ABSTRACTBrief, impromptu essays written for the 1990 administration of the College Board's English Composition Test (ECT) were randomly sampled for four different groups of examinees. These essays were subjected to further holistic ratings beyond those conducted for the ECT administration, and analytical ratings were also obtained. The holistic scores were correlated with the analytical scores to determine which essay characteristics were most closely associated with high holistic scores. The results indicated that overall organization, use of supporting materials, noteworthy ideas, rhetorical strategy, and thesis statement were the strongest correlates. Comparisons with the results of a previous study of 1979 ECT essays, written on a different topic, showed that these same characteristics were also most closely associated with high holistic scores in that study. These characteristics correlated strongly with high holistic scores for all subgroups randomly sampled, and for males and females.Additional factors associated with better holistic scores but studied only for the 1990 topic included following instructions, completion of the task, use of a standard essay format, use of literary references, and proficiency in the English language.Certain types of essay content were also associated with higher scores on the 1990 topic. Essays combining current affairs with literature and history or combining literature and history received slightly higher scores on average than essays based only on current affairs, literature, history, or personal experience. Nevertheless, some essays receiving high scores were based solely on personal experience, did not contain literary references, and did not use a standard essay format. The analyses suggest that some practice with this type of brief, impromptu essay, particularly under strict time constraints, may be useful as preparation for taking such essay examinations.