This study aims to address potential costs of using incorrectworked examples in teaching mathematics. While suchpractice has been shown to be effective in educationalresearch, previous findings in the memory literature suggestthat exposure to an incorrect solution may lead students tolater believe that it is correct due to increased familiarity. Wedesigned a two-session experiment with 1-week delay inwhich students studied correct and incorrect worked outexamples. We found only small changes in students’ ability tosuccessfully distinguish between correct and incorrectsolutions over time. Students did rate the previously studiedincorrect examples as being more correct after the 1-wkdelay, but this did not affect their correctness ratings of newcorrect and incorrect worked examples or their problemsolving accuracy. We conclude that the unique nature ofmathematical problem solving may protect students from thedangers of using incorrect worked examples.