National Archive of Criminal Justice Data
This dataset is maintained and distributed by the National Archive of Criminal Justice Data (NACJD), the criminal justice archive within ICPSR. NACJD is primarily sponsored by three agencies within the U.S. Department of Justice: the Bureau of Justice Statistics, the National Institute of Justice, and the Office of Juvenile Justice and Delinquency Prevention.
Evaluation of SAFEChildren, a Family-Focused Prevention Program in Chicago, Illinois, 2006-2010 (ICPSR 33101)
Principal Investigator(s): Tolan, Patrick, University of Virginia; Henry, David, University of Illinois at Chicago; Gorman-Smith, Deborah, University of Chicago. Chapin Hall; Schoeny, Michael, University of Chicago. Chapin Hall
Schools and Families Educating Children (SAFEChildren) is a family-focused program designed to aid families residing in high risk communities with child development during the child's transition to school. The program has the goal of building protection and impeding risk trajectories for aggression, violence, and school failure. The program utilizes multiple family groups (four to six families) combined with reading tutoring for the child. The SAFE Effectiveness Trial (SAFE-E) involved community providers delivering the family group intervention and upper grade students delivering the tutoring program. The trial took place between 2006 and 2010, and involved two age cohorts of children. Collaborating with two community mental health agencies and six elementary schools serving high-poverty, high-crime neighborhoods in Chicago, Illinois, families were randomly assigned to intervention groups of four to six families during their child's first grade year. Children also received tutoring from tutors selected from the upper grades of the child's school. Assessments were collected prior to, during and after the intervention to assess developmental influences, fidelity, process, and implementation characteristics that might affect impact. The purpose of these assessments was to examine the relation of implementation qualities to variation in intervention effects. Quality of implementation was expected to affect short and long-term impact of the intervention, focusing on three primary areas: (1) fidelity of implementation of the program, (2) provider characteristics, such as tutors' reading levels, and attitudes and orientation of the family intervention providers, and (3) quality of support for implementation. The data are from fidelity and process measures developed for this study and measures completed by parents, teachers, and children over four waves of measurement spanning two years, beginning in the fall of each child's first grade year.
One or more files in this study are not available for download due to special restrictions ; consult the restrictions note to learn more. You can apply online for access to the data. A login is required to apply for access. (Instructions on YouTube.)
Access to these data is restricted. Users interested in obtaining these data must complete a Restricted Data Use Agreement, specify the reasons for the request, and obtain IRB approval or notice of exemption for their research.
Tolan, Patrick, David Henry, Deborah Gorman-Smith, and Michael Schoeny. Evaluation of SAFEChildren, a Family-Focused Prevention Program in Chicago, Illinois, 2006-2010. ICPSR33101-v2. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2015-05-12. http://doi.org/10.3886/ICPSR33101.v2
Persistent URL: http://doi.org/10.3886/ICPSR33101.v2
This study was funded by:
- United States Department of Health and Human Services. Centers for Disease Control and Prevention
- United States Department of Justice. Office of Justice Programs. National Institute of Justice (2006-JP-FX-0062)
Scope of Study
Subject Terms: AIDS, alcohol, behavior problems, child development, crime prevention, delinquent behavior, families, family relations, HIV, inner city, intervention, parental attitudes, parental influence, process evaluation, program evaluation, risk, tutoring
Smallest Geographic Unit: none
Date of Collection:
Unit of Observation: Individual, Family intervention group
Universe: All children attending kindergarten during the recruitment period, which began in Fall of 2006, at six participating schools in Chicago, Illinois, their families, and their teachers. Seventh and eighth graders recommended by teachers at the six participating schools. Family intervention providers at two community agencies providing mental health services in the neighborhoods in which the schools were located.
Data Types: survey data
Data Collection Notes:
This study consisted of two distinct components. The first was an effectiveness evaluation comparing a control group of children and parents to a treatment group that participated in the SAFEChildren program. The second was a process and fidelity measurement which looked only at the experiences of the treatment group. The Sample Data and Twin ID Data (Datasets 1 and 2) provide data on participants, and link the two components of the study. Datasets 3-13 include data exclusively associated with the process and fidelity measurement. Datasets 14-25 include data exclusively associated with the effectiveness evaluation.
