Pittsburgh Study: Sample and Methodology

The Pittsburgh study of teenage fatherhood is part of the ongoing Pittsburgh Youth Study on delinquency in young urban males, which is being conducted by researchers at the University of Pittsburgh. A recent OJJDP Fact Sheet on the Program of Research on the Causes and Correlates of Delinquency (Browning et al., 1999) provides a summary of the research design for the Pittsburgh Study. The specific study of the precursors of teenage fatherhood was conducted by Magda Stouthamer-Loeber and Evelyn H. Wei. The 506 teenagers analyzed in the teenage fatherhood study are the oldest of the boys in the three age cohorts from the Pittsburgh public schools that make up the sample for the larger Pittsburgh Youth Study. (For details on sample selection, see Stouthamer-Loeber and Wei, 1998.) These participants were 12-13 years old at the beginning of this study in 1988 (they are now in their early twenties). The racial composition of the sample was equally distributed among African Americans and whites and was comparable to the racial composition of the Pittsburgh public schools. About 44 percent of these boys lived with a single parent, and 36 percent of their families received public assistance.

Participants in the study were questioned by researchers in regular interviews or assessment waves, beginning with a screening wave and followed by assessment A. These assessments were conducted at 6-month intervals (five times), and subsequently at yearly intervals. Data from the first nine assessment waves (screening through assessment K) were used for this study. Information was also collected from parents and teachers, and police, court, and school records; whenever possible, information from different informants (the participant, the parent, and the teacher) was combined.

Teen fatherhood was defined in the Pittsburgh Study as becoming a father before age 19 years. Two measures of delinquency were employed. The first, serious delinquency, was defined as having committed any of the following offenses by the time of assessment A: car theft, burglary, strong-arming, attack to seriously hurt or kill, or rape. The second measure of delinquency, called varied serious delinquency, covered the 4-year period from assessment waves B through I and was defined as having committed a delinquent act in two of the above categories within a 1-year period. Thirty-seven percent of the participants had committed an act of serious delinquency by assessment A, and slightly more than one-fifth of the sample (22 percent) was classified as having engaged in varied serious delinquency subsequently, during waves B through I.

Independent variables measured in the study included participants’ behaviors, attitudes, demographics, family, peers, and school performance. Statistical analyses were performed to obtain odds ratios and confidence intervals.1 Two stepwise logistic regressions2 were also run. For the second part of the study, investigating the impact of fatherhood on delinquency, the 62 fathers in the sample were matched with 62 nonfathers in terms of age, race, and neighborhood. For cases that reported fatherhood at wave I, data from wave K were used to examine delinquency in the year after fatherhood.

1 An "odds ratio" expresses the likelihood of an outcome, given the presence of a risk factor, as compared with the absence of that risk factor. The "confidence interval" that accompanies a statistic describes the degree of certainty that the statistic is precise.

2 A "stepwise logistic regression" is employed to compare two groups of individuals with respect to a particular outcome (e.g., teen fathers and nonfathers); taking multiple risk factors into account, it develops a model that predicts whether the outcome will occur.

Teenage Fatherhood and Delinquent Behavior Juvenile Justice Bulletin   ·  January 2000