This study aims to address potential costs of using incorrect worked examples in teaching mathematics. While such practice has been shown to be effective in educational research, previous findings in the memory literature suggest that exposure to an incorrect solution may lead students to later believe that it is correct due to increased familiarity. We designed a two-session experiment with 1-week delay in which students studied correct and incorrect worked out examples. We found only small changes in students’ ability to successfully distinguish between correct and incorrect solutions over time. Students did rate the previously studied incorrect examples as being more correct after the 1-wk delay, but this did not affect their correctness ratings of new correct and incorrect worked examples or their problem solving accuracy. We conclude that the unique nature of mathematical problem solving may protect students from the dangers of using incorrect worked examples.