Users should be aware that there is a slight case count discrepancy between the final report for this study (Henry, Tolan, Schoeny, Zwaniger, and Kim, 2012; NCJRS 238972) and the data provided to ICPSR. The final report references 384 children participating, including 191 students/families randomized to the treatment condition. However, one family withdrew consent after the intervention (consistent with the procedure detailed in the consent form), and their data were removed from the database, leaving data on a total of 383 children, including 190 in the treatment condition. Additionally, the report indicates that there were five participating schools. However because the specific schools involved changed between the two cohorts, the Therapist Fidelity Data (Datasets 5, 7, and 9) include a variable with values for 6 schools. At any given time no more than five separate schools participated.
The final report for this study (Henry, Tolan, Schoeny, Zwaniger, and Kim, 2012; NCJRS 238972) includes a Cost-Benefit analysis. The data used in that analysis are from the SAFE Follow-up Study (discussed in detail on pages 36-46 of the final report), and are not available as part of this collection.
The purpose of the study as described by the principal investigators includes exploring network processes, and the final report for the study (Henry, Tolan, Schoeny, Zwaniger, and Kim, 2012; NCJRS 238972) includes a social network process assessment. Data pertaining to network processes are not included in this collection.
The final report for this study (Henry, Tolan, Schoeny, Zwaniger, and Kim, 2012; NCJRS 238972) refers to questionnaires completed by family providers from the community mental health centers on demographic information and their experience in conducting preventive interventions similar to the intervention implemented in SAFEChildren. Data from these questionnaires are not available as part of this collection.
Study Purpose: The purpose of this study was to evaluate the effectiveness of SAFEChildren when implemented with community providers of the family intervention and tutors recruited from the student bodies of the participating schools (the SAFE Effectiveness Trial). Specifically, the first goal was to test the effects of the SAFE Effectiveness Trial intervention. The second goal was to explore network processes within the groups and their relations to outcomes. The third was to explore pre-existing provider attitudes and process and fidelity measures, and their relations to outcome.
The SAFE Effectiveness Trial (SAFE-E) involved community providers delivering the family group intervention portion of the SAFEChildren intervention and upper grade students delivering the tutoring program. The primary target sample for this study was the children (and their parents/caregivers) enrolled in first grade at six elementary schools in disadvantaged neighborhoods in Chicago, Illinois. In addition, teachers were enrolled to provide sources of data about the children; tutors and mental health service providers were also included as participants in the study. Individual children and their families were randomly assigned to treatment or control conditions. In addition to the 1st grade participants, 278 7th and 8th graders who had been nominated by teachers were recruited to be trained and to participate as tutors for the 1st grade children in the treatment condition. Because of the principal investigators' prior involvement with Chicago area inner-city, several schools were interested in participating prior to the beginning of the project. The two community mental health agencies that were recruited were agencies that served the communities in which the schools were located.
A two-stage procedure was used for recruitment and consent of families. The participating schools provided the researchers with the names, addresses, and phone numbers of all children attending kindergarten. Initial information was sent to all families. Following these letters, program staff contacted each family by phone or home visit. The research program was explained to families and informed consent was obtained. Eighth-grade tutors were nominated by their teachers. A consent procedure similar to that of the first-grade participants was followed to secure parent consent and child assent for participation. The tutors received 40 hours of training prior to beginning the intervention and received 30 minutes of supervision each week during the course of the intervention. Providers were recruited and hired by their respective agencies, and trained by university personnel.
The study consisted of two components, a process and fidelity measurement looking at the parents, therapists and tutors who participated in the SAFEChildren program (Datasets 3-13), and an effectiveness evaluation that compared a control group of children and parents to a treatment group participating in the SAFEChildren program. Demographic and administrative data were collected for both the control and treatment groups (Dataset 1) as well as data about any twins participating in the study (Dataset 2). All interviews were conducted using laptop computers and were checked for accuracy and completeness.
The Process and Fidelity Characteristics portion of the study (Datasets 3-13) used a sample which consisted of the participants who were (1) randomly assigned to receive intervention, and (2) in attendance at the specific sessions wherein the measures were administered. Dataset 3 provides data on attendance at each session. This included the parents or caregivers of children in the treatment group (Datasets 3, 6, 8, and 11), the therapists (Datasets 5, 7, 9, and 12), and youth tutors (Datasets 10 and 13) participating in the study. Process assessments were completed at four week intervals during the interventions. Data were collected from families and group leaders for three constructs: Fidelity of Provision of Services (e.g., group activities); Group Intervention Processes (e.g., attitudes about the group leader, family involvement); and Provider Characteristics, (e.g., education, attitudes about prevention). At four sessions, tutoring supervisors completed ratings of the tutors and process measures were administered to the tutors. Tutoring supervisors rated tutors on "Quality of Relationship and Quality of Tutoring". The process measures completed by tutors comprised scales for "Tutees Relationship with the Tutor" and "About the Program".
