--- - |2+ Background Anxiety disorders are associated with an overactive action monitoring system as indexed by a larger error‐related negativity (ERN). This study tests whether ERN magnitude changes following treatment, predicts response to treatment, and varies by treatment type. Methods The sample included 130 youth (9–14 years): youth with an anxiety disorder (ANX; n = 100) and healthy control (HC; n = 30) youth with no lifetime DSM‐IV disorders. ANX youth were randomized to either a manualized cognitive‐behavior therapy (CBT) or a comparison child‐centered therapy (CCT). The ERN was assessed before and after 16 sessions of treatment and within a comparable interval for HC. Subjective ratings about making errors on the task were obtained following each testing session. The ClinicalTrials.gov identifier is NCT00774150. Results The ERN was larger in ANX than HC youth but ERN magnitude did not significantly change following treatment in the ANX youth, regardless of treatment type, and baseline ERN did not predict treatment response. Post‐task ratings revealed that ANX youth worried more about task performance feedback than HC. Like the ERN, mean ratings did not significantly change following treatment. However, these ratings were not correlated with ERN amplitude. Conclusions Findings of greater ERN in pediatric anxiety disorders are replicated in a larger sample. More importantly, findings from this randomized control trial show that a larger ERN and feeling worried about performance feedback remain unchanged following treatment and are unrelated to treatment response. Such findings suggest that action monitoring systems remain overactive in anxious youth treated with psychotherapy, suggesting the need for future investigation of whether novel complimentary cognitive and emotional training programs can modify these systems would be warranted. - Journal of Child Psychology and Psychiatry, Volume 59, Issue 11, Page 1152-1161, November 2018.