The validity of the International Headache Society (IHS) classification system for college-aged students with headache was examined using cluster analysis. Undergraduate college student volunteers (N = 369) underwent a structured diagnostic interview for headaches, and the sample was divided into two subsamples for purposes of replication. A hierarchical cluster analysis (Ward's method) of the headache characteristics reported by the first subsample suggested a statistically distinct three-cluster solution, and the solution was replicated using the second subsample. It appeared that one cluster was tensionlike, while the other two were migrainelike. Nonhierarchical cluster analyses (K-means) of the cases from each subsample revealed a similar pattern of a tensionlike and two migrainelike clusters. Identical three-cluster solutions were found for the second subsample both by using cluster centers from the first subsample and by clustering the cases independently, suggesting that the cluster solution was not a random finding. The IHS classification system appears to lack adequate specificity and sensitivity for college-aged students with headache who report migrainelike symptoms. Thus, the generalizability of research results using college-aged students with headache to the adult population may be questionable.