For the effectiveness evaluation, family interviews were scheduled around convenient times for the family. The interviewer completed the child interview (Datasets 14, 17, 20, and 23) prior to interviewing the caregiver(s) or parent(s) (Datasets 15, 18, 21, and 24). If two caregivers were present and agreeable, both were interviewed. At random, 5 percent of families were contacted and re-interviewed (briefly) to assure reliability and validity. Each child's teacher was asked to fill out questionnaires on the child's behavior, social competence, and self-regulation at each wave (Datasets 16, 19, 22, and 25). Teachers were paid $10 per child for each assessment.
The sample of parents and children were participants in a family-focused group-based intervention from a 50 percent random sample of individual families with children entering first grade in six Chicago public schools. The sample of therapists was selected from mental health service agencies serving the communities in which the schools were located. The participating schools provided the principal investigators with the names, addresses, and phone numbers of all children attending kindergarten. Initial information was sent to all families. All information sent to families was translated into Spanish and included with English versions if the school served potentially Spanish speaking families. Following these letters, program staff contacted each family by phone or home visit. The research program was explained to families and informed consent was obtained. Eighth-grade tutors were nominated by their teachers. Providers were recruited and hired by their respective agencies, and trained by university personnel. There were a total of 1,178 participants in the study. 267 tutors were recruited and 15 individual therapists conducted groups. The principal investigators were able to recruit 383 subjects who completed the study, 190 were randomized to the treatment condition, and 155 consented to participate in the intervention. The Sample Data (Dataset 1) gives information on the 383 subjects who were recruited. The Twin ID Data (Dataset 2) gives identifying information for the 16 child participants who had twins in the study.
For the process and fidelity measurements, family attendance data (Dataset 3) is available for 2,058 cases, where each case represents a single parent's attendance at a single session (for up to 16 sessions). For the fidelity measurements, data are available for 61 parents who participated in session 3 (Dataset 4), 60 parents who participated in session 8 (Dataset 6), and 60 parents who participated in session 13 (Dataset 8). Fidelity data were provided by the therapists for 19 groups in session 3 and session 8 (Datasets 5 and 7) and 17 groups in session 13 (Dataset 9). Tutor Fidelity Data was gathered by supervisors evaluating 218 tutors at a total of 685 tutoring sessions (Dataset 10). For the process measurements, data are available from 86 parents (Dataset 11) at up to 3 family sessions each, resulting in a total of 184 cases. Process data on 83 children at up to 3 family sessions each were also provided by 11 therapists (Dataset 12), resulting in a total of 181 cases. Finally, process data are available from 224 tutors (Dataset 13) on 160 children at up to 5 tutoring sessions each, for a total of 906 cases. Data are not available for every participating parent, tutor, or therapist at each session because only parents who attended the sessions got the opportunity to complete process and fidelity measures; no makeup measures were administered. Also, not all 8th grade tutors or their supervisors completed tutoring process measures and the therapists were not required to complete the process assessments.
Due to non-participation and attrition among children, parents, and teachers during the span of the effectiveness evaluation, data are not available for every child in each component of the wave. The datasets corresponding to each wave of child interviews (Datasets 14, 17, 20, 23) have data on 380, 319, 308 and 318 children respectively. The datasets corresponding to each wave of parent interviews (Datasets 15, 18, 21, 24) have data regarding 383, 324, 315, and 322 children respectively. The datasets corresponding to each wave of the teacher questionnaires (Datasets 16, 19, 22, 25) have data on 379, 378, 337, and 347 children respectively.
Time Method: Cross-sectional , Longitudinal: Panel: Interval
Mode of Data Collection: computer-assisted personal interview (CAPI), computer-assisted self interview (CASI), cognitive assessment test, paper and pencil interview (PAPI), self-enumerated questionnaire, on-site questionnaire
Four waves of in-person interviews with participating children and their parents (separately).
Four waves of questionnaires filled out by each participating child's teachers.
Six waves of questionnaires for participating parents, therapists and tutors.
Reading test scores.
Description of Variables:
- The Sample Data (Dataset 1) has 27 variables regarding sample characteristics for each participating child, including assigned group (treatment vs. control), cohort, gender, ethnicity, dates of interviews of children, teachers, parents (recoded by principal investigators to number of days into the study to protect confidentiality), the child's age at each interview and the parent's age at each interview.
