A study of serious adolescent offenders as they transition to adulthood and out of crime

Out of Community Placements

Description

Specific Information: Out of Community Placements

Self-reported participation in sanctions and interventions involves an assessment of both residential and community-based social services. This information was obtained using a modified version of the Child and Adolescent Services Assessment (CASA) (Burns et al. 1992). The CASA was designed to assess the use of mental health and social services via self-report from youth age 8 to 18 and their parents. It explores services in both the justice (juvenile and adult systems) and mental health sectors.

The Pathways study asked participants if, during the recall period, they had an overnight stay in seven different types of placements: Drug/alcohol/detox unit, psychiatric/medical hospital, residential treatment unit or group home, foster home, detention/jail/secure facility, shelter, and any other type of placement (e.g., halfway house). These individual residential facilities were recoded into groups based on their general mission and target population. The groupings were derived from several discussions with service providers and policy-makers from both sites and other locales (including juvenile court service administrators, practitioners, lawyers and judges). These experts considered the seven general categories used in the CASA for residential placements as overly broad, and the groups were modified as a result. Juvenile court administrators and service providers from each site assisted in classifying particular facilities when their group membership was questionable. Ten categories of residential care settings were examined:

• Drug or alcohol treatment unit. These are facilities where the primary focus is the provision of substance use treatment services. Detoxification and longer-term substance use treatment programs are both included, with the vast majority being longer-term treatment facilities.
• Psychiatric hospital or psychiatric unit of a general hospital. These are settings which provide inpatient acute care to evaluate and stabilize individuals with mental health or behavioral problems.
• Foster Home. These are defined as placements with non-kin ordered by child welfare or dependency court
• Shelter. These facilities provide short-term, non-secure, temporary out-of-home care.
• Jail and prison. These settings have incarceration as the main goal, and are almost all adult settings. Jails, usually locally run, hold people until trial or for relatively short sentences after trial. Prisons are usually state-run and hold people for a longer term after trial. Federal adult detention centers and Immigration and Naturalization Services (INS) detention centers are also included in this category, although these constitute a very small portion.
• Detention. These are facilities where adolescents await their adjudication hearing or more permanent placement after adjudication and disposition. These can be adult or juvenile facilities.
• Pennsylvania Youth Development Centers (YDC) / Arizona Department of Juvenile Corrections (ADJC). These facilities are state-run, secure juvenile facilities, formerly characterized as "state training schools". They provide secure custody, education, and treatment to committed youth.
• Contracted residential treatment (general). These settings provide residential care within a structured environment. A range of services may be offered, usually centered on a specific model of intervention (e.g., peer culture, physical challenge), and there may be varying amounts of security and access to the community.
• Contracted residential treatment (mental health). These settings have an integrated residential program of therapies and activities. The primary focus of treatment is on the youth's mental health needs, and the facility targets mentally ill adolescents.
• Other. This includes any residential setting (adult or juvenile) not captured by the above categories such as a residential military-style high school or YMCA.

Specific timepoint information obtained for each facility

If an overnight stay was acknowledged, specific information was obtained, including:

• For each month of the recall period: we recorded the facility name, a four-digit identifying facility code, the length of each stay (in days), the number of days on a home pass from each facility, and whether they had a stay lasting at least seven nights in a row at each facility.
• For each of the ten different setting types: whether they had a stay in any facilities of each type; number of different facilities within each type; total number of days spent in each setting type summing across all facilities within that type; and an overall rating of the helpfulness of the programs.
• For stays lasting seven or more days across the recall period, additional information was obtained for each facility, including: type of facility; whether it is a juvenile, adult or specialized setting; total length of the stay; total number of times in and out of the placement; length of the longest stay; presence of a primary counselor and a release counselor and the frequency of meeting with both; ratings of safety; presence of caring adults at the placement; whether any services were received and if so, the number of sessions received per day as well as their assessment of how helpful the service was to them; peer offending; and items assessing the participant's perceived helpfulness of the facility in preparing them for their release.
• Cumulative summary variables: total number of days spent in each setting type (summing across all facilities within that type) and the proportion of time spent in each setting type.

Proportion of time in the community and facility

Additional variables summarizing the participant's time in a facility and on the streets are also computed. Because our subjects have recall periods that vary in length, it is important to consider the subject's time in a facility and time in the streets along a common metric. For this reason, a proportion score representing the subject's time in the streets and time in a facility is calculated for each recall period. There are two variations of these scores: one which includes all ten setting types and one which includes only stays in seven settings without access to the community (Drug/Alcohol, Psychiatric, Jail/Prison, Detention, Ydc/Adjc, Contracted Residential, and Contracted Residential Mental Health). The version of proportion score used should depend on which is more appropriate for the analytic question being explored. The following items are available at the recall, monthly and yearly levels:

• All setting types considered: Total number of days spent in all settings, proportion of time in all setting types, and proportion of time on the streets considering all setting types
• Settings without access to the community: Total number of days spent in facilities without access to the community, proportion of time spent in these facilities, and proportion of time on the streets considering these settings.

