This study used publicly available information to identify institutional controllable variables that influence graduation rate and to group colleges using these factors. The dependent variable for the correlation analysis was the six-year graduation rate of the fall 1993 freshman cohort (as reported by US News and World Report) at more than 1,400 bachelors-degree granting institutions. Factor analysis identified five components that accounted for 78.8 percent of the graduation rate variance: student academic ability (defined primarily by average combined Scholastic Assessment Test score, freshman retention rate, percent of freshmen from top 25 percent of their high school class, percent of applicants admitted, and number of colleges in geographic area); presence of nontraditional students (percent of part-time students, average age, percent male, percent living in student housing, student/faculty ratio); presence of disadvantaged students (percent minority, percent taking only remedial courses); institution sector (including percent of students on financial aid); and institution mission (liberal arts/comprehensive curriculum focus, doctoral degrees awarded, federal research dollars). Cluster analysis grouped 971 colleges on the five principal components but was found unsuitable for establishing an objective grouping of peer institutions for graduation rate comparisons. (Contains 12 references.) (CH)