Ability estimates yielded by the one- (1PL), two- (2PL), and three-parameter (3PL) models and the nominal response model (NRM) were compared with the number-right (NR) scoring model using items not susceptible to test-wiseness (NTW) and items susceptible to the ID1 test-wiseness strategy. These items were contained in grade 12 diploma examinations for social studies and chemistry. The results were compared for high-, middle-, and low-ability examinees. Differences were found between pairs of ability estimates obtained when 2PL, 3PL, and NRM scores were used in place of NR scores. The differences tended to be greater for chemistry than for social studies, and with the exception of high-ability students in social studies, for the subtest containing items with absurd options than for the subtest containing nonsusceptible test-wise items. It appears at least for the two subject areas considered in the present study, that the scoring models cannot be used interchangeably to obtain estimates of examinees’ abilities, particularly when a test contains test-wise susceptible items.