Abstract: Bandler and Grinder (1975) proposed the construct `most highly valued representational systems` to describe auditory, visual, and kinesthetic perceptual orientations among individuals. They set forth a predicate method for determining an individual's primary representational system (PRS). Contrary to Bandler and Grinder's construct, experimental evidence consistently supports the principle of visual dominance among humans. In order to test the effectiveness of Bandler and Grinder's constructs, 221 volunteers were screened by an alternate predicate method. 22 subjects were selected as 'highly auditory' and 19 subjects were selected as 'highly visual'. These 41 subjects were asked to give their first impressions to 75 simultaneous, audio-visual presentations of letters, numbers, and words. Twenty-five of the simultaneous presentations were discrepant. Subjects were then individually interviewed. Their verbal responses to standardized questions were tape recorded. Four questions were examined. (1) Do the verbs, adverbs, and adjectives expressed by subjects in their taped interviews yield discrete classifications of primary representational systems by Bandler and Grinder's predicate method? (2) Do the predicate- method classifications correlate with the screening classifications derived by the alternate predicate method? (3) Do predicate-method classifications indicate whether individuals will make auditory of visual responses to the twenty-five discrepant presentations of auditory and visual stimuli? (4) Will subjects classified as auditory or visual respectively resort to auditory or visual responses with greater frequency as task and stimulus difficulty are increased over the five series of discrepant stimulus presentations? Predicate tabulations indicate that the majority of individuals could not be effectively classified by Bandler and Grinder's predicate method. A point biserial correlation was performed using the auditory and visual screening classifications and the proportion of auditory predicates expressed during the experimental interviews. No relationship was found between the two methods of classification. Attentional responses were analyzed by a 2x5 repeated measures ANOVA. No significant relationships were found between predicate-method classifications and responses to the discrepant stimulus presentations. Though a significant interaction resulted, the pattern of the means across the five tasks was not the pattern predicated. There was no significant relationship between predicate-method classifications and attentional responses across increasingly difficult tasks. The major finding of the study was a visual dominance effect. Thirty-six of the thirty-seven subjects responded more frequently to the visual stimuli than to the auditory stimuli on discrepant presentations. 92% of the possible 925 responses were visual, supporting the generalizability of the visual dominance principle to an experimental paradigm in which televised linguistic symbols are presented without measuring the reaction times of subjects' responses as in previous studies. This further evidence for visual dominance in human perceptions casts doubt on the constructs of Bandler and Grinder, but reasserts the importance of future research relating conscious attentional mechanisms to the ways visual perceptions gain control of these mechanisms.