Friday, November 11
Questionnaire Design
Fri, Nov 11, 10:30 AM - 11:55 AM
Hibiscus A
Questionnaire Design for Sensitive Topics and Confidentiality

Cognitive Testing of the Questionnaire for the Swedish School Survey on Crime (303541)

*Anna Frenzel, The Swedish National Council for Crime Prevention 

Keywords: school survey on crime, questionnaire design, cognitive interviews

The Swedish National Council for Crime Prevention (Brå)---an agency under the Ministry of Justice---is a center for research and development within the judicial system in Sweden.

One of many surveys Brå conducts is the school survey on crime among students in year nine (aged 15). The survey includes questions regarding alcohol and drug use, relationship with parents and friends, and participation in and exposure to crime. The survey started in 1995, and the ninth survey was conducted in 2015.

A revision of the questionnaire was made prior to the last data collection. The revision was performed in several steps using an internal and external reference group, and was followed by an expert review, made by Statistics Sweden, which aim was to ensure that the questionnaire collects the correct information. Thereafter followed a cognitive test by Statistics Sweden to identify problems in the interpretation of the questions and how the covering letter is perceived. The cognitive test was conducted through interviews with five students (aged 15--16). The experts aimed to find test subjects with e.g., divorced parents, who participated in and/or were exposed to crime, which smokes and/or use alcohol.

The result of the cognitive interviews showed that the test subjects were doubtful regarding the honesty of the respondents, despite the anonymity when participating in the survey. The interviews also showed that the test subjects tend to steer their answers towards what is perceived as a social acceptable answer, and the questions are interpreted subjective by the test subject’s experiences and conception of current norms. The problem with respondents giving social acceptable answers is common in all surveys containing sensitive questions, which is important to keep in mind when interpreting the data.

The presentations focus will be on the above presented problems and how these were dealt with in the final questionnaire.