Background: Most classroom observation studies have documented significant deficiencies in the classroom attention of children with attention-deficit/hyperactivity disorder (ADHD) compared to their typically developing peers. The magnitude of these differences, however, varies considerably and may be influenced by contextual, sampling, diagnostic, and observational differences. Methods: Meta-analysis of 23 between-group classroom observation studies using weighted regression, publication bias, goodness of fit, best case, and original metric analyses. Results: Across studies, a large effect size (ES = 0.73) was found prior to consideration of potential moderators. Weighted regression, best case, and original metric estimation indicate that this effect may be an underestimation of the classroom visual attention deficits of children with ADHD. Several methodological factors--classroom environment, sample characteristics, diagnostic procedures, and observational coding schema--differentially affect observed rates of classroom attentive behavior for children with ADHD and typically developing children. After accounting for these factors, children with ADHD were on-task approximately 75& of the time compared to 88% for their classroom peers (ES = 1.40). Children with ADHD were also more variable in their attentive behavior across studies. Conclusions: The present study confirmed that children with ADHD exhibit deficient and more variable visual attending to required stimuli in classroom settings and provided an aggregate estimation of the magnitude of these deficits at the group level. It also demonstrated the impact of situational, sampling, diagnostic, and observational variables on observed rates of on-task behavior.