‘Pick any’ measures contaminate brand image studies
Sara Dolnicar and John R. Rossiter
University of Wollongong
Bettina Grün
Johannes Kepler University Linz
Introduction
Consumers’ brand-attribute associations recorded with the commercially popular ‘pick any’ measure are very unstable (Castleberry et al. 1994): the average repeatability of associations on a retest is only 50%. This means that, if respondents are asked twice in a row whether they associate Coca-Cola with refreshing the first time, the likelihood that they will associate Coca-Cola with refreshing again in a second survey is only 50%.
If this low stability...