A study finds that an unrewarded-exploration group later chooses correct paths, but participants report having studied a posted map during the break. What is the main problem for interpreting the result as latent learning?