"Pick-any" measures contaminate brand image studies

Brand image measures using the typical "pick-any" answer format have been shown to be unstable (Rungie et al., 2005). In the present study, we find that the poor stability results are mainly caused by the pick-any measure itself because it allows consumers to evade reporting true associations. Using a forced-choice binary measure, we find that stable brand attribute associations are in fact present with much higher incidence (70%), thus outperforming both the measures predominantly used in industry (pick-any, 41%) and academia (7-point scale measure, 59%). Under simulated optimal conditions the forced-choice binary measure leads to 90% stability of brand-attribute associations and is therefore recommended as the optimal answer format for brand image studies.