The advantages and disadvantages of new software for market segmentation analysis are discussed, and the application of this new, chi-square based procedure (CHAID), is illustrated. A comparison is presented of an earlier, binary segmentation technique (THAID) and a multiple discriminant analysis. It is suggested that CHAID is superior to earlier data analysis techniques such as the Automatic Interaction Detector (AID) in some important respects. CHAID uses a chi-square statistic to measure the strength of association between a given predictor and the criterion to be maximized. Because the chi-square statistic is nonparametric, CHAID requires fewer assumptions than does AID. Many of the advantages of CHAID stem from its use of a chi-square optimizing procedure, which is useful for segmenting applicant and prospective applicant pools. For college planning, the approach serves as a guide for distributing scarce resources. The example of a segmentation analysis identifies prospective students who have some chance of being admitted, based on questionnaire data. The CHAID technique is shown to have additional flexibility, both in the variety of variables that may be optimized and in the way the technique can identify multiple segments at any given step. (SW)