Web Survey Bibliography

Title Using computational linguistic resources to evaluate and improve survey questions
Year 2014
Access date 10.12.2014
Abstract Purpose of the study: Writing good survey questions is a complex task and several studies have shown the importance of optimally wording survey questions. Our aim is to to explore how computational linguistics resources that are available online can help survey designers to evaluate question wording. Design/methodology/approach: A case study was conducted on a web questionnaire for international exchange students. We evaluated word frequencies in linguistic corpora and searched for synonymous wordings in the WordNET database. We made 23 different wording changes that resulted in two versions of the questionnaire, one with low-frequency and one with high-frequency wordings. To compare the two versions we conducted a split-ballot experiment on three universities. Findings: The groups that responded to the version that was improved with computational linguistics resources; we observed decreased drop-out rates from 30% to 20%. Moreover, students found the questionnaire less difficult and less words were not understood in the improved version. On the other hand, there are no significant differences between the two versions in terms of item-nonresponse, straightlinging, response times, enjoyment of completing the questionnaire and difficulty of generating answers to questions. However, for some indicators the results differ between different sub-populations (incoming – outgoing student, country, language). Originality/value: Research into designing survey questionnaires has mainly focused on structural characteristics of questions such as the type, format and order of questions and response categories. On the other hand, there is substantially less research on the linguistic aspects of words used to form a question. Moreover, the originality of the study is in the use of online computational linguistics resources in pre-testing web survey questionnaires. Research limitations/implications: The limitation of the study is that it was conducted on a student sample. Further experiment are needed on different populations and for different languages. Practical implications: The results will help improve the web application that we are developing for evaluating survey questionnaires. We hope it will help researchers to write better survey questionnaires.
Year of publication2014
Bibliographic typeConferences, workshops, tutorials, presentations
Print

Web survey bibliography - Vehovar, V. (139)

Page:
Page: