Strengthening Washington DC Families (SWFP) Project, 1998 - 2004 (ICPSR 34425)

Version Date: Dec 10, 2012 View help for published

Principal Investigator(s): View help for Principal Investigator(s)
Karol Kumpfer, University of Utah; Denise Gottfredson, University of Maryland

Version V1


The Strengthening Washington DC Families (SWFP) Project examined the effectiveness of an evidence-based prevention program implemented on a sample of 715 families across mulitple settings in an urban area. The study area also included suburban Maryland. SWFP was set up as a true experimental design with families being randomly placed into one of four treatment conditions:

  • child skills training only
  • parent skills training only
  • parent and child skills training plus family skills training
  • minimal treatment controls

Entire families were assigned to one of the four treatment conditions. Data were collected from all family members who participated in the program. Thus the individual data files contain more than 715 records. The parent file contains 796 cases and the child file contains 961 cases.

The Strengthening Families Program is based on cognitive-behavioral social learning theory and family systems theory targeting elementary school-aged children. In this program parents receive training in parenting skills, children receive training primarily in social skills, and families receive family skills training. The aim of the program is to effectively reduce parent, child, and family risk factors for substance use and delinquency.

Kumpfer, Karol, and Gottfredson, Denise. Strengthening Washington DC Families (SWFP) Project, 1998 - 2004. Inter-university Consortium for Political and Social Research [distributor], 2012-12-10.

Export Citation:

  • RIS (generic format for RefWorks, EndNote, etc.)
  • EndNote
United States Department of Health and Human Services. National Institutes of Health. National Institute on Drug Abuse. (DA010825)

Users are reminded that these data are to be used solely for statistical analysis and reporting of aggregated information and not for the investigation of specific individuals or facilities.

Inter-university Consortium for Political and Social Research
1998-11 -- 2004-04
1998-11 -- 2002-08 (Pre-test), 1999-02 -- 2002-10 (Post-test), 1999-08 -- 2003-04 (6 month follow-up), 2000-08 -- 2004-04 (18 month follow-up)
  1. Entire families were asked to participate in the program. For this study family was not limited to just include biological parents. Family could include foster parents, step-parents, boyfriends or girlfriends of a parent living with the child, grandparents, or other relatives who lived in the home and who helped take care of the child(ren). To be eligible families needed to have a child who was between the ages of 7 and 11. Another requirement was that participants needed to be able to read, speak, and understand English. Family members with severe mental disorders or who required salient treatment were excused from participation since it would be hard or inappropriate to include them.

  2. Minimal disclosure protections were required to protect the confidentiality of the data files. On both the parent and child files two measures were enacted. The first was to de-identify the location of each site. The second was to collapse the race/ethnicity variable into fewer categories. Nothing else was done on the child file; however, on the parent file specific medications are asked about on question 15. These have been grouped together into one value.

  3. On both files there are instances where some demographic information, such as gender, may have changed across the four waves of data collection. In the case of the parent file this is alright due to the fact that the responder may have changed from one wave to the next. For example, a mother may have been the respondent in wave 1 but then in wave 2 the father completed the survey. However, for the child's file this should not be the case. The same child is represented across all four waves. If there is a discrepancy one of the entries is in error. This could be due to the fact some of the children were quite young during their participation.

  4. There is only one version of the questionnaire that parents were asked to complete. However, there are two versions of the child's questionnaire. As a general rule children who were 8 years old or younger completed the version shorter in length. But there is only one data file for children. The variable order in the file follows the longer version. Variable labels begin with the question number given in the older child's questionnaire. When a question was also asked on the younger child's questionnaire the variable label contains that question number in parentheses directly following the first question number. For the most part the question text found in the codebooks comes from the older child version of the questionniare. In some instances the younger child version contained additional text prefacing the questions that was included in the question text.