The proportion scores were computed using the following steps:

• The numerator for the proportion represents the days in a facility (either all setting types, or setting types without access to the community).
• The denominator is the number of days covered in the recall period. This value represents the number of days in each month that is represented in the calendar data. The values for the last month of the recall period, however, are reduced if the interview was actually conduced in that month. For example, if the current interview is conducted on 02/18/05, the value for the last month of the recall period is 18.

Specific Information: Number of Periods Out of the Community

To assess the patterns of sanctions and interventions that involved removal from the community, we created a set of variables that reflect the number of distinct periods during which the youth was in an out-of-home placement for consecutive months, without time in the community. A full explanation of the coding rules implemented for the creation of these variables can be found in the detailed documentation for this calendar. The following variables are available:

• For each linear month: a running count of the number of known stays (this includes stays that were known to the research interviewer to occur during a missed timepoint interview; examples include situations when the interviewer knows the participant is in prison but the participant can't complete the interview because he/she is in restricted housing and not permitted to complete the interview); a running count of the number of reported stays (this is a count based only on available self-reported timepoint data, thus stays that occurred in a missing time point are not included).
• Cumulative through seven years: the total number of distinct stay episodes, both known and reported.

Specific Information: Facility Stay Series

A separate dataset called the "Facility Stay Series" provides an overview of each individual stay reported by the subject through the end of 84months. A stay here is defined as a continuous run of months at a single facility. The stay begins in the month the subject first reports being at that facility and ends in the month where they no longer report being there (either because they moved to a different facility or because they were released to the community). A full explanation of the rules for creating these variables can also be found in the detailed documentation for this calendar. Each stay is characterized by a set of 62 variables:

• Identifying facility code; whether the facility is an adult, juvenile or specialized setting; type of facility (drug/alcohol vs. psychiatric, etc.); recall period months (s#m#) spanning the stay; length of the stay in months and days; presence of a primary counselor and a release counselor at any point during the stay and the frequency of seeing both; overall safety rating; presence of caring adults and domains of social support; whether any services were received and if so, the number of different services received, total number of sessions received (for each individual service as well as total across all services), and the number of sessions received per day (for each individual service and total across all services); peer delinquency (antisocial behavior); and future orientation of the program.
• An additional variable provides the total number of stays through the end of follow84.

Specific Information: Variables characterizing each month

This series of variables provides basic information related to the facility in which the subject had the longest stay, defined here as the "main facility", for each linear month. The following variables are available:

##Identifying facility code; whether the facility is a juvenile, adult or specialized setting, or community month; type of facility; and number of days spent at the facility

Additional sections related to this calendar

Other sections of the Sanctions and Interventions calendar offer variables that supplement the information contained herein (see "Calendar Data/Sanctions and Interventions" for additional codebook sections). Specifically:

• Community based services. Information regarding services the subject received while in the community, including the frequency each service by month, focus (for drug/alcohol treatment, anger management/social skills training), and circumstances (whether the service was court-ordered, whether the youth attended alone or with family).
• Court monitoring services. Information regarding four types of court monitoring services, including probation/parole, community intensive supervision, drug court programs, and court-ordered groups.
• Contacts with the justice system. Information regarding five different types of contact with the legal system, including whether the youth was picked up by the police, arrested and charged, had a court appearance, received a summons, or received police or court assistance as the victim of a crime.
• Medication. Information regarding prescription medication the subject has taken for emotional and/or behavioral problems, as well as information regarding medication compliance.
• Official Court Record Information. Information regarding arrests and court involvement, including a monthly plotting of arrests based on official record information as well as juvenile facility stays recorded in the juvenile court records from each locale (this information is not available for period of time in the adult criminal justice system). This can be found under "Codebook -- Official Court Record Information".

Be sure to review the descriptions for each of these codebook sections in order to obtain a complete understanding of the information that we have available regarding sanctions and interventions.

Validity of self-report services data

As noted in Mulvey, Schubert and Chung (2007), we have reasonable confidence in the accuracy of the self-reported service data for two reasons. First, test-re-test reliability conducted by the authors of the CASA found very high reliability for reports of outpatient services (kappa = .8) and a moderate to high range of reliability (kappa = .6 to 1.0) for inpatient, out-of-home, and juvenile justice services (Ascher et al. 1996; Farmer et al. 1994) Our approach mirrors that used by these investigators, and we would expect our reliabilities to be about the same as a result.

