A questionnaire for parents was used to examine the characteristics of their purchase attitude toward children’s shoes. They were asked the reasons for their purchase, and could choose any number of options. Responses obtained from 58 respondents were encoded into 58 binary vectors in 16 dimensions, encoding the choice of an option as a “1” and a non-choice as “0.” A cluster analysis of these vectors using the method of Baroni-Urbani and Buser as the criterion for similarity was used to classify the tendencies of the respondents in terms of choice patterns. The characteristics of the choice patterns classified in each cluster were considered from the perspective of the number of choice reasons included in the classified vector and the place of purchase, method of decision, and daily policy, which were answered in other questions.