Those items or test characteristics that are likely to result in differential item functioning (DIF) across accommodated test forms in statewide tests have received little attention. An examination of elementary-level student performance across accommodated test forms in a large-scale mathematics assessment revealed DIF variations by grades, accommodation conditions, graphic inclusion, and item types (i.e., constructed response versus multiple choice). In general, items with a lower grade, graphic, constructed-response format appear to benefit accommodated examinees for a statewide mathematics test. (Contains 3 tables and 4 figures.)