The effects of superficial gender-related item wording changes on the performance of male and female examinees were studied through mathematics; discrete English items; and an English passage created in neuter, male, and female gender versions. Units of items were administered to randomly equivalent samples of about 250 examinees taking American College Testing examinations. A log-linear analysis was conducted using four factors: (1) item gender; (2) examinee sex; (3) examinee ability; and (4) response. No items were found to have sex bias, which is interpreted as the interaction of item gender with examinee sex. Mathematics items did not have varying difficulty or discrimination in the different gender versions. English items, while without varying discrimination levels in the different versions, had a dissimilar difficulty level in the gender-related versions. These differences were not systematic. Eight tables are presented, and four appendices provide the item versions. (SLD)