This study examined the effect of two ways of question phrasing as related to three styles of expository writing on the test performance of elementary school children. Multiple-choice questions were developed for sets of passages which were written using three different syntactic structures; and which had different levels of difficulty. The questions were phrased using the complete and incomplete stem mode. A total of six different tests were compiled based on stem mode and syntax of the passages. These tests were randomly given to children in two schools (grades two, three, and four). No significant difference in performance was found for sex, school, or grade level for the different syntactic structures. A significant difference in performance was found for the complete versus the incomplete stem mode for the multiple-choice items. Performance was superior on the items with a complete stem; this was true for both schools, both sexes, and for all three grade levels. As might be expected, performance generally increased with grade level. (Author/GDC)