- The Twin ID Data (Dataset 2) has three variables: a variable indicating that the child has a twin in the study and the ID for that twin
- The Family Attendance Data (Dataset 3) has five variables regarding the attendance of parents at each group session.
- The Parent Fidelity Session 3 Data, Parent Fidelity Session 8 Data, and Parent Fidelity Session 13 Data (Datasets 4, 6, and 8) have variables about parent satisfaction with the session and parent perceptions on the extent to which specific activities were completed. Because the number of items included regarding activity completion varies between datasets, Dataset 4 has 24 variables, Dataset 6 has 20 variables, and Dataset 8 has 22 variables.
- The Therapist Fidelity Session 3 Data, Therapist Fidelity Session 8 Data, and Therapist Fidelity Session 13 Data (Datasets 5, 7, and 9) have variables on the therapist's self-report on extent to which session objectives were met, specified activities were completed, and adults and children were engaged in the session. Because the number of items included regarding session objectives and activity completion varies between datasets, Dataset 5 has 28 variables, Dataset 7 has 24 variables, and Dataset 9 has 27 variables.
- The Tutor Fidelity Data (Dataset 10) has 20 variables assessing the tutor-student relationship and the extent to which tutors adhere to tutoring guidelines and structures. It includes variables about tutor engagement and tutoring structure.
- The Parent Process Data (Dataset 11) has 61 variables assessing impressions from parents on therapist alliance/satisfaction and parent satisfaction with the program. These include: trust/positive feelings of parent, trust/positive feelings of child, therapist helpfulness to parent, therapist helpfulness to child, program helpfulness to parent, parent engagement, child engagement, and effects of program on parent/family.
- The Therapist Process Data (Dataset 12) has 51 variables assessing impressions from providers on therapist alliance/satisfaction and parent satisfaction with the program. These include: trust/positive feelings of parent, trust/positive feelings of child, therapist helpfulness to parent, therapist helpfulness to child, program helpfulness to parent, parent engagement, child engagement, and effects of program on parent/family.
- The Tutor Process Data (Dataset 13) has 21 variables assessing the tutor relationship with the first-grade student, the tutoring process, and the student's progress. These include: trust/positive feelings of child and motivation/engagement of child.
- Each of Datasets 4-13 also include calculated scales based on the questions asked. For information about scale calculation in these datasets, please see the Implementation Measurement Manual included in the Documentation PDF included as part of this collection.
- The Effectiveness Trial Wave 1 Child Data, Effectiveness Trial Wave 2 Child Data, The Effectiveness Trial Wave 3 Child Data, and Effectiveness Trial Wave 4 Child Data (Datasets 14, 17, 20 and 23) include subscales from the Behavioral Assessment System for Children: Child Self-Report Interview, and the number of correct items on four tests from the Woodcock Diagnostic Reading Battery. The Behavioral Assessment System consists of questions about child's perceptions about school and their teacher. Wave 1 and Wave 3 (Datasets 14 and 17) each have 30 variables; Wave 2 and Wave 4 (Datasets 20 and 23) also include a variable for the child's current grade, and have a total of 31 variables.
- The Effectiveness Trial Wave 1 Parent Data (Dataset 15) has a total of 1,028 variables. These include: demographics, variables assessing prenatal development, variables about the parent's current relationship with a partner, an inventory of depression symptoms, variables about the child's behaviors, the quality and safety of the neighborhood the child lives in, family relationships, parent's involvement in schools, parenting practices, parental stress, and self-reports of delinquency.
- The Effectiveness Trial Wave 1 Teacher Data, Effectiveness Trial Wave 2 Teacher Data, Effectiveness Trial Wave 3 Teacher Data, Effectiveness Trial Wave 4 Teacher Data (Datasets 16, 19, 22, and 25) each have 96 variables, including 60 variables from the Behavioral Assessment System for Children and 21 variables from the Parent Involvement in School Scale. The Behavioral Assessment System consists of questions about whether the child is disruptive, how he/she interacts with other children, interactions with other children, and how well he/she performs in school. The Parent Involvement in School Scale consists of questions about frequency of communication between teacher and parent, parent's participation in school events, the parent's attitudes about education/goals for children.
