Early Identification of the Serious Habitual Juvenile Offender Using a Birth Cohort in Philadelphia, 1958-1984 (ICPSR 2312)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Data Collection Notes View help for Data Collection Notes

Study Purpose View help for Study Purpose

Beginning in the mid-1980s, the Office of Juvenile Justice and Delinquency Prevention (OJJDP) funded the creation of Habitual Offender Units (HOUs) in 13 cities. HOUs were created to prosecute habitual juvenile offenders by deploying the most experienced attorneys to handle these cases from start to finish. By targeting the earliest points in the career sequence of the juvenile offenders, the greatest number of serious offenses can potentially be averted. Selection criteria to qualify for priority prosecution by an HOU usually encompassed one or more generic components relating to aspects of a juvenile's present and prior offense record. In Philadelphia, to be designated a serious habitual offender and to qualify for priority prosecution by the Philadelphia HOU, a youth had to have two or more prior adjudications or open cases for specific felonies and a current arrest for a specified felony. Specific felonies included criminal homicide and first degree felonies, such as rape, robbery, aggravated assault, burglary, arson, involuntary deviate sexual intercourse, and drug offenses (delivery or possession with intent to deliver). In other words, an accumulation of a prespecified number of serious delinquent acts would trigger a specialized prosecutorial response. Each city's HOU adopted its own selection criteria based also on a combination of legal and administrative considerations. Legal considerations centered on the seriousness of the specified charges and the mounting criminal liability engendered by multiple arrests or adjudications. Administrative considerations centered on the number of cases generated by the selection criteria. However, none of the cities adopted selection criteria on the basis of systematic research and almost no research was conducted on how well these HOUs' selection criteria identified serious habitual offenders. The following questions arose about the Philadelphia selection criteria: (1) The specified crime charges were for serious offenses, but should other crime charges have been adopted? (2) Some of the specified crime charges could be related to the more repetitive types of criminal behavior (e.g., robbery, burglary), but was a single charge better than a combination of charge types for identifying serious habitual offending if a relationship between charge type and future offending exists? (3) Two prior adjudications or pending cases represent a mounting criminal record and appear to reflect a commitment to continued criminal activity, but is there a reason to adopt another threshold, preferably one case rather than two? and (4) Should the prosecutors consider other information, such as unofficial information, to identify the serious habitual offender? The main objective of the study was to determine which variables, reflecting current selection criteria, criminal histories, and personal characteristics, were most strongly and consistently related to selected aspects of a juvenile's future criminal career. With the capacity to identify accurately serious habitual offenders, prosecutors would be in a better position to allocate limited staff, money, and technology to those cases that are most likely to yield the greatest payoff. Intervention programs, whether based on incapacitation, deterrence, and/or rehabilitation, would also work best if the right group of juveniles were selected for priority prosecution.

