The prevalence of major depressive episodes (MDEs) in the general population varies across surveys. The short form of the "Composite International Diagnostic Interview" (CIDI-SF) was used less than a year apart in two representative surveys of the French population aged 15 to 75 (2004–2005). The methodology was identical except for the calculation of the 12-month prevalence of MDEs. The objectives of this article are to examine the effects of this methodological difference on prevalence estimates and the factors associated with MDEs. Materials and Methods—The Baromètre Santé and Anadep surveys are conducted using a two-stage sampling method. Households are randomly selected from the telephone directory, and then an individual is chosen from each household. The sample sizes were 16,883 and 6,498 individuals, respectively. The Baromètre survey focused on disorders experienced during the past year, while Anadep asked about lifetime prevalence and then about the past year, using the CIDI-SF in both cases. Results—The EDM prevalence estimates differed: 7.8% in the Baromètre and 5.0% in Anadep. The difference was marked for mild or moderate episodes, whereas the prevalence of severe episodes was closer, at 3.2% and 2.6%, respectively. However, the profiles remained equivalent. In both surveys, being female, widowhood, divorce, unemployment, disability, or sick leave were the circumstances most strongly associated with major depressive disorder. Discussion - This difference in prevalence estimates reflects the methodological challenges of measuring depression. Despite these differences, the factors associated with major depressive episodes remain largely the same in both surveys, allowing for the targeting of prevention policies. [author’s abstract]