvement among this cohort, above and beyond criminal behavior.

Long-Term Reach of Childhood Structural Conditions and Individual Social Adaptation

The Woodlawn Study has informed our understanding of the influence of a variety of proximal life experiences across the life course on criminal and substance use outcomes, such as the relationship between racial discrimination experiences on violent offending by young adulthood (e.g., McCord & Ensminger, 2003) and the protective role of positive ethnic-racial socialization on adolescent and young adult outcomes (Gaston & Doherty, 2018). In this section, we focus on findings regarding the long reach of childhood’s impact on midlife outcomes, as the length of the Woodlawn Study is one of its greatest strengths. These results also highlight a second major strength of the Woodlawn Study, which is its ability to consider within-race heterogeneity in one’s early socio-structural circumstances.

Multiple Woodlawn studies have found that several indicators of early structural adversity (e.g., family structure, residential mobility, household and neighborhood poverty) are related to a higher risk of various outcomes, including drug use, crime, social isolation, educational attainment, homelessness, and poor psychological health (Bowie et al., 2006; Fleming et al., 1982; Fothergill & Ensminger, 2006; Fothergill et al., 2009, 2012; Green et al., 2013; Juon & Ensminger, 1997; Juon et al., 2003; Lo, 2010; Williams et al., 2004). Moreover, these studies highlight the complexity of these socio-structural factors. For instance, early Woodlawn studies challenged existing stereotypes at the time about the detriments of father-absent families in Black American communities, finding that the cohort children were living in 86 different family structures (Kellam et al., 1982a, b, c) and that children from mother-grandmother families did not differ from children in mother-father families on key outcomes, such as school achievement (Kellam et al., 1977). These findings highlight that it is not the absence of a father per se that is detrimental but rather the absence of a second adult in the family.

Another consistent finding has been the importance of early school adaptation and attainment for later life outcomes. For instance, Woodlawn studies have found that teacher-rated first-grade aggressive behavior is a sturdy predictor of substance use and serious adult offending, substance use treatment, educational achievement, and school dropout (Crum et al., 1998, 2006; Ensminger et al., 1983, 1996, 2002; Evans-Polce et al., 2014; Fothergill et al., 2008; Juon et al., 2002, 2006; Kellam et al., 1982b). Interestingly, Woodlawn studies have shown that while shy behavior is protective for adolescent and adult substance use (Ensminger et al., 2002), the combination of shyness and aggression for males is especially problematic for these outcomes (Ensminger et al., 1982, 1983, 2002; Kellam et al., 1982b).

Short- and Long-Term Social and Health Consequences of Substance Use, Arrest, and Incarceration

We have published extensively on the consequences of substance use and criminal involvement in contribution to this body of knowledge (e.g., Doherty et al., 2022; Fothergill & Ensminger, 2006; Fothergill et al., 2009, 2016; Green et al., 2017). For instance, Woodlawn research has found that early, frequent, and/or persistent cannabis users were significantly more likely than later, less frequent, or nonusers to be unemployed and unmarried in young adulthood, to have had children outside of marriage (Green & Ensminger, 2006), to have dropped out of school (Ensminger et al., 1996), to live below the federal poverty line in midlife (Stuart & Green, 2008), and to engage in HIV risk behavior (Juon et al., 2011). Men with an arrest history were significantly more likely to engage in health-risk behaviors, such as smoking, binge drinking, and illegal drug use at age 42, than those without an arrest history (Doherty et al., 2022).

Woodlawn studies on premature mortality have identified being male, being raised by a single mother, being in foster care, and engaging in aggressive behavior in first grade as predictive of death in young adulthood and midlife (Juon et al., 2003, 2014). More recent studies have identified experiencing persistent poverty across the life course as a key predictor of mortality into the 40s and 50s, with higher education and social role participation moderating this impact (Green et al., 2022), and being arrested for violent offenses by age 32 as placing someone at heightened risk of later mortality to age 58 compared with non-violent only offenders and non-offenders. This increased risk of mortality among violent offenders is largely due to their extensive involvement with the criminal legal system, primarily for non-violent arrests (Doherty & Green, 2022).

Protective Factors and Success in Adulthood

While adversity and problem behaviors among Black Americans have often been a focus of study, an important contribution of the Woodlawn Study has resulted from its efforts to advance the field’s understanding of protective factors and life course success. Social integration (e.g., marriage, community engagement, religious involvement) and educational success are two dimensions that have been consistently associated with positive adult outcomes.