Second, using official records in one of the data collection sites (the ProDES system in Philadelphia), we found high agreement between this information and the self-report data regarding the occurrence and timing of the receipt of residential services. The ProDES information system is a well-established and longstanding cooperative effort between the Philadelphia Department of Human Services and the Crime and Justice Research Center to track service involvement for youth in the juvenile justice system (Jones, Harris, and Fader 1999). We compared the ProDES reports of service involvement over a two-year period to our reports in the Pathways study data set for the sample used here. Our self-reported stays in settings other than jail and detention facilities (these are not covered by the ProDES system) were corroborated 96% of the time in the ProDES system (n=521). Conversely, of the participants who overlapped across the two studies, 97% of the stays recorded in ProDES were also present in our self-report data (n=343). In addition, there was high agreement about the timing of residential facility stays. We found 97% agreement regarding the intake and discharge month if we allowed for a two-month discrepancy in the reports (n=175) and 90% agreement if we allowed for only a one-month discrepancy (n=175). Although we do not have access to parallel validation data for the Maricopa County site, it seems reasonable that these results would generalize to the reports of service use from that site as well.

Descriptive Information: Monthly Data Characterizing the Recall Period

As a standard practice, the specific calendar information will be accompanied by two variables which describe the recall period. This information is important for the user to consider when attempting to use data characterizing the recall period (e.g., measures) in conjunction with the monthly-level calendar data. In addition, this information is useful if the user is viewing events from a developmental perspective. These variables include:

• The actual number of days in each month that is represented in the calendar data.
• The calendar month and year mapped to the recall month (e.g., if S1M04 covers February 2003, the 'number of days' value for that month will be 28, and the calendar month associated with S1M04 will be 02/2003, "RealDate").
• The subject's age at each month of the recall period, available in two forms: 1) the subject's age truncated to a whole number, and 2) the age as a continuous variable. (The subject age at the time of the interview is available in the "Demographics" codebook, available under "Measures").

General Information: Use of monthly life calendar data

Data regarding the participant's self-reported residential stays is captured using a monthly life-calendar approach (Belli, 1998; Caspi, Moffitt, Thornton, & Freedman, 1996), where the research participant is provided with a visual calendar that contextualizes the recall of research data by anchoring information to salient events. Specifically, individuals are first asked to recount salient events which occurred in the recall period (e.g. birthdays, deaths) and this information remains visible to the participant as an anchor point for the timing of events in each of several life calendar domains. This approach thus creates an integrated view of activities in all of the domains examined, has firm roots in the science of how people remember events and life situations (Bradburn, Rips, and Shevell, 1987; Belli, 1998), and capitalizes on these processes to generate accounts of past events. On a practical level, it provides researchers with a richer set of data points. Instead of simply getting a summary measure of life changes over an extended recall period, the monthly life-calendar places these changes at specific points in time, opening up the possibility of examining sequences of events and potential causal mechanisms within individuals (Fals-Stewart, 2003; Mulvey, et al., 2006).

General Information: Conversion of data to linear months

The monthly Pathways data, in its raw form, is not suited for some kinds of analytic approaches (e.g., trajectory analysis). Each time point interview allows for a maximum of eight or 14 months in the recall period, depending on the follow-up wave (eight months was the maximum for time points 6-36 and 14 was the maximum for time points 48-84). This means that there is a corresponding variable in the dataset reflecting events occurring in each of those months through a maximum of 14 months (the outer limit of any of the possible months covered). If however, the recall period did not include the maximum number of months (as is most often the case), there will be variables with no data. For example, subject 1 has a recall period of five months for follow-up 12 so this means he/she will have data in five monthly variables but not in the remaining nine. Subject 2, has seven months in the recall period for follow-up 12 so he/she will have data in all but seven of the monthly variables. The recall length is set by programming code based on the current date in relationship to the date of the previous interview (see "Interview Information" under Measures for a more detailed description of how the length of the recall period is determined). However, the programming code did permit the interviewer to "reset" by hand the length of the recall period. This was done infrequently, but in some instances it created a situation where we obtained two reports of the same month. For example, follow-up 6 covered months January to June and the interviewer resets the follow-up 12 recall period to start with June (leading to two different reports for the month of June). A series of data cleaning decisions (described in the "Out of Community Placements Calendar Documentation") were implemented to correct these situations but we note them here because they are relevant to the conversion of the data to linear months.

