The purpose was to ascertain whether there existed differential achievement effects associated with the factors and levels of context, form, item-stem and item-response format, size of common number, and school grade, as they relate to two variations of distributivity. Twelve 9-item tests were constructed and given to intact 4th-, 5th-, 6th-, and 7th-grade classes from two midwestern city school districts; each pupil worked with one of the 12 randomly distributed tests. Conclusions were that pupils' sensitivity to the use of distributivity was relatively low, that ability to complete correctly examples of types tested tended to increase from grade four to five to six to seven, that "regrouping sets" examples were less difficult than "multiplication-addition" examples and that"right-distributive" examples were easier than "left-distributive" examples, and that pupils had only a limited tendency to respond in the same way across a set of examples within a test. (DT)