The present study investigated the effects of the questionnaire length upon response quality. In study 1, we manipulated the location of items in a questionnaire by putting the identical items either at the first part of the questionnaire or at the la...
The present study investigated the effects of the questionnaire length upon response quality. In study 1, we manipulated the location of items in a questionnaire by putting the identical items either at the first part of the questionnaire or at the last. The qualities of responses to items at different locations were measured using several statistics (individual entropy, Lempel-Ziv complexity, reliability, respondent`s likelihood based on IRT) which were expected to reflect the response quality. In study 2, we tested whether the item location effect is also observed in real data. The results of study 1 show that individual standard deviation, individual entropy, Lempel-Ziv complexity decrease as the same scales were placed towards the end of the questionnaire. But the reliability and response to filtering questions showed no significant difference. The length of answers to open questions went shorter as the questions were placed towards the end. In study 2, we could find that the individual entropy and the Lempel-Ziv complexity decreased and maximum length of straight response increased in the long survey in which the median duration was 51 minutes.