NotesFAQContact Us
Collection
Advanced
Search Tips
ERIC Number: ED490473
Record Type: Non-Journal
Publication Date: 2005
Pages: 24
Abstractor: Author
Reference Count: 61
ISBN: N/A
ISSN: N/A
The Effects of Grammar Testing on the Writing Quality and Reduction of Errors in College Freshmen's Essays
Davis, Wes; Mahoney, Kelley
Online Submission
This experimental, statistical study investigated the effects that the testing of grammar and writing mechanics would have on the overall quality and reduction of errors in college students' essays for freshman composition. In the experimental group of 42 students, the professor assigned several exercises in grammar and mechanics as a review related to composing skills and then gave two major tests on proofreading essays for grammatical errors. However, the other professor did not give these grammar tests to the 41 students in the control group. The study used "T-tests" for statistical analysis on pretest and posttest essays, which each of the 83 students had written. On overall writing quality, the faculty raters holistically scored the students' essays, using a scale from 1 (failing) to 4 (superior). Since the two raters scored each student's pretest and posttest essay, each essay had a combined score resulting in a scale from 2 (failing) to 8 (superior). The results showed that the 42 students of the experimental group who tested on grammar had a mean pretest essay score of 2 and a mean posttest essay score of 4.53, showing a gain of 2.53. Statistically, these students made very highly significant gains in overall writing quality. The 41 students of the control group not tested on grammar had a mean pretest essay score of 2.66 and a posttest score of 4.49, showing a gain of 1.83. These students also made very highly significant gains in overall writing quality, although the experimental group's posttest essay scores were still significantly higher than the control groups. On traditionally serious grammatical errors, the experimental group had a mean number of 1.5 serious errors on the pretest essay and a mean number of 0.93 error on the posttest, a reduction of 0.57 which was statistically significant. The control group's mean number on the pretest essay was 1.23 serious errors and a mean posttest number of 0.64 error, a reduction of 0.59 which was also significant; however, there was no significant difference between both groups in the reduction of serious errors. For the less serious "minor" errors, the experimental group had a mean number of 7.4 minor errors on the pretest essay and a mean number of 5.12 errors on the posttest essay, a reduction of 2.28 errors which was highly significant. The control group had a mean pretest essay number of 12.5 minor errors and a mean posttest essay number of 7.42 errors, a reduction of 5.08 errors which was very highly significant. Moreover, statistics showed a significant difference between both groups' posttest essays in the reduction of these minor errors, with the control group making more significant reductions in the minor errors than the experimental group. The researchers concluded that the two major grammar tests on proofreading two essays for errors may have had some effect on the experimental group's gains in overall writing quality for correctness. However, these tests appeared to make no difference between both groups, by having very little if any effect on the students in the experimental group to reduce the number of errors significantly in their essays. (Contains 4 tables.)
Publication Type: Information Analyses; Reports - Research
Education Level: Higher Education
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A