Objective: As much of adult ADHD diagnosis is based on self-report, it is imperative to examine accuracy of self-reported symptoms; yet few ADHD measures include validity scales. We examined the effect of ADHD malingering on the Connors Adult Attention Rating Scales (CAARS), including a recently developed embedded infrequency measure (CII). Method Undergraduates were randomly assigned to malinger ADHD to access stimulant medication, malinger ADHD to receive extended time accommodations, or give best effort. They were given a large battery of self-report and neuropsychological tests, including the CAARS. Results: Groups scored significantly differently on all CAARS subtests, except D; they also scored significantly differently on CII. Follow-up tests showed that medication simulators scored higher than the other two groups on A, and both simulator groups scored higher than best effort on B and C, while all three groups were significantly different (medication higher than extra time higher than best effort) on E, F, G, H, and CII. 57% of medication malingerers, 25% of extra time malingerers, and 7% of best effort scored at T > 80 on scale E (which the CAARS manual identifies as potentially indicative of overreporting) (p < .005); 31% of medication malingerers, 12% of extra time malingerers, and none of best effort scored high on CII (p < .05). Conclusion: Results show that individuals can easily malinger high scores on self-reported ADHD scales, but that use of indices for noncredible report (CAARS T>80 or the CII) can help to identify at least some of those who are malingering, with reasonable specificity.