Study Design View help for Study Design

This study attempted to identify variables, reflecting HOU selection criteria, criminal histories, and personal characteristics, at each of a juvenile's felony police contacts to determine which were most strongly and consistently related to the frequency and seriousness of future juvenile and young adult offending. The first three police contacts in a Philadelphia juvenile offender's record were of special interest because they included the earliest point (i.e., the third contact) at which a youth could be prosecuted in the Philadelphia HOU, under their selection criteria. Analyses were conducted separately on the first, second, and third police contacts to (1) assess Philadelphia's current selection criteria, which were triggered at the third felony police contact, (2) establish whether there were variables that were consistently related to a juvenile's future criminal career across the first three felony transitions, and (3) determine whether there were grounds for prosecutors to select juveniles for prosecution by the HOU at the first or second felony police contacts rather than waiting for the third. To accomplish this, an assessment was conducted using a group of juveniles born in 1958 whose criminal career outcomes were already known. Applying the HOU selection criteria to this group made it possible to determine the extent to which the criteria identified future habitual offending. Data for the analyses were obtained from a birth cohort of Black and white males born in 1958 who resided in Philadelphia from their 10th through their 18th birthdays. For each birth cohort member with at least one felony police contact during the juvenile years, spanning the 10th through 18th birthdays, a juvenile felony sequence was created starting with the juvenile's first felony and ending with the juvenile's last felony. Then arrest and judicial histories for the juvenile and young adult periods were merged to form a continuous sequence, spanning ages 10 through 26. This sequence was the basis for computing the dependent variables. At each of the sequential points, or transitions, data were provided for the Black and white youngsters separately and for both races together. The pooled data files retained only up to three police contacts. In order to assess the sensitivity of the findings to sampling error, statistical analyses were first conducted on a randomly partitioned construction sample comprising 70 percent of the subjects. The final statistical results at each felony transition based on the construction sample were then reexamined at the corresponding felony transition in a validation sample comprised of the remaining 30 percent. In response to the prosecutors' interest in narrowing the scope of eligible juveniles to just the "high" group (i.e., most frequent/serious), the researchers noted that percentages assigned by the prosecutors to identify the "high" group converged on the top quartile. Therefore, from an operational standpoint, juveniles with one or more specified felonies who fell into the top 25 percent with respect to future repetitive serious offending would be included "in" the HOU. Juveniles who fell below that point would be directed "out" of the HOU. To reflect this in-out decision format, dependent variables were dichotomized, using the 75th percentile to split the variables into a top 25 percent and a bottom 75 percent. To get some idea of the robustness of findings and to investigate how alternative definitions of the "high" and "low" group would affect the selection criteria used to make the in-out decision, final results based on the 25/75 percentile split were reanalyzed using a 10/90 split.

Sample View help for Sample

The study used a cohort of 13,160 males born in 1958 who resided in Philadelphia from their 10th through their 18th birthdays. Cohort members were restricted to those who had lived in Philadelphia without interruption from age 10 to 17.

Universe View help for Universe

All juvenile offenders in Philadelphia, Pennsylvania.

Unit(s) of Observation View help for Unit(s) of Observation

Data Source View help for Data Source

Data Type(s) View help for Data Type(s)

Description of Variables View help for Description of Variables

Dependent variables represent aspects of a juvenile's future criminal career (i.e., aspects that occurred subsequent to the point in the felony sequence under analysis). Criminal careers represent police contacts for the juvenile years and arrests for the young adult years for which police contacts and arrests are synonymous. Dependent variables were computed in five ways, reflecting different criminal career aspects: (1) the rate per year of police contacts, (2) the average seriousness per police contact, (3) the average seriousness per year of police contacts, (4) the average number of criminal charges per police contact, and (5) the average number of criminal charges per year. These five dependent variables were computed for four crime type groups, in increasing order of restrictiveness: (1) all offenses (i.e., crime events), (2) specified felonies (i.e., serious crimes), (3) Uniform Crime Reports (UCR) index crimes, and (4) UCR violent index crimes (i.e., dangerous crimes). The five criminal career aspects and four crime type groups were computed for two age intervals: (1) the remaining juvenile years (through the youth's 18th birthday) and (2) the remaining juvenile years and the young adult years (through the youth's 27th birthday). This resulted in a total of 40 dependent variables. The data also contain various dummy variables related to prior offenses, including type of offense, number of prior offenses, disposition of the offenses, weapon used, whether the offense was gang-related, age at first prior offense, and seriousness of first prior offense. Dummy variables pertaining to the current offenses include type of offense, type of injury to victim, number of crime categories, number of charges, weapons used, number of offenders, gender, race, and age of offenders, type of intimidation used, number of crime victims, gender, race, and age of victims, type of victimization, characteristics of offense site, type of complainant, and police response. Other variables include offender's socioeconomic status, age at first prior offense, age at most recent prior offense, age at current offense, and average age of victims.

Response Rates View help for Response Rates

Not applicable.

Presence of Common Scales View help for Presence of Common Scales

None.

Original Release Date View help for Original Release Date

1998-10-15

Version History View help for Version History