Analysing the test quality of an online questionnaire - a tool to evaluate the target groups´ (men who have sex with men) reception and benefit of an HIV/STI prevention campaign of the Deutsche AIDS-Hilfe e.V.

Abstract

Kein Abstract vorhanden

Introduction: In Germany, gay men and other men who have sex with men (MSM) are considered to be a key population group for HIV and other Sexually Transmitted Infections prevention measures. The Deutsche AIDS-Hilfe e.V. runs a campaign (“ICH WEISS WAS ICH TU“) which is evaluated regularly. The survey of 2014/15 pursued the main goals to evaluate the online appearance of the campaign, in particular the online tool users´ perception of the tools and their benefit in terms of HIV/STI prevention. Quality analysis of the questionnaire was still pending and was the scope of this thesis. Methodology: The three main criteria of test quality analyses are objectivity, reliability and validity. These criteria were analysed in detail. Guidelines for objectivity and content and face validity analyses were considered, hence a qualitative analysis line of argument emerged. Reliability and construct validity analyses were conducted using inference statistical methods. A few additional quality criteria were analysed such as economic efficiency and ethical appropriateness. Results: The criteria for objectivity can be rated to be of high standard. Reliability had diverse outcomes. The scales where users of the online tools gave feedback about their perception showed high values of reliability. The scales for benefit in terms of HIV/STI prevention had only moderate to low values of reliability. Content validity showed good theoretical fundamentals, however construct validity could not confirm that the items indeed capture what was intended. Conclusion and recommendation: The goals of a survey define which quality criteria are of greater importance. Hence, the part of the survey where users rated the online tools can be considered to be of appropriate quality. Content validity and reliability are of adequate quality and, as this part does not aim to display a theoretical construct, one does not have to focus on the poor results of construct validity analysis. However, it would be advisable to exclude or revise items that proved to be weak or redundant. The other part of the survey is based on theoretical assumptions therefore the rather poor results of construct validity have a severe impact on quality. It became obvious that the theoretical grounding and the items needed to be revised.