False positives arise when people without disorders are diagnosed as having disorders. Various approaches for avoiding false positives have been suggested. This review critically assesses the roles of zones of rarity, the threshold problem (the problem of determining the boundary of disorder in cases that shade into normality), and the Diagnostic and Statistical Manual of Mental Disorders (DSM) criterion that requires that a disorder cause clinically significant impairment or distress (the harm criterion). The lack of zones of rarity in much of psychiatry gives rise to the threshold problem. The DSM harm criterion is frequently presented as offering a solution to the threshold problem. However, I argue that the harm criterion cannot offer a general solution to the threshold problem, as harm is not always correlated with the intensity and frequency of symptoms. Still, the harm criterion is essential to ensure that people who are merely different are not diagnosed as having a disorder. The threshold problem can be addressed by selecting symptom-based cut-off points to distinguish between disorder and normality. These cut-off points are frequently arbitrary in the sense that they often reflect no natural division between disorder and normal, but they may be more or less wisely chosen. Where possible, the thresholds should be set so that the advantages of diagnosis can be expected to outweigh the disadvantages.