The purpose of this study is to evaluate the evidence that cognitive behavioral therapy improves exaggerated appraisals of threat among people with a primary anxiety disorder. The current meta-analysis of 19 randomized clinical trials (1594 participants) tests whether cognitive behavioral treatments for anxiety disorders improve threat reappraisal to a greater extent than wait-list control conditions, non-cognitive behavioral treatments, and active control conditions. A statistically significant and large cumulative effect (g=0.76, 95% CI: [0.51–1.01], p<0.01) indicated that cognitive behavioral therapy produces greater improvements in threat reappraisal relative to a comparison condition. Heterogeneity of the effect was high (I²=80.92%). A moderator analyses showed this effect was significant when compared to waitlist control conditions (g=1.08) non-cognitive behavioral treatments (g=0.37), and active control conditions (g=0.29). The effect of cognitive behavioral treatment on threat reappraisal was greater for studies on panic disorder (g=1.06) relative to social anxiety disorder (g=0.60) and generalized anxiety disorder (g=0.35). Findings from this meta-analysis support cognitive models of anxiety, which assert that cognitive behavioral therapy improves people's ability to make more realistic appraisals of threat.