A spatially distributed instruction format (i.e., when information sources are presented side-by-side) has been found to be generally beneficial for learning statistics (Jang, Schunn, & Nokes, 2011). In a follow-up classroom study, we examined whether students generally selected the better format (i.e., faster problem solving with better understanding of materials; distributed format in this study) when given the choice and whether individual differences affect students’ instruction format preferences. Students were found to prefer the instruction format that matches to their ability (an adaptive choice): Students with high mental rotation and verbal learning ability preferred the spatially stacked format of instruction to a distributed format.