For the last four decades, researchers have developed methods of quantitatively assessing the acceptance of rape myths – cultural tropes that explain and justify sexual violence. From their initial conceptualization, rape myths were thought to exist at a cultural level and to contribute to a broader ideology supportive of sexual violence. However, prevailing statistical methods for modeling rape myth acceptance are inconsistent with this theoretical understanding. Here, I argue that rape myth acceptance can be conceptualized as a network of discursive expressions – that is, a person’s current willingness to voice these cultural myths, not necessarily an indicator of any persistent belief. If this view is correct, network models should be a better representation of rape myth acceptance than common cause models, and people’s endorsements of rape myths may be unstable over time. Analyzing data from twelve previous studies with quantitative measures of rape myth acceptance (total N = 5,882), I find support for this understanding of rape myths, specifically that (1) network models of rape myth acceptance generally outperform factor models; (2) rape myth acceptance positively correlates with instability over time; and (3) networks of rape myth acceptance are more strongly connected when related attitudes, such as sexism, are weaker. These results are broadly inconsistent with approaches that cast rape myth acceptance as persistent individual beliefs and suggest new avenues for research and practical interventions.