© 2005 Behavioral Health Research Institute of the Southwest
612 Encino Place NE / Albuquerque, NM 87102 / 505.244.3099 / www.bhrcs.org

Drinking, Driving, and Drugs:
A Comparative Study of Juvenile Substance Abuse Offenders
in Bernalillo County, NM
By Nancy Owen Lewis, Ph.D. and Iyiin Chang, M.S.

2. METHODS

Table 1.
Study Population

The study population consisted of 1,518 offenders referred to the Juvenile Probation and Parole Office in Albuquerque, NM, (AJPPO) between July 1994-June 1995 for a substance abuse offense. Offenders were classified into one of three study groups – DWI, alcohol, and drug – depending on their initial or index referral during this period (Table 1). Five percent of these referrals (N=82) involved two or more substance abuse offenses.

Of these 82 referrals, 24 included both DWI and alcohol offenses. Three involved DWI and drug offenses, while four included DWI, alcohol, and drug offenses. Fifty-one referrals consisted of both alcohol and drug offenses (Table 2). In short, 26% (N=32) of the DWI referrals, 14% (N=79) of the alcohol cases and 6% (N=58) of the drug referrals included an additional substance abuse offense. In cases where multiple classifications were possible, offenders were assigned to a study group in the order of DWI, alcohol, and drug. Based on index referral, these study groups consisted of 118 DWI offenders, 531 alcohol offenders, and 869 drug offenders.

Table 2.
Substance Abuse Offenses Included in the
Index Referral (7/94-6/95)

The 531 alcohol referrals consisted of 432 cases for possession of alcohol, 40 for possession of alcohol in the vehicle, 26 for open container, 23 for minor procuring alcohol, 6 for minor under the influence, and 4 for drinking in the vehicle. The 869 drug referrals included 517 for marijuana possession, 144 for possession of drug paraphernalia, 58 for narcotics possession, 56 for narcotics distribution, 44 for marijuana distribution, 26 for drug possession, 20 for drug distribution, and 4 for solvent abuse (Table 3).

Data Collection and Coding

Each juvenile referred to the AJPPO is assigned both a file number and referral number. Subsequent referrals are assigned consecutive referral numbers. Consequently, each offender has one file number, which may include multiple referral numbers. In this study, referral is used to refer to a case, while offense refers to a specific incident within that case. One referral may include several offenses.

Information on each referral is coded by office staff on a two-page Client Tracking Referral System Master Record and the information entered into a computerized database. This form is comprised of three sections: 1) referral assignment and intake, which includes the offender’s name, other identifiers, and description; 2) preliminary inquiry, which includes demographic information; 3) formal and informal disposition; and 4) final disposition. In summary, each offender has a computerized file, which includes demographic information as well as offense history. Computerized files for the study population were obtained from the NM Children, Youth, and Families Department (CYFD), which oversees AJPPO. To comply with confidentiality requirements, CYFD staff removed all identifiers as requested.

Certain DWI information – including blood alcohol concentration (BAC) and crash involvement – was not available through this database. To obtain these data, project staff conducted chart reviews of 28 case files. However, since most DWI offenders were overage ( 3 18) at the time of the study, their records were not available. Instead, this information was obtained from the Division of Government Research at the University of New Mexico, which accessed the DWI Citation and Tracking System maintained by the NM Motor Vehicle Division (MVD) to obtain BAC and crash involvement data for the DWI offender group. Juvenile data was linked with MVD records using offender identifiers. If an offender had three identifiers, at least two had to match the MVD records. For some offenders, the only identifiers available were date of birth and date of arrest. In those cases, both variables had to correspond with the information contained in the MVD records to establish the record as a match.

Table 3.
Study Groups by Offense Type

Analyses

Chi-square tests were used to compare the three study groups in terms of the following characteristics:

1. Demographic – gender, age, ethnicity, education, household income, number of siblings, position in family, household , parent’s marital status, and residence
2. Arrest circumstances – referral agency, incident date (day, month, and time), and time between incident and referral
3. Response by juvenile justice system – time between referral and disposition, disposition type (formal, informal, temporary), and specific disposition
4. Offense history – number and type of prior offenses (conduct offenses, property offenses, offenses against persons, substance abuse offenses, and other offenses)
5. Subsequent offenses – number and type of subsequent offenses (conduct offenses, property offenses, offenses against persons, substance abuse offenses, and other offenses).

Data on gender, age, ethnicity, referral agency, incident date, time between incident and referral, and offense history were available for all subjects (N=1,518). Complete information on all variables was obtained for 78.1% of the offenders (N=1,186 offenders). Data on one or more variables were missing for the remaining 332 subjects (21.9%). A chi-square test was conducted to determine whether subjects with missing variables differed in terms of gender, ethnicity, or age from those with complete data. No significant differences were detected.

To determine recidivism rates, subsequent referrals to the AJPPO were monitored through June 1997 or until the offender reached age 18, which ever came first. Recidivism was analyzed using generalized linear models (Nelder and Wedderburn 1972) and involved the use of statistical software SAS PROC GENMOD (SAS Institute 1990). The dependent variable was the number of re-referrals, while the independent variables were age, gender, ethnicity, education, household, and study group. The purpose of the analysis was to determine which variables were predictive of recidivism.