Research with adults has shown that variations in verbal labels and numerical scale values on rating scales can affect the responses given. However, few studies have been conducted with children. The study aimed to examine potential differences in children's responses to Likert-type rating scales according to their anchor points and scale direction and to see whether or not such differences were stable over time. A total of 130 British children, aged 9 to 11 years, completed six sets of Likert-type rating scales, presented in four different ways varying the position of positive labels and numerical values. The results showed, both initially and 8-12 weeks later, that presenting a positive label or a high score on the left of a scale led to significantly higher mean scores than did the other variations. These findings indicate that different arrangements of rating scales can produce different results, which has clear implications for the administration of scales with children. (Contains 3 tables.)