Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Marketing (Business Administration)

First Advisor

William R. Darden


An empirical assessment of the moderating effect of combinations of ethnically dissimilar shoppers and sellers on the nature of retail sales encounters is undertaken. Following the similarity-attraction paradigm (Bochner 1982; Byrne 1961), it is theorized that ethnically similar shoppers and sellers are more attracted to each other than ethnically dissimilar ones, resulting in shoppers' perceptions of sellers' behaviors and personality traits being adversely effected. Operationally, the entire model of sales encounters is moderated by combinations of ethnically dissimilar shoppers and sellers (global moderation hypothesis), or dyadic ethnic dissimilarity. A 2 x 2 experimental design is employed, with the levels of factor 1 being shoppers and sellers and levels of factor 2 being Anglo and Cajun ethnicity. To test the global moderation hypothesis, a model of sales encounters is fitted to (1) four ethnically distinct shopper-seller contrasts (i.e., samples comprised of ethnically similar and ethnically dissimilar shopper-seller combinations), and (2) an overall, ethnically heterogeneous sample (i.e., a grand sample combining the four ethnically distinct samples). Comparing goodness-of-fit for the ethnically undifferentiated sample (a single-group model) with the four ethnically distinct samples (a multi-group model) yields an increase in the goodness-of-fit, which suggests some initial support for the thesis of the global effect of moderation of combinations of dissimilar shoppers' and sellers' ethnicities. The moderation of individual relationships is then examined by comparing the pattern and/or directionality of the structural coefficients calculated on each of the ethnically distinct samples.