Social Integration

Despite low rates of marriage among this population (Green et al., 2012a), men who marry have a substantial (28%) reduction in criminal arrests when they are married compared to when they are unmarried (Doherty & Ensminger, 2013) with offending increasing upon divorce among those living in structurally disadvantaged neighborhoods (Doherty & Bersani, 2016). The protective effect of marriage is less clear among women (Doherty & Ensminger, 2013) with more economically-disadvantaged women reaping greater benefits from marriage in the form of a reduced probability of offending than their more advantaged counterparts (DiPietro et al., 2018). Importantly, the low rates of marriage among the Woodlawn cohort have directed our research efforts to identifying non-marital experiences that can lead to success in adulthood. For instance, social integration, defined as engaging in a higher number of adult socializing life events (i.e., marriage, parenthood, employment, and high school graduation), is related to desistance from offending (Doherty & Cwick, 2016), and religion and close family ties are protective against smoking, drug use, and drug use problems in both men and women (Bowie et al., 2006; Fothergill et al., 2009; Green et al., 2010a; Juon et al., 2002).

Education

Woodlawn studies also highlight the importance of educational factors in protecting against drug and alcohol use and disorders (Crum et al., 1998, 2006; Fothergill et al., 2008) and criminal outcomes. For instance, Doherty and Bersani (2020) explored the individual, familial, and contextual factors that distinguished men identified as at high risk of criminal legal system involvement but who were not arrested from similarly-situated men who were arrested. Findings reveal that graduating from high school on-time and having a sibling who graduated from high school emerged as particularly strong protective factors, pointing to the importance of promoting education for the whole family. Highlighting the long-term effects of education, Strong et al. (2016) find education mediates the relationship between adolescent smoking and adult substance use, and Green et al. (2022) report on education buffering the impact of early adversity on premature mortality.

Woodlawn researchers and students have produced more than 100 publications utilizing these longitudinal data, which now captures prospectively-collected snapshots of the lives of the Woodlawn cohort that spans over 50 years. These studies have mapped cumulative disadvantage, substance use, and criminal offending and have identified key risk and protective factors of adversity, resilience, and success across the full life course. In turn, these findings have informed life course theory and policy for a population that experiences significant criminal and health disparities.

What Are the Main Strengths and Weaknesses?

One key strength of the Woodlawn Study is that it is an epidemiologically-defined community-based population from one racial group who experienced diverse economic backgrounds early in life. This characteristic of the data allows for an examination of social and economic factors within one racial group and supports investigations into the natural heterogeneity of life course pathways among Black Americans. Moreover, because essentially all Woodlawn first-grade children and their families participated in the study, there is little sample selection bias in this community cohort. Comparisons with national estimates suggest that this cohort is comparable to nationally representative samples of Black Americans with respect to key factors, such as engagement in most social roles, substance use, and criminal arrest (e.g., Anthony et al., 1994; Doherty & Ensminger, 2014; Ensminger et al., 2016). A second considerable strength of this study is the extent of the early information from multiple sources (e.g., cohort members, their mothers and teachers, and clinicians), as well as contextual data (i.e., census data) and administrative records (i.e., school, criminal justice, and death records), which contributes to the ability to consider the individual and interactive influence of a wide range of early and later life factors on life course outcomes and carefully isolate effects. Perhaps the biggest strength of the Woodlawn Study is the length of follow-up, which is over 50 years with high rates of retention, making this the longest prospective community cohort study of urban Black Americans in the United States.

Despite these considerable strengths, there are several limitations to the study. One weakness of the study is the lower sample size during the adolescent follow-up leading to missing data when examining the full life course. There are also notable limitations in the measures during the adolescent assessment, such as the deviant or prosocial nature of peers and individual experiences of racial discrimination. A third weakness is the length of time between assessments. Although the data span the full life course, there is no prospectively-gathered self-reported data between ages 6 and 16 or between ages 16 and 32, which are developmental stages known for myriad biological, cognitive, and social changes. As a result, the Woodlawn Study is best suited to examining the long-term impact of childhood experiences on adult outcomes and to modeling behavioral pathways across, rather than within, the life stages of childhood, adolescence, young adulthood, mid-adulthood, and later life. Finally, similar to many longitudinal studies, such as the Woodlawn Study, the generalizability to other racial groups and more contemporary contexts is unclear.

Can I Get Hold of the Data? Where Can I Find Out More?

