Testing the robustness of best worst scaling for cross-national segmentation with different numbers of choice sets

The aim of the study is to showcase how cross-cultural research can take advantage of the measurement invariance of best–worst scales. The study utilises best–worst scaling (BWS) to assess the importance of environmental sustainability among other experience and credence product attributes for the purchase of wine across five countries. Three consumer segments with distinct product preferences were identified across all five countries, which differ in their relative size in each market. This case study demonstrates different analysis methods suitable for the analysis of BWS data on aggregated and segment level when respondents perform complete or partial main effects designs. Furthermore, a sensitivity analysis is conducted to explore how the reduction of individual response load affects aggregated and segmented scale estimates and errors. The results provide researchers with guidelines on the degree to which best–worst designs can be divided among several respondents to reduce survey length and respondent load, depending on the research question pursued. The advantages and limitations of the BWS method are discussed, which are important for researchers implementing BWS in food consumer research.

► Best–worst scaling produces a scale that is invariant across cultures.
► Standardised best–worst scales allow a ratio-level comparison of attribute importance.
► Segments of consumers with similar attribute importance can be identified across countries.
► Discrete choice models are a flexible method to analyse partial main effects BWS designs.
► Reducing individual choice sets hardly affects aggregated level results but impacts latent class choice model estimates.