A comparison of measures of the attained writing proficiency of 45 college-level students of English as a second language is reported. Students were tested by two indirect measures, the Test of the Ability to Subordinate (TAS) and the Revision and Editing Test (RET), and their compositions were evaluated by a direct measure, holistic evaluation. The TAS consists of 45 questions requiring students to combine two or more sentences with any form of subordination. The RET asks questions requiring an understanding of units larger than the sentence and the ability to improve sentences or sentence groups. Results of the two tests were compared with the holistic composition evaluation, the criterion measure, and it was determined that: (1) the TAS has higher reliability and validity than the RET for these subjects; (2) neither test has concurrent validity with the holistic evaluation; (3) the revision and editing subparts are of equal difficulty for this group but assess different constructs as they were designed to; and (4) this study's findings on the holistic evaluation's composition profile differ markedly from previous researchers' conclusions. Despite finding number 2, it was concluded that the holistic scoring profile guide and RET provide similar criterion-related information for assessment, placement, exemption, certification, and promotion. Statistical analysis results and references are included. (MSE)