We are committed to sharing this rich data set with life course researchers. To this end, we utilize procedures that encourage collaboration and ensure broad access while also maintaining confidentiality to the research participants. We have included study documentation and interview schedules from each data collection effort on the Woodlawn website (www.woodlawn.edu) to allow researchers to gauge the data’s potential and the research questions the data can address. We have archived the child and adolescent Woodlawn data (1966–1976) with the Institute for Quantitative Social Science, Henry A. Murray Research Archive at Harvard University (https://murray.harvard.edu/). The de-identified adult interview data, administrative data, and associated documentation are available to users under a data-sharing agreement that provides for (1) a commitment to use the data only for research purposes and not to identify any individual participant, (2) a commitment to secure the data using appropriate computer technology, and (3) a commitment to destroy or return the data after analyses are completed, among other protections. Procedures require that researchers submit an abstract to the Woodlawn Study team (woodlawn@umd.edu). This procedure initiates communication with the requesting researcher to prevent duplication of efforts and to ensure full comprehension of the data. Once Institutional Review Board approval to use the data has been received by the requesting researcher, Woodlawn team members will prepare a dataset with the requested variables from the codebook. The researcher will coordinate with a Woodlawn Study investigator before submission of any publications or presentations to ensure accuracy, participant protection, and sensitivity to the population.

Profile in a Nutshell.

• The Woodlawn Study is an epidemiologically-defined community cohort study of 1242 Black Americans (51% female and 49% male, modal birth year = 1960), who were in first grade in 1966–67 in the Woodlawn neighborhood of Chicago, Illinois, USA.

• The study comprises extensive interview data over the life course, including self-, mother-, and/or teacher-reported assessments at ages 6, 16, 32, 42, and 62 (in progress), administrative records (i.e., education, crime, and death records), and census data.

• The data cover a wide range of focal areas across the life course, including family environment, socioeconomic indicators, education, social integration (e.g., marriage, community engagement, religious involvement) and social support, employment, racial discrimination, substance use, crime/victimization, and mental and physical health, including mortality.

• The study has been funded predominantly by the National Institutes of Health since 1975.

• The child and adolescent data (1966–1976) are archived with the Institute for Quantitative Social Science, Henry A. Murray Research Archive at Harvard University. Researchers interested in utilizing the non-archived information can request to use the data by submitting an abstract to woodlawn@umd.edu.

Appendix

Table 6.

