The purpose of this theoretically-based study was to examine the effects of yellow-highlighting "relevant" words and units within math word problems. Initial differences were documented between 10 girls at-risk for ADHD and 10 comparisons on the performance of group and individual assessments of math computations and word problems, as had previously been reported for boys. To address these deficits (faster speed and lower accuracy) for the at-risk group, these fourth-grade girls were randomly assigned word problems under two experimental conditions (highlight vs. nohighlight) within a counterbalanced condition- and form-order design. Intervention gains were (a) slower rate of performance, (b) decreased off-task behavior, and (c) improved problem solving performance, differentially for girls at-risk for ADHD when their materials were first presented with highlighting--the effects of which carried over to a non-highlight condition. The theoretical, practical, and future research implications of this study are addressed. (Contains 1 table and 3 figures.)