Best-Worst Scaling of Values: More Substance than Style?

Research on response styles typically focuses on response styles being bias that requires purging from the data. In the literature several procedures have been proposed to detect response styles such as acquiescence, extreme response style and midpoint responding. The main focus has been on post hoc detection of response styles and then correcting for their presence. For instance, in measuring Schwartz’ values, the commonly accepted procedure is centering the values that are measured on rating scales before subsequent analyses are done.

A main question is whether these response styles are bias or whether they are a communication style that needs no correction. However, also communication styles might differ across cultures and thus might affect cross-national comparisons. A way to circumvent response styles might be another way of measurement. Best-worst scaling (BWS) of values might be such an approach. In best-worst scaling respondents are given number of sets of values and for each set they are asked indicate which specific value in a set they consider of highest importance to them and which one of lowest importance. After answering all sets the researcher can deduct the respondent’s value profile. A question is whether these profiles mainly contain substance or whether other response styles contaminate the profiles. To answer this question we use data from 5 nations worldwide and we link the Schwartz values profiles to data on behaviour.