Evaluating the Inclusion of Words and/or Pictures in Best-Worst Scaling Experiments

Abstract Two best-worst scaling (BWS) experiments, one including only pictures and one including both pictures and descriptive words, were used to rank consumer preference for tomatoes with varying attributes in a nationally representative sample of 1,200 US residents. To the authors’ knowledge, this is the first implementation of a fully pictorial-based BWS experiment. Relative preference among pictures of tomatoes in different retailing/production scenarios were analyzed and results between pictures that did and did not include descriptive words were compared. The experiment resulted in relative rankings that were statistically different between the BWS experiments. Further analysis evaluated word associations of respondents for the tomato pictures, establishing the foundation that, at least in one instance, lack of agreement existed regarding pictorial content. The potential for varying ranking of products or attributes when pictures versus text are used to convey related meaning/messaging has implications for the marketing and conveyance of product attributes.