Purpose: To assess neurocognitive and behavioral performance in children with idiopathic epilepsy (CWE, n = 74), their siblings without epilepsy (control, n = 23), and children with migraine (CWM, n = 13), and to identify medical factors related to learning or behavioral problems in CWE.

Methods: Subjects, ages 8–13 years with IQs of ≥80, completed a neurocognitive test battery annually for ≤3 years. For CWE, age at seizure onset, most recent EEG results, seizure type, seizure frequency, current antiepileptic drug (AED), and most recent AED serum levels were documented at each visit.

Results: CWE and CWM had high rates of grade retention and placement in special education compared with sibling controls. CWE performed worse than controls on numerous neurocognitive variables. These differences persisted over time. CWE with abnormal EEGs scored lower than CWE with normal EEGs on reading and spelling measures, even with comparable IQs. Age at seizure onset, seizure type, and seizure frequency were not related to neurocognitive or behavioral test scores. CWE taking carbamazepine (CBZ) performed better than CWE taking valproate (VPA) on academic achievement measures, although the study lacked controls necessary to assess this finding thoroughly. CWM did not differ from CWE or controls in cognitive or academic achievement skills.

Conclusions: Long‐term risk of learning problems exists among CWE as compared with controls, even with normal IQs and well‐controlled seizures. Predicting learning problems in CWE based on medical factors remains elusive. Monitoring of educational progress and neurocognitive screening may be most effective in assessing academic risk for CWE.