Rozenblit and Keil (2002) claim that people are subject to an illusion of explanatory depth (IOED) whereby they believe they understand the world in greater detail, coherence and depth than they actually do. In the present research, we questioned Rozenblit and Keil’s conclusions in two ways. First, we tested whether people might overestimate their explanatory knowledge as a result of misconstruing how to initially rate their understanding of stimuli. We found that when directed to consider the physical-mechanical processes of stimuli instead of their functional affordances, participants did in fact offer more accurate estimates of understanding relative to their explanatory performance. Second, we tested whether the explanations participants proffer are misleadingly shallow. We predicted that by encouraging a more relational encoding of stimuli, participants would be able to produce better explanations. However, the results showed that participants’ explanations remained shallow after relationally encoding stimuli.