The purpose of this study was to determine whether item difficulty is significantly affected by language difficulty and response set convergence. Language difficulty was varied by increasing sentence (stem) length, increasing syntactic complexity, and substituting uncommon words for more familiar terms in the item stem. Item wording ranged from very simple sentences to sentences involving excess verbiage and syntactic transformations. The semantic similarity of item response options was also varied, creating three levels of response set convergence. Seventeen judges confirmed the orderings of language difficulty and response set convergence for the test items, and 990 undergraduate students at the University of Washington responded to the items. Analyses of item difficulty calculated for each level of each of the two factors included an additive conjoint method, repeated measures analysis of variance, and a test of proportions. Response set convergence was an effective manipulation in the ranges studies. Language difficulty had a non-significant effect on difficulty level. (CM)