Research suggests that troubleshooting activities that require students to reflect on teacher-crafted erroneous examples; i.e., erroneous solutions to problems that correspond to widespread naïve ideas, are beneficial to learning. One possible explanation to these beneficial effects is that troubleshooting activities encourage students to test the quality of their own naïve ideas, not only the ones driving the erroneous examples, thereby improving learning. Few studies have addressed this claim, and the results are inconsistent. These studies, however, were not designed to examine the extent to which students with different naïve ideas benefit from troubleshooting activities. Here, ten 9th grade classes took part in a field experimental study that applied a pre-post-test design after finishing a unit on exponents. Students in each class were randomly assigned to a troubleshooting (114 students) or a self-diagnosis activity (112 students). Self-diagnosis activities are considered to directly nudge students to examine the quality of their own naïve ideas by requiring them to reflect on their solutions. The troubleshooting and self-diagnosis activities both capitalized on the pre-test problems. Both groups increased their proficiency in exponents to a comparable extent from the pre-test to the immediate and the delayed post-test. Troubleshooting students with different naïve ideas detected the errors in the erroneous examples equally well, and their error detection significantly and positively correlated with their self-repair of their own naïve ideas. These findings suggest that all the students benefitted from troubleshooting activities, regardless of whether their own naïve ideas resembled the ones driving the erroneous examples or not.