The literature on the efficacy of eye movement desensitization and reprocessing (EMDR) for treating depression is heterogeneous due to research design, quality issues, and trials methodology. The current meta-analysis seeks to examine EMDR for depression with the aim of answering the aforementioned limitations. Thirty-nine studies were included for analysis after a review of the relevant literature. Univariate meta-regressions were run to examine dose-response and the effect of moderating variables. Subanalysis for primary and secondary depression showed a large, significant, and heterogeneous effect-size estimates, where EMDR significantly improved symptoms of depression in contrast to all control types. At post hoc, data were reexamined and a significant and large, yet heterogeneous, effect-size estimate emerged between the EMDR and control arm after the removal of two outliers [Hedges' g = 0.70, 95% CI =0.50–0.89, p-value < .01, I2 = 70%, K = 37]. This is the first meta-analysis examining for the effect of EMDR comparing to various control modalities on depression with dose-response. We found (a) that studies were balanced at onset in terms of depression severity, and (b) a large and significant effect of EMDR on depression at the end of trials. Additionally, the significance of the aggregate effect-size estimate at the end of trials was unchanged by the intake of psychotropic medications, reported demographic variables, or EMDR methodology.