Health statistics consistently show women to have higher rates of morbidity than men, particularly for mild conditions (Verbrugge 1976, 1985, 1989). One of the hypothesized explanations of this relationship is that women are more willing than men to report their physical and mental health problems. Although this explanation is inconsistent with most of the available evidence (see below), it needs to be taken very seriously because it raises fundamental questions about the observed gender difference in morbidity and, more generally, the techniques used in data acquisition. In a recent paper in this journal, Ritchey, La Gory, and Mullis (1991) look at the reports of health risks and 24 physical symptoms among 72 homeless men and 28 homeless women. According to Ritchey et al., homelessness is "a natural laboratory" because homeless men are at much greater risk of illness and, as a consequence, "gender differences in reported symptoms should be reversed, attenuated, or erased" (p. 45). Their position is that if women report more symptoms than men, "it would suggest that the higher female morbidity detected in all types of health surveys is influenced greatly by reporting bias" (p. 35). Ritchey et al. find that the number of symptoms reported by homeless women (x =7.86) was significantly greater than the number of symptoms reported by homeless men (x = 5.19). In their final analysis they report that they "isolate the direct effect of sex" by statistically controlling the risk factors related to the experience of physical symptoms (p. 44). The estimated effects of these statistical controls modestly increase the gender difference in the number of symptoms from 2.67 to 2.88. Ritchey et al.