Research DEFECT II - Design effects in complex surveys
In the previous DEFECT study we tried to separate the estimation of interviewer, design and method effects in complex samples. These effects were studied by an interpenetrated design with five nationwide surveys. Topic of the surveys was fear of crime, measured and explained by a rational-choice model.
The follow-up study DEFECT II had three goals:
- Detailed analysis of the auxilliary data collected in DEFECT
- Collecting additional 2-wave data for a panel of the DEFECT CATI survey
- Qualitative re-interviews of respondents with model inconsistent answers in the CATI-survey
Detailed analysis of the auxilliary data collected in DEFECT
DEFECT consists of five surveys (n=1200 each) with more than 136 variables. For each case up to 400 additional variables are available, depending on the interview mode. Since information on contacts (paradata), data on the environment and interviewer ratings are available even for nonrespondents, a huge number of interesting hypotheses on nonresponse and fear of crime will be examined and tested.
Collecting additional 2-wave data
Very little is known on long term stability and validity of fear of crime measures. Therefore, the CATI respondents of the DEFECT study were recontacted after one and two years after the initial interview. This 3 wave panel will be used to estimate a Wiley-model to separate stability and reliability.
Respondents with model inconsistent answers in the CATI panel survey were recontacted and interviewed with qualitative interview techniques. 30 respondents were selected by different criteria for model inconsistent answers. In each case a rational explanation for the deviant reponses were found. This research design yielded a unique combination of quantitative residual analysis with qualitative interviews.
Principal investigator: Rainer Schnell
Data management: Tobias Bachteler
Project coordination: Helene Vennigen
Qualitative interviews: Christiane Sander
German Research Foundation (DFG)
Available on request to the principal investigator
The DEFECT questionnaire is availabe via the download page.