We experimentally investigated the influence of context-based biases, such as prestige and popularity, on the preferences for quotations. Participants were presented with random quotes associated to famous or unknown authors, or with random quotes presented as popular, i.e. chosen by many previous participants, or unpopular. To exclude effects related to the content of the quotations, all participants were subsequently presented with the same quotations, again associated to famous and unknown authors, or presented as popular or unpopular. Overall, our results showed that context-based biases had no, or limited, effect in determining participants’ choices. Quotations preferred for their content were preferred in general, despite the contextual cues to which they were associated. We conclude discussing how our results fit with the well-known phenomenon of the spread and success of misattributed quotations, and we draw some more general implications for cultural evolution research.