Text analysis of open-ended survey responses: a complementary method to preference mapping

Abstract

The present study illustrates the use of computer-aided text analysis to evaluate the content of open-ended survey responses. During an in-hall test, different varieties of mayonnaise were evaluated by 165 respondents on a 10-point liking scale, with the option to freely comment on these assessments. An expert panel assessed the main sensory characteristics of the mayonnaises. Usually, preference mapping is applied to find out which sensory attributes drive consumer preferences for products. As an alternative, we will use the consumer statements to investigate this relationship. By counting words and looking at word combinations, we will use correspondence analysis to construct a visualization that is comparable to a preference map. Results from the analyses of the verbal responses are compared with the results from preference mapping. The agreement between the correspondence map and the preference map is striking, with the additional advantage being that the correspondence map is stated in terms of consumer language. The textual map is also in line with consumer understanding about creaminess. It appears that respondents find it easier to comment on their dislikes than on their likes. Some advantages and disadvantages of analyzing open-ended survey responses are also addressed.