Current pscyhometric assessments, which are based on test score predictors providing information only on products of performance, fail to account for satisfactory amounts of variance in academic achievement or other criterion measures of interest. To corroborate and extend previous work on information processing measures, by examining the differential predictive power of a process based on eye movement data compared to test score information, two studies were conducted. The eye movements of 39 college students (19 in the first study; 20 in the second study) were recorded during the solution of a set of 3 x 3 figural analogies from the Advanced Progressive Matrices (APM). A stimulus was presented on 35 millimeter slides with eight possible responses. The subject's head was positioned against chin and forehead rests. A calibration procedure preceded eye movement data collection. An analysis of the results from both studies showed that the process model based on eye movement data yielded greater prediction of academic achievement than did models based on test score data. Indices reflecting the amount of information that can be processed at a given time, and the allocation of processing resources to activities involving rule application accounted for significant amounts of variation in grade point average (GPA). Specifically, individuals earning relatively high GPAs were characterized as processing information in relatively large units with little redundancy, and expending a small percentage of their total processing resources on rule application activities. (BL)