ABSTRACTUsing data collected from a retrospective study of the records of 110 patients diagnosed as having convergence insufficiency, a model has been developed which allows the classification of a patient into one of two categories of success (total or partial/none) on the basis of an initial examination. The subjects were treated with orthoptics and classified into either the total success group (no objective or subjective visual difficulties) or the partial/no success group. Stepwise discriminant analysis was used in the data reduction phase in which 6 of 14 possible variables were selected which best explained the variation in the degree of success. These six variables were: (1) the AC/A ratio; (2) the negative vergence blur value at 40 cm; (3) the recovery value of the positive vergences at 6 m; (4) the break value of the positive vergence at 6 m; (5) the amplitude of accommodation; and (6) the frequency of the deviation. Subjects were more likely to be successful when 1, 2 and 3 were high, 4 and 5 were low, and the deviation was latent. Discriminant analysis using these six variables allowed 25 of 33 (76%) of the patients to be correctly classified in terms of their success. The application of the model for classifying patients with convergence insufficiency is presented and briefly discussed.