Impact of questionnaire length and complexity on survey time

Comparison of two business web questionnaire versions

Abstract

In case of web surveys, researchers struggle to achieve satisfactory response rates and to collect data of good quality. Those two problems could be mitigated by a thorough approach to the questionnaire design and by changing questionnaire elements that could have an impact on the questionnaire length and complexity levels. In the paper, the questionnaire length and complexity are quantified by observing survey times or the times that respondents need to complete the provided questionnaire. The questionnaire length and complexity levels have an impact on survey times and consequently on response rates. The aim of the paper is to investigate how different questionnaire lengths and complexity levels influence survey times. The selected questionnaire length and complexity determinants are observed using the case of two different questionnaire versions. Those questionnaire versions have been used in a business web survey of Croatian enterprises. The results have shown that on average the long questionnaire version is twice longer and 142% more complex than the short questionnaire version. However, due to the interaction effect of length and complexity determinants, the long questionnaire version has on average 119% higher questionnaire length and complexity level than the short questionnaire version. The analysis has shown that this increase of the questionnaire length and complexity had a different impact on enterprises with different characteristics. Furthermore, the increase of the questionnaire length and complexity level did not lead to a proportionate increase in survey time at the overall enterprises level.