Succes en falen van beleid en non-respons: een empirisch onderzoek naar het terugzenden van schriftelijke enquêtes door Nederlandse gemeenten

Samenvatting

To investigate survey nonresponse, questionnaire data were solicited by mail from Dutch municipal governments (N = 286) in 1987, & supplemented with information gathered on both respondents & nonrespondents from independent sources. Very successful & very unsuccessful municipalities were least likely to respond, while intermediate communities responded the most often. This curvilinear relationship between performance & nonresponse can be seen as an interaction between interest in the research topic & evaluation fears. Thus, the relationships among variables become somewhat biased, & so are corrected by poststratification weighing techniques, subsequently showing that the assumption of equal response probabilities is not met. Weighing by one or two independent variables does not lead to better results any more than does weighing by policy performance alone. Only weighing by both an independent & a dependent variable leads to acceptable results. 6 Tables, 35 References. Adapted from the source document.