There is a change in false belief task performance across the 3-5 year age range, as confirmed in a recent meta-analysis [Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory mind development: The truth about false-belief. "Child Development," 72, 655-684]. This meta-analysis identified several performance factors influencing success, including manipulations that highlight the salience of the initial belief content (such as asking where Sally will look "first" for the marble). However, because a proportion of variance in performance remained unexplained even when identified performance factors were controlled for, the authors concluded from the standpoint of a "theory--theory" account that children's improvement is the result of conceptual change. Further, the meta-analysis showed that manipulations such as "look first" improve performance only in children who are in the older part of the 3-5 year range, and thus plausibly operating with a "transitional" theory of mind--just on the point of realizing conceptual change. Here, we present three studies systematically investigating the "look first" manipulation which showed that: (i) the advantage for the look first question can be demonstrated in children across different cultures, (ii) look first has an effect that is additive to the improvement with age; there is no interaction such that older children gain more benefit from younger children, (iii) performance in younger children can be, but is not always, elevated to levels that are statistically above chance. These results challenge the theory-theory account and are discussed in terms of models of belief-desire reasoning in which "both" conceptual competence "and" performance factors play central roles.