Birnbaum's three-parameter logistic item response model was used to study guessing behavior of low ability examinees on the Graduate Record Examinations (GRE) General Test, Verbal Measure. GRE scoring procedures had recently changed, from a scoring formula which corrected for guessing, to number-right scoring. The three-parameter theory was used to assess (1) the effect of this scoring change on the probability of a correct response; (2) differences in the probability of correct response for each of the four item types (analogies, antonyms, sentence completion, and reading comprehension); and (3) prediction of guessing according to differences in probabilities of correct response. The LOGIST computer program was used to estimate item, person, and c-parameters. Analysis of variance indicated that differences attributable to scoring instructions were small and not significant. For three of the four item types, the mean c-parameter was 15 to 20 percent lower than what would have occurred from random guessing. For the antonym item type, however, the mean c was equal to the probability expected from random guessing. Although some issues were raised suggesting further research needs, it was concluded that item response c-parameter theory was suitable for studying guessing. (GDC)