Using a new method for studying the development of false-memory formation, we examined developmental differences in the rates at which 6-, 7-, 9-, 10-, and 18-year-olds made two types of memory errors: backward causal-inference errors (i.e. falsely remembering having viewed the non-viewed cause of a previously viewed effect), and gap-filling errors (i.e. falsely remembering having viewed a script-consistent event that was not actually witnessed). Previous research suggests that backward causal-inference errors are supported by recollection, whereas gap-filling errors are supported by familiarity. We hypothesized that age differences in these errors would parallel the developmental trajectories of these processes. As predicted, age-related increases in backward causal-inference errors were observed, while gap-filling errors were age-invariant, suggesting that recollection-based memory distortions increase with age while familiarity-based memory distortions are relatively stable from middle childhood through adulthood.