The cultural learning concept of Credibility Enhancing Displays (CREDs) concerns the extent to which behavioral models consistently live out their professed ideals. While researchers have suggested that past CRED exposure is an important variable for predicting who does and does not become a religious believer, it is unclear how CREDs relate to when a person rejects the religious beliefs modelled to them during their upbringing. Using a large sample of formerly believing atheists, two analyses assessed the ability of CREDs to predict the age at which an individual became an atheist. In the first analysis (n = 5,153), CREDs were positively associated with a delay in Age of Atheism, with family-level religious variables (Religious Importance, Religious Choice, and Religious Conflict) moderating this relationship. In the second analysis (n = 3,210), CREDs remained a stable predictor of Age of Atheism while controlling for demographics, parental quality, religious variables, relational variables, and institutional variables. Overall, while findings support a robust relation of CREDs to atheistic outcomes even when controlling for many other variables that influence religious transmission processes, they also highlight the importance of considering how such other variables modify the impact of CREDs on (non)religious outcomes.