In our prior research, average recognition memory response bias tended to be conservative when stimuli were paintings, whereas bias for common English words tended to be liberal or neutral. Efforts to understand the mechanism(s) underlying this materials-based bias effect (MBBE) have yielded new questions but no definitive answers. Here, we report a set of studies exploring the possibility that participants respond more conservatively to paintings because they expect the novel, visually rich paintings to evoke a strong, detailed memory experience at test, whereas the more familiar, visually similar words are not expected to produce this kind of vivid recollection as often. In three studies using variations of the remember/know procedure, we found that correctly recognized paintings were more often reported as "remembered" than were recognized words. There were also parallel materials-based differences in the reported bases for "new" responses. But we did not observe the expected relationships between response bias and these subjective reports. We discuss the implications of these results for accounts of the MBBE, and the more general issue of the role of stimulus materials in recognition memory response bias. (PsycInfo Database Record (c) 2022 APA, all rights reserved).