Eye movement analysis should provide a basis for differentiation of humans in terms of developed preference for certain kinds of visual presentations over other kinds. A study was conducted to examine this position. A total of 40 children from grades 8 through 12 participated as subjects. Each subject was presented with a synchronous audio-visual paired-associate learning situation, with one group receiving pictorial representation and the other receiving print representation: in the picture condition, the subject viewed a picture of the objects, and in the verbal condition, a printed sentence identifying the objects. A pretest and posttest evaluated by means of an eye camera the subject's use of his eyes, recall tests were given at the end of each learning trial, several standard ability measures were administered, and demographics were obtained. Results showed that pictorial supported audio-visual associative learning is superior to printed word supported learning, and that adding eye movement variables to prediction equations using more conventional predictors greatly increases the amount of variance accounted for. (SH)