"Self-diagnosis tasks" are aimed at fostering diagnostic behavior by explicitly requiring students to present diagnosis as part of the activity of reviewing their problem solutions. Recitation groups in an introductory physics class of about 200 college students were distributed into a control group and three intervention groups in which different levels of guidance were provided for performing self-diagnosis activities. We investigated how well students self-diagnose their solutions in the different interventions and examined the effect of students' self-diagnosis on subsequent problem solving in the different intervention groups. We found that in the context of an atypical quiz, while external support altered the self-diagnosis performance, the self-diagnosis score was not correlated with subsequent problem-solving performance on a transfer problem. We discuss possible explanations for our findings. (Contains 16 tables, 10 figures and 2 footnotes.)