The "linear months" data set-up corrects these two situations. The "linear months" data reformats the variables so that each variable is a sequential representation of life event data for each month of the research participant's life from the baseline interview forward. In this format, variables that were place-markers for months not covered in the recall period are eliminated and situations where there were two reports for the same month are corrected. Thus, "linear month 8" actually represents eight calendar months from the baseline and "linear month 16" is actually 16 months past the baseline interview.

A specific list of variables available in the linear format is provided below. Also provided is a "map" to link the linear month back to the recall period and month in which the information was originally collected. This is important to know when recall-level data is being used in conjunction with the monthly event calendar data.

Data Issues

General

• The follow06 through follow36 timepoints allow for a maximum of 8 months in the recall period, while follow48 and later will have a maximum of 14. Note that while variables for months 9 through 14 are included in the datasets for follow06 through follow36, these are essentially place markers and in no instance will any case have data for these months until the follow48 period (see description above).
• Interviews completed with versions 01.07 and earlier captured a limited amount of information regarding out of community placements. Cases completed with these early versions will have data regarding the number of days they stayed overnight at each facility (for each month of the recall period), and a four-digit identifying code assigned to each unique facility, however, these will not have any information about the services received during the stay and the youth's perceptions of the experience (e.g., feelings of safety and caring adults present in the facility).

Beginning with version 01.08 through the latest interview versions, the calendar was expanded to capture more specific monthly information about each facility in which the subject had an overnight stay (including whether they were released on a home pass that month), as well as information regarding the subject's experiences (feelings of safety, caring adults at the facility, etc.) and services they received during their stay (drug/alcohol treatment, group therapy, etc.).

In the follow06 timepoint, 425 cases were completed with an interview version of 01.07 or earlier; in the follow12 timepoint, there are 128 cases. All cases in follow18 and later were completed with an interview version of 01.08 or later.

Adult versus Juvenile Facilities

• At the time of the interview, there was no distinction made between adult and juvenile facilities. After data collection was completed, two additional variables were created (S#DetA_Stay -- had a stay in an adult detention; and S#OtherACC_Stay -- had a stay in an adult community corrections facility) to identify those cases that had a stay in an adult facility. Since these facilities were not separated during the interview, the majority of variables in this calendar will be grouped into the ten setting types (all "Detention" variables will include both adult and juvenile, and all "Other" will include both juvenile and adult community corrections).
• Adult detention facilities can be found both within the Detention and JailPrison sets of variables. These can be identified using either the "facility type" variable (option 25 = adult detention), or by the facility code. Facilities with the following codes have been identified as adult detention facilities:
• 1015, 1130, 1188, 1243, 1258, 1282, 1283, 1291, 1290, 1295, 1354, 1407, 1408, 1409, 1481, 1482, 1483, 1526, 1542, 1595, 1646
• Adult community corrections (facility type, option 26 = adult community corrections) can be found within both the Contracted Residential (general) and Other sets of variables. Facilities with the following codes have been identified as adult community correction settings:
• 1374, 1450, 1480, 1531, 1573, 1603, 1588
• In some instances, a youth may go to a juvenile facility after a stay in an adult facility. This happens in about 10% of the cases in each site and we were unsure if this situation was an error in the self report data or if it could actually happen. Consultation with a Philadelphia public defender indicates that this scenario does actually occur in practice. For example, a youth may return to the juvenile system after being in the adult system for a particular purpose (e.g., restoring competency to stand trial). Unfortunately, we are not able to sort out the rationale for a juvenile stay after an adult system confinement, we can just observe that it occurred.

Identifying Facility Code

• A four-digit code is associated with each unique facility, assigned by staff at the Coordinating Center. It should be noted that attempts were made to ensure that unique codes represent distinct facilities; however, it is possible that, due to the participants limited knowledge, we have inadvertently coded the same facility with multiple codes if various participants provided distinct names for the same facility.
• We have identified the following codes that actually represent the same facility, and can be merged together:
• Codes 1075, 1052, and 1391 (all are type contracted residential mental health)
• Codes 1096, 1103, 1154 (all are type drug/alcohol)
• Codes 1097, 1104, 1153, 1479 (all are type psychiatric hospital/unit)
• Codes 1101, 1118 (both are type jail/prison)
• Codes 1107, 1120 (both are type jail/prison)
• There are also instances where multiple facilities are run by the same umbrella organization, or, a single facility offers more than one type of treatment (such as two separate units for drug/alcohol and psychiatric treatment). We have assigned separate codes in both situations but we are listing these codes here so that they can be merged together if desired.
• Codes 1022, 1664, 1669, 1670 (one location that offers more than one type of treatment)
• Codes 1005, 1016, 1665 (one location that offers more than one type of treatment)
• Codes 1058, 1059 (one location)
• Codes 1038, 1039, 1066, 1067, 1068, 1069, 1205, 1259, 1347 (umbrella organization)
• The following codes represent Pennsylvania Department of Correction facilities:
• Codes 1047, 1351, 1545, 1257, 1014, 1413, 1132, 1425, 1406, 1051, 1240, 1449, 1184, 1477, 1136, 1457, 1459, 1421, 1131, 1473, 1414, 1241, 1412, 1127, 1478, 1125
• These codes represent Arizona Department of Correction facilities:
• Codes 1085, 1083, 1091, 1079, 1089, 1088, 1080, 1081, 1086, 1090, 1087, 1084, 1379, 1539, 1363

