How should you frame questions to measure user attitudes accurately? An experimental design study

Abstract

Attitudes are most frequently measured through responses to questionnaires. The validity of the results is strongly dependent on the quality of the instrument. Questionnaires employed to evaluate users' attitudes toward issues, products, or services frequently tend to have several biases yielding inflated responses. A 2 × 2 experimental design study established that rating attitudinal items on a Likert scale in a questionnaire with all statements framed in positive valence tend to produce higher attitudinal scores than when statements are balanced with both positive and negative valence. © Springer-Verlag Berlin Heidelberg 2007.