- The Effectiveness Trial Wave 2 Parent Data (Dataset 18) has a total of 787 variables. These include: demographics, variables assessing the parent's current relationship with a partner, an inventory of depression symptoms, variables assessing the child's behaviors, variables about family relationships, parent involvement in schools, parenting practices, parental stress, and self-reports of delinquency.
- The Effectiveness Trial Wave 3 Parent Data and the Effectiveness Trial Wave 4 Parent Data (Datasets 21 and 24) each have a total of 862 variables. These include: demographics, variables assessing the parent's current relationship with a partner, an inventory of depression symptoms, variables assessing the child's behaviors, variables about the quality and safety of the neighborhood the child lives in, family relationships, parenting practices, parental stress, and self-reports of delinquency.
- Datasets 14-25 also include recoded variables and scales calculated by the principal investigators. Information about recodes can be found in the SAFE Children Effectiveness Evaluation Documentation Memorandum in the Documentation PDF accompanying this collection and calculation of scales can be found in the SAFE Children Effectiveness Evaluation Scale Information section of the Documentation PDF.
Response Rates: Overall, 84 percent of eligible families consented for assessment. Of the 190 students/families randomized to the treatment condition, 155 (81 percent) consented to participate in the intervention. 98 (53 percent) attended at least 1 family group session, and 67 attended 50 percent or more of the sessions.
Presence of Common Scales:
Several likert-type scales were used. In addition, the following scales were used:
- Modified GREAT Schools and Families Fidelity Measures (Miller-Johnson et al., 2004) (Datasets 4-10)
- Adapted Fidelity of Implementation Rating System (Forgatch et al., 2005) (Datasets 5, 7, 9)
- Modified GREAT Schools and Families Process Measures (Miller-Johnson et al., 2004) (Datasets 11-13)
- Woodcock Diagnostic Reading Battery (Woodcock, 1997) (Datasets 14, 17, 20, 23)
- The Quality of Relationship Scale (Capaldi and Crosby, 1997) (Datasets 15, 18, 21, 24)
- Conflict Tactics Scale (Strauss, 1979) (Datasets 15, 18, 21, 24)
- Beck Depression Inventory (Beck, Emery, Rush, and Shaw, 1979) (Datasets 15, 18, 21, 24)
- Behavioral Assessment System for Children (Reynolds and Kamphaus, 1992) (Datasets 14-25)
- Children's Behavior Questionnaire (Rothbart, Ahadi, Hershey, and Fisher, 2001) (Dataset 15)
- CYDS Neighborhood Scale (Tolan, Gorman-Smith, and Henry, 1996) (Datasets 15, 21, 24)
- Fear of Crime Scale (Tolan, Gorman-Smith, and Henry, 1996) (Datasets 15, 21, 24)
- Family Relationships Scale (Revision of Tolan, Gorman-Smith, Huesmann, and Zelli, 1997) (Datasets 15, 18, 21, 24)
- Parent Involvement in School Scale (Eccles and Harold, 1993) (Datasets 15, 16, 18, 19, 21, 22, 24, 25)
- Parenting Practices Questionnaire (Revision of Gorman-Smith, Tolan, Zelli, and Huesmann, 1997) (Datasets 15, 18, 21, 24)
- CYDS Stress Measure (Tolan and Gorman-Smith, 1992) (Datasets 15, 18, 21, 24)
- Revised Self-Report of Delinquency (Elliott, Huizinga, and Menard, 1989) (Datasets 15, 18, 21, 24)
Extent of Processing: ICPSR data undergo a confidentiality review and are altered when necessary to limit the risk of disclosure. ICPSR also routinely creates ready-to-go data files along with setups in the major statistical software formats as well as standard codebooks to accompany the data. In addition to these procedures, ICPSR performed the following processing steps for this data collection:
- Created variable labels and/or value labels.
- Standardized missing values.
- Checked for undocumented or out-of-range codes.
Original ICPSR Release: 2013-12-20
- 2015-05-12 The ID Link Data (originally Dataset 2) were removed from the collection. As a result all other datasets were re-numbered. Datasets 3, 4, 6, 8, 10-13 were updated to remove the original ID variable, CHILDID, and replace it with the ID variable AID. All links between files were maintained. Cases corresponding to invalid CHILDID's were dropped in the Family Attendance Data (Dataset 3), Tutor Fidelity Data (Dataset 10) and Tutor Process Data (Dataset 13).
- Citations exports are provided above.
Export Study-level metadata (does not include variable-level metadata)
If you're looking for collection-level metadata rather than an individual metadata record, please visit our Metadata Records page.