The purpose of this study was to examine the means of scores on three forms of a standardized reading comprehension test taken by community college students in developmental reading classes. The three forms of the test were administered as a timed multiple-choice test, a constructed response test, and an un-timed multiple-choice test. Scores on the Group Embedded Figures Test (GEFT) were used to classify the students who participated in the study as having field dependent (LOW GEFT), mid-field dependent/independent (MID GEFT), or fieldindependent (HIGH GEFT) tendencies. The paired samples test was used to analyze the scores among the students classified as LOW GEFT, MID GEFT, and HIGH GEFT for mean differences in scores on the three test formats. The data revealed that for LOW GEFT students, the format of the test impacted their scores, with the mean of the scores of the un-timed multiple-choice test being significantly higher than the timed multiple-choice test and the constructed response format. The data also showed that for the MID GEFT students the mean of the scores for the un-timed multiple-choice test was significantly higher than the means for the timed multiple-choice test scores and the constructed response test scores. However, no significant mean difference was found between the timed multiple-choice test scores and the constructed response test scores. For the HIGH GEFT students, significant mean difference existed only between the un-timed multiple-choice and the timed multiple choice scores. The means of reading comprehension test scores on the three formats between the LOW GEFT, MID GEFT, and HIGH GEFT students indicated significant mean difference between the timed multiple choice test scores but not between the means of the scores for the constructed response and the un-timed multiple-choice test scores. Demographically, when the means of the reading test scores were analyzed with ethnicity as the controlling variable, the Hispanic students had a significantly higher mean on the scores for the constructed response test format. No other significant mean differences were found between the scores of the African American, Caucasian, Hispanic, or Native American students. When the means of the reading test scores were analyzed with gender as the controlling variable, no significant mean difference was found between the reading comprehension scores of the men and women. This study indicated that cognitive style had more impact on students’ performance on a standardized test of reading comprehension than did ethnicity or gender. The un-timed multiple-choice format also had an equalizing effect on the means of the scores for these students.