The low rate of agreement among readers of College Entrance Examination essays suggested the need to examine the qualities in student writing which caused wide variance in grading. To study this question, 300 homework papers by freshmen at three universities were graded by 60 distinguished readers in six fields. The following factors, by rank, seemed to influence readers: ideas expressed, grammar, punctuation, spelling, handwriting, organization, analysis, wording, phrasing, and "flavor." These factors reduced to "general merit" and "mechanics," in addition to three possible ratings of "high,""medium," or "low," were used to grade monthly test papers of English pupils in 17 high schools for 1 year. From this trial period, a means of measuring student growth in writing ability was developed. All students in a span of three grades would simultaneously write on the same topic several times a year. The unidentified papers would be graded, and the students' scores compared over a 3-year period, would indicate ther progress. (JM)