Word processing's impact on student performance in required writing courses is not clear. The literature up to 1984 reflects early enthusiasm for the potential that writing teachers saw in the technology. Heavily testimonial and speculative, this work is full of commonsensible reflections on personal experience, and projections about likely positive interplay between word processing's central features and the writing process (Bridwell, Nancarrow, & Ross, 1984). Since about 1984, however, such informal endorsements have been modified. Performance gains as measured do not always meet expectations (Nichols, 1986; Redman, 1985). When revision behavior is studied, the findings disappoint enthusiasts fairly consistently (Bridwell, Sirc, & Brooke, 1985).

Both the early speculative articles and the more recent controlled studies are limited. Speculation and testimony carry the obvious limits of any untested inquiry. The quantitative studies, on the other hand, frequently look at small samples, study writers early in their word-processing history, or base their observations on limited machine time. What the studies have in common is their acknowledged inconclusiveness about the ways in which student writing might or might not improve in a word-processing environment.

This study shifts discussion away from improvements in student writing when that writing is done using a word-processing package and asks a pragmatic, administrative question: Do students who enroll in sections of required first-year writing courses taught in a microcomputer laboratory/classroom persist to completion of required courses at rates different from the rates of persistence observed in pen-and-paper sections of the courses? An earlier study (Moore, 1985) reports a significant increase in course completion rates among students using word processing in a first-year composition course.

That students should persist toward completion of the courses in which they are required to enroll is of interest for a number of reasons. Efficiency in use of resources, morale between students and teachers, and programmatic accountability come to mind as the most obvious. Considerable expense, trouble, and disruption attend the introduction of computers into a writing curriculum (Selfe, 1987); and in view of the mixed results on student writing reported in the literature, administrative concerns of other sorts must be addressed . Retention of students who arc moving toward degrees is certainly such a concern.

This study compares the course completion rates of two groups: the first, a cohort of students enrolled in sections of required first-year writing courses taught in a microcomputer-equipped classroom; the second, students enrolled in all other sections of the course. Both groups enrolled in a two-quarter writing course sequence satisfying the lower-division writing requirement in the non-competitive admissions unit of a large urban research university. Students enrolling in this unit completed high school in the lower half of their class and reported college entrance examination scores in the lower two quartiles. The study was based on 931 students enrolled in the two courses during the winter and spring quarters of 1986. Students selected their own sections.

Except for the fact that study sections were taught in a microcomputer-equipped classroom, conditions of instruction were constant across sections. All sections of the courses met in 10-week quarters. All were taught from a staff-generated common syllabus using staff-written texts. The common syllabus called for an emphasis on writing workshop methods. In the first course, emphasizing informal and narrative writing, all students completed the same sequence of nine assignments. In the second course, emphasizing public and academic writing, there was less homogeneity of assignment sequence, but there were similar objectives and methods. Staff training provided for consistency in approach and in course content across all sections. A free, walk-in tutorial center for supplemental instruction was available to all students.

In all sections, students attended four hours of class each week. In microcomputer sections, instructors required students to use a word-processing package in completing assignments in and out of class. All class sessions for these sections were held in the computer classroom. Students in microcomputer sections also had access to the lab when classes were not in session. Students in other sections were allowed to use word processors if they so wished, but they did not use the microcomputer-equipped classroom or its lab time. To the extent that isolating a single variable is possible when multiple teachers are involved, the difference across sections was the microcomputer classroom environment (logistics prevented the same instructors from teaching study and control sections).

The microcomputer classroom was equipped with fifteen nonnetworked microcomputers, a ratio of one computer for every two students per section. Five each of Apple IIe, Macintosh, and IBM PC computers were available (a secondary analysis of data indicated that machine type produced no differential outcome). In addition to computer stations, the microcomputer classroom contained tables for group discussions and individual off-computer work spaces like those found in typical classrooms. Apple IIe and IBM computers were supplied with the BANK STREET WRITER word-processing package; Macintosh computers were supplied with MACWRITE. Spelling-checker programs were available. Students who had a previous history with, and a preference for, another software product were allowed to use it. Students enrolled in the microlab sections were required to attend an hour of training prior to the start of the course.

Students were determined to have "completed" the course if they received a letter grade for the course, including "earned" failing grades. Students who received grades of W (indicating withdrawal), or students who were assigned failing grades because they did not attend class or did not complete assigned work were counted as not completing the course. Students who received an initial grade of I (incomplete as of the last day of class with permission to make up missing work) were counted as having completed the course if they made up missing work and received a passing letter grade within six months of the end of the course. Students with incomplete grades unchanged after six months were counted as not completing the course.

The students in the microcomputer sections completed the two courses at rates higher than did those enrolled in all other sections.

First Course
Students enrolled in microlab sections of the first course (n=72) completed the course at a rate of 95.8%. Those enrolled in all other sections of the first course (n=418) completed the course at a rate of 83.7%. The completion rate for all sections (micro and others, N=490) was 85.5%. See Table 1.

Second Course
Students enrolled in microlab sections of the second course (n=75) completed the course at a rate of 93.3%. Those enrolled in all other sections of the second course (n=366) completed the course at a rate of 86.8'~o. The completion rate for all sections (micro and other, n=44] ) was 88.()%. See Table 2.

In an earlier study, Moore (1985) reported completion rates for first-year composition classes using microcomputer word processing against rates for students in traditional sections. Of Moore's students who used word processing, 83.5% completed the course while 56.2% of those who did not use computers completed the course (p. 59). Owing to differences in conditions, direct comparisons between results of Moore's study and this study are not possible. But, it is worth noting that both studies report increased retention of students to course completions when the course is taken in a microcomputer word-processing setting.

Are these findings significant? We think so. As reviewed above, the literature can be read to suggest that word processing has doubtful impact on the writing performance of students when performance is assessed through study of revision or scores on writing samples. But, to the extent that first-year writing programs are funded to provide service courses through which students meet a college or university requirement, "performance" can and should be defined more broadly to address the seemingly mundane administrative question of student success in completing and passing required first-year writing courses. The evidence is that students do indeed persist at higher rates when their writing instruction in required first-year courses includes use of word processing. And, such information can help writing teachers and program administrators justify the expense and trouble that accompany shifts toward computer-enriched writing environments for our students.

Terence Collins, Nancy Engen-Wedin and William Margolis teach at the University of Minnesota.

Bridwell, L., Nancarrow, P., & Ross, D. (1984). The writing process and the writing machine: Current research on word processors relevant to the teaching of composition. In Beach, R., & Bridwell, L. (Eds.), New directions in composition research. New York: Guilford.

Bridwell, L., Sirc, G., & Brooke, R. (1985). Revising and computing: Case studies of student writers. In Freedman, S. (Ed.), The acquisition of written language: Revision and response. Norwood, NJ: Ablex.

Holdstein, D., & Redman, T. (1985). Empirical research in word processing: Expectations versus experience. Computers and Composition, 3(1), 43-54.

Moore, W. (1985). Word processing in first-year composition. Computers and Composition, 3(1), 55-60.

Nichols, R. (1986). Word processing and basic writers. Journal of Basic Writing, 5(2), 81-97.

Selfe, C. (1987). Creating a computer lab that composition teachers can live with. Collegiate Microcomputer, 5(2), 149-158.