Among individuals with a diagnosis of autism spectrum disorders (ASD), conversation topic preference could influence social skills in many ways. For example, an individual with advanced vocal-verbal skills, but just learning to join a conversation, might be less inclined to participate if the topic chosen is not preferred. However, commonly used preference assessment procedures have not been applied to evaluating conversation-topic preferences. Therefore, the purpose of the current experiment was to conduct three different types of assessments that varied in efficiency, the degree of certainty they allow, and clients with whom they are likely to be applicable and acceptable. In particular, we conducted a self-report preference assessment, a multiple-stimulus-without-replacement (MSWO) preference assessment, and a response restriction conversation assessment (RRCA). Each assessment identified a preferred topic of conversation, but the RRCA was the only assessment that was able to differentiate which topics would maintain a conversation. Implications for assessment and intervention procedures related to complex social skills are discussed and directions for future research are proposed.