Phases two and three of an evaluation of an innovative approach to English Composition instruction at Long Beach City College are reported. In the innovative approach, each class was divided into two groups of 15 students each, which met twice each week to discuss topics for writing, to read and revise first drafts of papers, and to discuss basic writing problems as they arose. In addition, the students were to spend one hour each week working on writing problems with teachers and teacher aides. In Phase Two, a comparison was made of 136 final essays of 150 students (67 TBA, i.e., experimental, and 69 traditional). All of the essays were written under controlled conditions. The results were tabulated on an essay score sheet which allowed tabulation of 7 rhetorical variables and 14 non-standard features; in addition a tabulation was made of total errors, error-word ratios, and error-sentence ratios. Analysis of the data is provided. The results of the analysis show that the non-TBA students were more successful in eliminating errors than were the TBA students. In Phase Three, performance in transfer-level English Composition (English 1A) was studied. Criterion variables used were: the tendency to enroll in transfer English Composition; the tendency to complete or persist in English 1A; the ratio of achievement grades earned; and the ratio of penalty grades earned. Results of the study showed that the traditional population demonstrated a greater tendency to enroll in the transfer composition course, the difference being at the .05 level of significance. It is concluded that both types of courses equally prepare students for transfer Freshman English. The Essay Score Sheet and tables are provided. (DB)