A market segmentation analysis was conducted on students at a large mi
dwestern urban university using two forms of hierarchical cluster anal
ysis on student characteristics: an agglomerative procedure using a ma
tching-type association measure and a divisive chi-square-based automa
tic interaction detection (CHAID). The resulting segments were compare
d for their ability to distinguish among students according to six sat
isfaction scales and measures of students' priorities for college stud
y derived from a general satisfaction survey. As expected, the CHAID c
lusters discriminated better among students according to their several
measures of satisfaction, one of which was the criterion variable for
the analysis. However, both procedures produced differences across on
ly two of six satisfaction scales. The matching-type measure clusters
resulted in significant differences on 11 of 18 college study priority
items compared to only 6 of 18 for the CHAID clusters. Final discussi
on describes the usefulness of market segmentation strategies for plan
ning, evaluating, and improving academic and student support programs.