•Virtual reality exposure therapy shows a large effect size compared to control conditions.•Virtual reality exposure therapy did not show significant differences from in vivo exposure therapy.•Effect sizes for disorder types were relatively consistent. Trials of virtual reality exposure therapy (VRET) for anxiety-related disorders have proliferated in number and diversity since our previous meta-analysis that examined 13 total trials, most of which were for specific phobias (Powers & Emmelkamp, 2008). Since then, new trials have compared VRET to more diverse anxiety and related disorders including social anxiety disorder (SAD), posttraumatic stress disorder (PTSD), and panic disorder (PD) with and without agoraphobia. With the availability of this data, it is imperative to re-examine the efficacy of VRET for anxiety. A literature search for randomized controlled trials of VRET versus control or in vivo exposure yielded 30 studies with 1057 participants. Fourteen studies tested VRET for specific phobias, 8 for SAD or performance anxiety, 5 for PTSD, and 3 for PD. A random effects analysis estimated a large effect size for VRET versus waitlist (g = 0.90) and a medium to large effect size for VRET versus psychological placebo conditions (g = 0.78). A comparison of VRET and in vivo conditions did not show significantly different effect sizes (g = −0.07). These findings were relatively consistent across disorders. A meta-regression analysis revealed that larger sample sizes were associated with lower effect sizes in VRET versus control comparisons (β = −0.007, p <  0.05). These results indicate that VRET is an effective and equal medium for exposure therapy.