The present study adapts the typical eyewitness misinformation paradigm into an academic context. Unbalanced English–Spanish bilinguals (N = 81) listened to a lecture in English (L1) or Spanish (L2), read notes in L1 or L2, and completed a forced‐choice recognition test in the lecture language. Unlike prior studies with proficient bilinguals, unbalanced English‐dominant participants showed greater recognition memory accuracy for material presented in English only than did material presented in Spanish only. English misinformation had a greater impact on memory for the Spanish lecture than vice versa. Most importantly, the modified misinformation paradigm is an effective tool to investigate academic misinformation effects and could be used in bilingual and monolingual research. Copyright © 2017 John Wiley & Sons, Ltd.