Description of measures for key domains

Domain	Description of data by age
School adaptation	Childhood (C): Based on teacher feedback regarding the major developmental tasks expected from first graders, five scales of social adaptation were created. Teachers rated each student on these five areas: achievement, aggression, immaturity, inattention/restlessness, and shyness on a 4-point scale from adapting to severely maladapting using the Teacher’s Observation of Classroom Adaptation (TOCA) (Kellam et al., 1975). Teachers rated classroom conduct and math and reading achievement on a 4-point scale (excellent to unsatisfactory). Students took the Metropolitan Readiness Test; IQ tests were conducted in the 9 public schools Adolescence (A): Grades, aspirations and expectations, bonds, teachers’ perceptions, and adaptation (e.g., truancy) were based on mother- and self-reports. Chicago Board of Education records included standardized test scores, graduation, and grade retention
Socioeconomic resources	C&A: Mothers reported parental education, number of residential moves, number of adults and number of children in the household, family income, sources of income, home ownership, parental employment, and welfare receipt Young Adult (Y) & Midlife (M): Participants reported the number of residential moves, household composition, household income, sources of income, sources of financial stress, home ownership, and welfare receipt; educational history (i.e., years of schooling, degrees, graduation); and employment history (e.g., timing of first steady job, years in current job, quality and type of employment, the number of employers since their first steady job, and periods of unemployment)
Family relationships	C&A: Family discipline, affection, involvement, and supervision were reported by mothers in both assessments and by adolescents Y&M: Participants reported on marital relationships: number of marriages, the age each marriage started and ended, and how each ended (e.g., separation, divorce, widowed); Parenting: age at first birth and the number of children (both in and out of the household), extent of household warmth and conflict; Grandparenting (M only): number of grandchildren; and Family: frequency of contact with mother, father, children, siblings; caregiver roles of family members (M only)
Friendships	A: Self-reported items on satisfaction with friends, popularity with friends, friendliness, and importance of friends Y&M: Number of total & close friends, frequency of contact; number of people outside your household that you do something with in a typical week
Social support	A: Friendship network support (e.g., closeness, support, belongingness) Y&M: Sources of support (partner, family members, neighbors, co-workers, church members, professionals) for 8 different needs (e.g., when sick, when need money)
Community engagement	A: Number of activities, organizations, or teams a member of at school and outside of school Y&M: Type, frequency, and duration of involvement in social organizations, political participation, volunteer activities, and religious services
Religiosity	C: Religious membership, importance of religion A: Church membership, frequency of church attendance Y&M: Involvement in religious activities (e.g., frequency of attendance at places of worship) M: Spirituality, closeness to God, spiritual coping
Neighborhood	C&A: Mother-reported quality of housing, schools, transportation, shopping, safety, churches, jobs, and friendliness of neighbors A: 1980 Census data matched to participant address, including percentage below poverty, mother-only households, home ownership, and college-educated residents Y&M: Self-reported perceptions of neighborhood quality, physical and social disorder, adequacy of services, indicators of collective efficacy, and a measure of social cohesion Y&M: 1990 and 2000 Census data matched to participant address for each interview, including percentage below poverty, mother-only households, home ownership, and college-educated residents
Substance use, disorders, and treatment	A: Self-reported age of onset and frequency of lifetime and past 2-month use of tobacco, alcohol (beer/wine as well as liquor), marijuana, inhalants, cocaine, and 5 other substances. Maternal reports of frequency of self-use and household use of alcohol and drugs Y&M: Use: Adult self-reported tobacco, drug (and alcohol use, including age of first and last use, age of heaviest use, past year use, frequency of use, mode of administration, and problems associated with use; Disorders: Modules developed for the National Comorbidity Survey from the Composite International Diagnostic Interview (CIDI) (Anthony et al., 1994; Kessler et al., 1994). DSM III-R criteria guided the development of substance use abuse and dependence in young adulthood; DSM IV criteria were used for midlife for alcohol and 11 drugs, including tobacco dependence; Treatment: types of programs attended, nature of the provider, timing, and frequency of treatment for substance use
Criminal offending	A: Self-reported delinquent behaviors using a standard delinquency measure (23 items). Self-reported involvement with the criminal justice system for nine illegal activities Y&M: Self-reported criminal behaviors, including violent, property, and drug-related crime (27 items). Violent crimes (10 items): Carried gun or other weapon; get into a serious fight; beat someone up to get money or other valuables; used a weapon in a fight; purposefully injure someone physically; beat up someone within your family; beat up someone not in your family; forced someone to have sex; been in a gang fight; used threats to get someone to give you something; Drug crimes (3 items): get drugs in exchange for sex; gave drugs in exchange for sex; sold illicit drugs; Property crimes (13 items): Stole something worth at least $10; broke into store or other business; broke into someone’s home; used a stolen credit card; wrote or sprayed graffiti on walls, buses, shelters; set fire intentionally to a building, car, or lot; charged something to someone else without permission; swindled or conned someone; snatched a purse or pick a pocket; stole a motor vehicle; knowingly bought, sold, or held stolen goods; stole something from someone in your family; shoplifted Annualized arrest records from ages 17 to 52 including date, type of crime (up to 3 unique charges), type and method of disposition, and sentence for each Y&M: Data on incarceration is available from the study participants’ self-reports of incarceration (timing and length) from both adult interviews and if the respondent was interviewed in jail or prison
Victimization experiences	Y&M: Frequency and timing of seven victimization experiences, including having been injured by someone on purpose, having something stolen from their home, or having been raped
Physical health	C&A: Mother-reported general health of child, serious and chronic illness problems (e.g., birthweight, hospitalizations, chronic conditions) Y&M: Self-rated health (1 = poor to 5 = excellent). Self-reports of current health status and specific health problems (asthma, chronic pain, diabetes, HIV status) and interference with daily activities M: Detailed questions about health and illness such as number and type of chronic conditions and satisfaction with access to health care and health insurance
Mental health	C: Teacher-reported mental health of child (12 items), focusing on fearfulness, sadness, internalizing behavior and bizarre behavior, and a global assessment. Clinicians assessed flatness, depression, anxiety, hyperkinesis, and bizarre behavior. Mothers completed a Mother’s Symptom Inventory on psychological distress (38 items). Mothers reported on their own symptoms of anxiety and depression A: Self-reported scales (14 items) on anxious and depressed mood (Petersen & Kellam, 1977). Mothers reported on their own symptoms of anxiety and depression Y&M: Modules from the CIDI assessed depression, panic attacks, and antisocial personality disorder in young adulthood and depression in midlife. Additional indicators of psychopathology include suicidal ideations and behavior, and anxious mood
Racial identity and discrimination	Y: Self-reported racial socialization items (e.g., things taught as a child about being Black) (15 items) Y&M: Self-rated perceptions of racial discrimination in various settings (8 items) M: Centrality subscale of racial identity (5 items), the perceived racism scale (8 items), and perceived discrimination in community settings (11 items)

C, childhood; A, adolescence; Y, young adulthood; M, midlife.