Error processing is frequently examined using the error‐related negativity (ERN), a negative‐going event‐related potential occurring after the commission of an error at frontal‐central sites, and has been suggested as a neural biomarker that may be useful in characterizing trajectories of risk for anxiety. While the ERN has been shown to have excellent psychometric properties in adults, few studies have examined psychometric properties of the ERN in children and adolescents. The current study examined the 2‐year test‐retest reliability of the ERN in a sample of children and adolescents, and the convergent validity of the ERN using a flanker and go/no‐go task. Results suggest that the ERN is both reliable and stable across 2 years and across tasks. However, results also indicate that the internal consistency obtained using the flanker task is greater than the internal consistency obtained using the go/no‐go task.