Grammar Checkers: Programs That May Not Enhance Learning

Much has been written about the use of word processors in college writing classes. Most of the research has been in English departments. Appleby (1983) reported that students who use word processors come to understand that writing is a process that involves revision. Barnhardt (1988) demonstrated that students who revised (on word processors) produced work that was judged significantly better than students who did not revise their work. But the research on quality is not conclusive (Hooper, 1987). One tool that aids revision is the grammar checker. Simon (1986) noted that “text analysis” programs allow the student to concentrate on style. She also noted that some think students might rely too heavily on these aids. In one of the first studies of grammar checkers for writing instruction, Smye (1988) noted that “a typical checking cycle might take 15-20 minutes” (p. 48). This delay made grammar checkers much less useful for classroom use. Nonetheless, Smye was optimistic that grammar checkers could be useful eventually to help students understand that writing involves revisions. Peek, Eubanks, May and Heil(l989) tested Rightwriter, a commercially available grammar checker for IBMs and compatibles, in writing exercises with two sections of accounting students. The investigators reasoned that the quality of a student’s writing would vary as a function of conscious choices made while the student was involved in the writing process. Grammar checkers, then, would have the potential of making writers more aware of problems. They used two written assignments to investigate this question with accounting students. The first assignment was done by hand and served as a pretest measure for writing quality. A second assignment was done with the aid ofa computer. The experimental group (N=35) submitted drafts on disk to the instructor who used Rightwriter to generate a list of computersuspected errors. Students in this experimental group then had the opportunity to revise the draft based on the computergenerated exceptions. Members ofthe control group (N=24) prepared their work with the aid of a computer but without benefit of the computerized grammar analysis. No students actually ran a grammar checker and the test involved only one feedback event. The hypothesis was non-directional and there is no mention of randomization in assigning students to groups or groups