In the fall of 1986, a pilot study was conducted to investigate the differential performance at the item level of Mexican-American students who spoke English as a second language (ESL) versus White native English speakers. This study was designed to replicate the pilot study and test the hypotheses based on that study. The test materials used were the English Usage and Mathematics Usage tests of the American College Testing Program Assessment. It was hypothesized that: (1) items that emphasize mechanics in the English Usage Test, such as grammar and punctuation, tend to favor ESL examinees; (2) items that focus upon style and structure in the English Usage Test tend to favor non-ESL students; and (3) mathematical items with the greatest verbal load tend to favor non-ESL examinees. Subjects included 471 Mexican-American ESL students and 1,000 White native speakers (non-ESL). Results indicate that none of the three hypotheses was supported. Although the mean score for the ESL students was almost a full standard deviation below that of the native speakers for both tests, it appears that the group difference in performance was reflected throughout most of the test items. Specific categories of items that were disproportionately easy or hard for either of the groups could not be found. Six graphs and four tables conclude the document. (TJH)