This study investigated the claim (e.g., Vosniadou & Brewer's, 1992) that children have naive "mental models" of the earth and believe, for example, that the earth is flat or hollow. It tested the proposal that children appear to have these misconceptions because they find the researchers' tasks and questions to be confusing and ambiguous. Participants were 6- and 7-year-olds (N=127) who were given either the mental model theorists' original drawing task or a new version in which the same instructions and questions were rephrased to minimize ambiguity and, thus, possible misinterpretation. In response to the new version, children gave substantially more indication of having scientific understanding and less of having naive mental models, suggesting that the misconceptions reported by the mental model theorists are largely methodological artifacts. There were also differences between the responses to the original version and those reported by Vosniadou and Brewer, indicating that other factors, such as cohort and cultural effects, are also likely to help explain the discrepant findings of previous research.