Natural language in measuring user emotions: A qualitative approach to quantitative survey-based emotion measurement

Natural language in measuring user emotions: A qualitative approach to quantitative survey-based emotion measurement

Tonetto, L.M.
Desmet, P.M.A.

Industrial Design Engineering

Industrial Design

2012-09-11

This paper presents an approach to developing surveys that measure user experiences with the use of natural everyday language. The common approach to develop questionnaires that measure experience is to translate theoretical factors into verbal survey items. This theory-based approach can impair the questionnaire’s ecological validity because items can suffer from a disparity between the abstractness of the theory and the concreteness of actual user experience. Moreover, items formulated by researchers may not be fully understood by respondents because they do not represent their natural language. In the proposed approach the items are (a) generated by respondents instead of by researchers, and (b) are domain-specific instead of general. The approach is tested in an application study in which a survey was developed to assess the affective responses of car owners to their cars. This resulted in survey with 71 items, each representing participants’ natural speech about car experiences. In an online platform, 538 people answered the survey. Reliability of data, tested with Cronbach alpha index, was 0.94, suggesting that the procedure was able to result in a survey which results are highly reliable to measure drivers’ appraisals to their cars.

measurement of emotion
survey development
car experience

http://resolver.tudelft.nl/uuid:88ba101e-dfb5-4801-88d1-98d216a4eba5

978-0-9570719-2-6

Out of Control: Proceedings of the 8th International Conference on Design and Emotion, London, UK, 11-14 September 2012

Institutional Repository

conference paper

(c) 2012 The Author(s)