  5. For the child's file, the original file deposited contained system missing data for not applicable cases in the case where the child did not participate in that wave of the data collection or the younger child version of the questionnaire was used. These blank cases were recoded into specific missing values. Those cases where participation in the wave did not occur were recoded first followed by those who filled out the younger child questionnaire.

  6. The variable, UNQFAMCD, provides a unique family identification number. It is an ID variable available in both files to link the two data files together.


Data for this study were collected in four waves. Each participant completed the questionnaire prior to the beginning of the training program. After the training ended participants were asked to complete the questionnaire again. Follow-up occurred at 6 and 18 months after the training ended.

The same questionnaires for parent and child respectively were used for all four periods of data collection. Thus the variable names repeat across each data file. The only difference being a number embedded in each variable name to designate which wave of data collection the variable belongs to.

Most of the surveys were completed on-site where the training occurred. In some cases, when needed, they were completed at home or an agreed upon public location. The surveys were completed in a small group setting. A research assistant was present to assist individuals fill out the questionnaire.

During registration in cases where the family had more than one child participating in the program the parents were able to identify the child with the most problematic behavior as the primary or target child.

This study utilized a convenience sample. Five organizations located around the Washington, DC area were selected to implement the Strengthening Families program. One of the five was a pre-release center which targeted families of incarcerated parents. The other organizations recruited families through multiple avenues such as health fairs, shopping malls, and knocking on doors. Culturally matched site coordinators also worked with schools and churches to recruit families.

The program targets youth who are at-risk for substance abuse or conduct problems. However, this study defined a youth being at-risk by their residence in an at-risk neighborhood.

The participating parents were predominantly African American (73 percent) and female (85 percent) at the time of the pre-test. The children participating in the pre-test were about 69 percent African American and 47 percent female. None of the demographic characteristics examined differed significantly by experimental condition.

Longitudinal: Cohort / Event-based

The parent file contains, with the exception of one variable, categorical variables that predominantly ask for a frequency (using multiple scales) of an occurrence. Major sections in the parent file include parent and child demographics, child behavior, parent and child substance use, parent's emotional well-being, self assessment of parenting skills, family dynamics, and discipline of children.

The child file contains all categorical varaibles. The questions are generally either True/False or ask for the frequency of an occurrence. Some of the major groups of questions include demographics, friends, substance use, social skills, impulsiveness, and parents/guardians.

During the recruitment phase approximately 1,400 families were approached for participation. Of those about 51 percent (715) registered for the program and completed the pre-test questionnaire.

A total of 796 parents and 961 children completed the pre-test questionnaire. The attrition rate for both groups was very similar. The post-test had about 65 percent in both groups complete a questionnaire. At a follow-up visit 6 months later approximately 54 percent of parents and children completed a questionnaire. At the final follow-up visit at 18 months after the program finished only 19 percent of the respondents completed a questionnaire.

Overall averaged across both groups about 25 percent completed only the pre-test questionnaire. Approximately 27 percent of the respondents completed 2 and 34 percent completed 3 questionnaires. There were about 14 percent of the respondents from both parents and children who completed all 4 questionnaires.

  • Family Environment Scale (Moos)
  • Parent Observation of Children's Activities (POCA) (Kellam)
  • Kumpfer Parenting Skills Scale (Kumpfer)
  • Social Skills Rating System (Gresham and Elliott)
  • Depression Inventory (Beck)
  • What About You? Survey (Gottfredson)



2018-02-15 The citation of this study may have changed due to the new version control system that has been implemented. The previous citation was:
  • Kumpfer, Karol, and Denise Gottfredson. Strengthening Washington DC Families (SWFP) Project, 1998 - 2004. ICPSR34425-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2012-12-10.

2012-12-10 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:

  • Standardized missing values.
  • Created online analysis version with question text.
  • Checked for undocumented or out-of-range codes.


  • The public-use data files in this collection are available for access by the general public. Access does not require affiliation with an ICPSR member institution.