Street time Proportion Scores

• When computing the street time proportion scores, in some cases, the sum of the number of days in all settings exceeds the number of calendar days covered in that recall period. This is largely due to the way in which the final month of the recall period is handled by the interview. Interviews that were conducted after the 15th of the month will have the last month of the recall period set to the number of days up through the day of the interview (e.g., the number of days in the final month for an interview conducted on 02/18/05 will be set to 18; however the total number of reported days summed across all settings exceeded 18). These cases were corrected to set the sum of the days in all settings equal to the number of calendar days covered by the interview to avoid producing a proportion score greater than one.

This correction was made only to the computed sum of all days that is used in this proportion score; the actual number of days reported by the subject for each month was not changed. As a result, summing both the recall level and monthly level self-reported number of days will not necessarily equal the sum of all days (S#NDaysAllSettings) used for these proportion scores.

• In addition, there are cases where the proportions at the recall level are inconsistent with the proportions at the monthly level. For example, the recall proportion may indicate the subject spent 100% of the time in a facility, but at the monthly level the proportions are less than 100. Or, the recall period proportion may reflect no time at all in a facility, but we get a non-zero proportion at the monthly level. This is again due to the number of days in all settings being greater than the number of days in the month (e.g., if S1M05 covers February, the number of days should not exceed 28; March should not exceed 31; April should not exceed 30, etc.).

Facility Stay Series Dataset

• A variable called "Additional_Information" has been included in the facility stay series dataset. This marker indicates situations in which other sources of information (e.g. the study tracking website, the living calendar, and the release interviews) provide information that wasn't fully captured in the residential services calendar. The markers could indicate that 1) there is a stay represented in one or more of the above sources that was not recorded on the services calendar at all, or 2) for a stay that is represented in the dataset, there are additional days that were not recorded on the calendar
• Items available regarding out of community placements

For an overview and a detailed list of the questions included with this calendar please select the link(s). In addition to providing an overview of the "flow" of the calendar and a detailed listing of the questions, this document notes version issues (i.e. questions/variables that are only present for a sub-sample due to their later addition to the interview) and provides other information that is critical to using and interpreting the data correctly. The table below gives you an overview of issues related to each construct noted above and it also provides you with the page numbers within our detailed document that address each of these constructs. Please be sure to consider this information carefully before moving forward with your analysis.

Certain questions listed in the table below are asked once for each facility, however only the SPSS variable names specific to stays in drug or alcohol facilities will be listed as an illustrative example and to save space. The variable names for the remaining setting types follow the same naming convention, where the "DAXX" portion of the variable name will be replaced with the following:

• PsychXX
• JPXX
• DetXX
• YdcXX
• CRXX
• CRMHXX
• OtherXX

Click here to download a detailed document in PDF format.

References

• Caspi, A., Moffitt, T., Thornton, A., Friedman, D., Amell, J., Harrington, H., et al. (1996). The Life History Calendar: A research and clinical assessment method for collecting retrospective event-history data. International Journal of Methods in Psychiatric Research, John Wiley & Sons, Inc., US, 6(2), 101-114.
• Belli, R.F. (1998). The structure of autobiographical memory and the event history calendar: Potential improvements in the quality of retrospective reports in surveys. Memory, 6(4), 383-406.
• Burns, Barbara J., Adrian Angold, Kathryn Magruder-Habib, Elizabeth J. Costello, and M. K. Patrick. 1992. The Child and Adolescent Services Assessment (CASA). Durham, NC: Duke University Medical Center.
• Mulvey, E. P., Schubert, C. A., & Chung, H. L. (2007). Service use after court involvement in a sample of serious adolescent offenders. Children and Youth Services Review, 29, 518-544.

This Web site is maintained by the Center for Research on Health Care (CRHC) Data Center. Contact the webmaster at dcweb@pitt.edu. All rights reserved.