Previous research has demonstrated the efficacy of two explanation-based approaches for increasing learning in educational games. The first involves asking students to explain their answers (self-explanation) and the second involves providing correct explanations (explanatory feedback). This study (1) compared self-explanation and explanatory feedback features embedded into a game designed to teach Newtonian dynamics and (2) investigated relationships between learning and individual differences. The results demonstrated significant learning gains for all conditions. There were no overall differences between conditions, but learning outcomes were better for the self-explanation condition after controlling for the highest level completed by each student. Analyses of individual differences indicated that certain threshold inhibitory control abilities may be necessary to benefit from the self-explanation in games.