This study attempts to account for disparities between the Anderson and Cuneo (1978), Leon (1982), and Lautrey, Mullet, and Paques (1989) studies in regard to children's area judgment. Two task variables were manipulated: stimulus distribution (biased/unbiased) and the type of response scale (graduated/ungraduated). Three age groups (5, 6, and 7 year olds) were tested. The mean integration pattern for 5-year-olds presented a negatively biased stimulus distribution, and an ungraduated response scale was highly convergent and suggested the use of a centration rule (replication of the Lautrey et al. results). When 5-year-olds were presented with an unbiased stimulus distribution and a graduated scale, the integration pattern was only slightly convergent (as in Leon). The effects of two factors (age and graduation) were significant and combined additively: The older the child, the more graduated the response scale and the more the integration pattern tended to form three ascending parallel lines (the Anderson & Cuneo results).