The first phase of a study to predict retention and withdrawal among disadvantaged students at an urban commuter institution is described. Variable selection for the study was guided by Tinto's (1975) and Bean's (1982) models. Pilot testing was undertaken to estimate some psychometric characteristics of Pascarella and Terenzini's institutional integration scales (that were developed to operationalize Tinto's model) for these students. Longitudinal tracking was undertaken with freshmen enrolled in a developmental studies program at the City University of New York. Pilot testing appears to suggest the factorial validity of the institutional integration scales for this group of students. The results revealed suspected reading comprehension problems and a possible halo effect, but when several items were revised and retested, the scales showed improved characteristics. Appended materials include a diagram of Tinto's conceptual schema for college withdrawal; a diagram of Bean's synthetic causal model of student attrition; a list of background, organizational, environmental, and outcome/attitudinal variables; and the Pascarella/Terenzini Institutional Integration Scales, modified versions. (SW)