Not liking the Likert? A Rasch analysis of forced-choice format and usefulness in survey design

Abstract

We have less time and focus than ever before, while the demand for attention is increasing. Therefore, it is no surprise that when answering questionnaires, we often choose to strongly agree or be neutral, producing problematic and unusable data. The current study investigated forced-choice (ipsative) format compared to the same questions on a Likert-type as a viable alternative. An established motivation questionnaire was administered in two versions, forced-choice and Likert-type, to 1088 first-year engineering students. Descriptive, non-parametric statistics and Rasch measurement models were applied to assess usefulness, validity and reliability. Results: The ipsative version had a higher response rate, less missing data, and the motivations emerged more clearly. Evidence for the reliability and validity of the forced-choice version was excellent. The forced-choice format is recommended as an alternative to the Likert types when collecting human or social survey data.