Development, Implementation, and Evaluation of a More Efficient Method of Best-Worst Scaling Data Collection

Abstract

Abstract Discrete choice experiments are used to collect data that facilitates measurement and understanding of consumer preferences. A sample of 750 respondents was employed to evaluate a new method of best-worst scaling data collection. This new method decreased the number of attributes and questions while discerning preferences for a larger set of attributes through self-stated preference “filter” questions. The new best-worst method resulted in overall equivalent rates of transitivity violations and lower incidences of attribute non-attendance than standard best-worst scaling designs. The new method of best-worst scaling data collection can be successfully employed to efficiently evaluate more attributes while improving data quality.