Purpose: In this study, the authors examined how variability in classroom social communication performance differed between children with fetal alcohol spectrum disorders (FASD) and pair-matched, typically developing peers. Method: Twelve pairs of children were observed in their classrooms, 40 min per day (20 min per child) for 4 days over a 2-week period. Coders documented classroom social communication during situations of Cooperation and following School Rules by recording performance on handheld computers using the Social Communication Coding System (SCCS). The SCCS consists of 6 behavioral dimensions (prosocial/engaged, passive/disengaged, irrelevant, hostile/coercive, assertive, and adult seeking). The frequency of occurrence and duration of each dimension were recorded. These measures were then used to examine variability in performance within and across days (changeability and stability, respectively). Results: Independent of classroom situation, children with FASD were more variable than their typically developing peers in terms of changing behavioral dimensions more often (changeability) and varying their behavior more from day to day (stability). Conclusions: Documenting performance variability may provide a clearer understanding of the classroom social communication difficulties of the child with mild FASD. (Contains 5 tables.)