Toward respondent validation of travel behavior segmentation results using a quantitative online self-assessment survey

Segmentation is widely applied in quantitative travel behavior research to categorize heterogeneous populations into more homogeneous subgroups to create a more nuanced understanding of travel behavior and facilitate targeted policy and planning. However, without proper validation, identified segments may reflect artifacts of the data or the statistical method used rather than meaningful mobility patterns, limiting their practical use. Respondent validation, which assesses whether individuals recognize themselves in the segments they were assigned to, has received little attention in quantitative travel behavior research. This paper introduces a quantitative online self-assessment survey as a method to pursue respondent validation of segmentation results. Using a case study in Munich, Germany, we evaluate whether respondents can comprehend, differentiate, and identify with segment profiles, how well their self-assessments align with statistical assignments and membership probabilities, and what reasons they provide for mismatches. The results indicate that the profiles of segments created as part of our method were generally perceived as appealing, comprehensible, and distinguishable, though only a minority of respondents fully identified with a single segment. ... mehrExact matches of self-assessments and statistical assignments to segments were rare, but self-assessments correlated positively with membership probabilities, suggesting that probabilistic segmentations may be better suited for quantitative respondent validation than deterministic assignments. To reveal potentials for refinement of the segments and their descriptions, our method lastly enabled respondents to attribute discrepancies between self-assessments and statistical assignments to specific segment characteristics. Factors specific to our case study limit the generalizability of our findings. Overall though, the findings suggest that our method can provide systematic insights into respondents’ evaluations of segmentation results and can thus contribute to the validation of segmentation results, especially strengthening their credibility, transparency, and practical relevance.