In two experiments, we examined the disruptive effects of a "can't answer" response option (CARO) on equivalence formation. The first experiment was a systematic replication of Duarte, Eikeseth, Rosales-Ruiz, and Baer (1998), in which participants in a CARO group and a No-CARO group performed conditional discrimination tasks with abstract stimuli using a paper-and-pencil format for training and testing of equivalence relations. The presence of the CARO led to the nonemergence of equivalence classes. In the second experiment, participants performed conditional discrimination tasks using standard matching-to-sample training and testing procedures on a computer with CARO available only during testing. Equivalence yields were also low, with participants using CARO more on transitive and equivalence trials than on symmetry trials. The results support previous reports of equivalence disruption by nonresponse options such as CARO and suggest directions for further research. (Contains 3 tables and 4 figures.)