Spatially-Referenced Choice Experiments: Tests of Individualized Geocoding in Stated Preference Questionnaires

Maps in stated preference surveys rarely identify the location of respondents’ homes. This standard approach is grounded in the assumption that respondents are aware of their exact household locations relative to mapped policy effects, and hence possess sufficient understanding of spatial relationships to support well-informed preference elicitation. The validity this assumption is rarely if ever tested. This paper evaluates this nearly universal practice of generic policy-area mapping in choice experiments. This is compared to a more information-intensive alternative in which individualized maps pinpoint the location of each respondent’s household relative to policy effects. The latter approach requires a unique map to be generated for each respondent. Methods and results are illustrated using an application to riparian land restoration in south coastal Maine. Comparison of the results from these two approaches illustrates the implications of stated preference survey design that provides additional cartographic detail, and suggests the potential limitations of generic policy area maps.