We examined 3- to 6-year-old children's understanding of their own and another's false beliefs in two experiments using diverse deceptive box and unexpected transfer tasks. Tasks in the first experiment required the child to think about an absent person's responses to questions, whereas those in the second experiment used a confederate present during testing. Children performed better on deceptive box tasks when the other person was present rather than absent; they performed equally on accidental and intentional unexpected transfer tasks. Children understood the other's false beliefs by 4-5 years of age, and their own by 5-6 years of age. Children aged between 4 and 5 years appear to have greater difficulty discounting or ignoring misleading information when thinking about their own beliefs than when thinking about another's, at least for the particular false belief tasks employed. Inconsistencies between our findings and those of a recent meta-analysis (Wellman, Cross, & Watson, 2001) are easily reconciled.