engleski

Impact of Different Questionnaire Design Characteristics on Survey Response Rates: Evidence from Croatian Business Web Survey

This paper aims to determinte whether or not certain questionnaire designs could lead to higher response rates compared to other questionnaire designs. To this end, the data from conducted business web survey on a sample of Croatian enterprises was used. In the survey nine different questionnaire designs were randomly distributed across the sample of enterprises. The difference between the questionnaire designs can be found in the number of questions that have been shown to respondents per questionnaire screen and in the kind of provided pictures. The conducted analyses have shown that, when all enterprises together are observed, questionnaire designs in which all questions were shown immediately to the respondents with “negative” pictures included and where questionnaire is divided into logical groups of questions with “positive” pictures, the highest response rates in the future business web surveys could be achieved. Thereby, the recommendation of questionnaire designs for enterprises stratified according to their size, legal form, main activity and location of their headquarters are also provided.

Business web survey ; Croatian enterprises ; log-linear model ; Pearson’s chi-square test ; response rate ; questionnaire design

This work has been fully supported by the Croatian Science Foundation within the project STRENGTHS (project no. HRZZ-IP-2013-11-9402).