音声ブラウザご使用の方向け: SKIP NAVI GOTO NAVI

WORD PROCESSING TOOLS FOR EDITING AND REVISING: IMPROVING THE WRITING SKILLS OF PERSONS WITH LEARNING DISABILITIES

Rena B. Lewis
Project Director
Enhancing Writing Skills Project
San Diego State University
6505 Alvarado Road, Suite 204
San Diego, CA 92120-1878
Voice/Message: (619) 594-8591
FAX: (619) 594-8592
e-mail: rlewis@mail.sdsu.edu
Internet: edweb.sdsu.edu/SPED/EWS.html

Tamarah Ashton
Assistant Project Director
Enhancing Writing Skills Project
e-mail: tashton@mail.sdsu.edu

Bonnie Haapa
Research Assistant
Enhancing Writing Skills Project

Carl Fielden
Research Assistant
Enhancing Writing Skills Project
e-mail: carl.fielden@sdsu.edu

Web Posted on: December 12, 1997


Persons with learning disabilities often encounter difficulties in writing, particularly in the editing and revising stages. Problems typically occur in monitoring and correcting errors in the more mechanical aspects of written language (spelling, capitalization, punctuation, and written grammar) as well as in making substantive changes in the content of the written product.

Enhancing the Writing Skills of Students with Learning Disabilities Through Technology is a three-year research project funded by the U.S. Office of Special Education Programs. It is sponsored by San Diego State University in collaboration with the North Coastal Consortium for Special Education in San Diego County. In the second year of the project, 1995-1996, a large scale study was conducted to investigate the effectiveness of computer-based editing/revising tools for students with learning disabilities in grades 4 through 12. These tools were spelling and grammar checkers and synthesized speech.


OVERVIEW OF THE STUDY

One hundred and twenty-two students with learning disabilities (LD) participated in this year-long study. All received writing instruction from their special education teachers using research-based methods, including teaching writing as a process and instruction in writing strategies. One group of students, the traditional group, continued writing by hand. All other students used the word processor Write This Way. In addition, 122 general education students made up a comparison group. Progress was assessed on three major variables: writing accuracy, quality of writing, and attitude toward the writing process.


THE WORD PROCESSOR

Write This Way was selected for study for several reasons:

  • (a) it contains checkers for spelling, grammar, and capitalization;
  • (b) it provides help with spelling, grammar, and capitalization errors;
  • (c) it offers synthesized speech; and
  • (d) it was designed for students with learning disabilities and those with hearing impairments

One advantage of this program is that teachers can choose which features to activate. For example, it is possible to turn the speech feature on and off, to choose whether the program will allow students to proof their work, and to determine when proofing occurs (continuously or only when requested). In addition, teachers can select the types of errors for which the program will search. Thus, it is possible to set Write This Way to detect spelling errors but not grammar errors or to search for only certain types of grammar errors.

In this study, Write This Way was configured in several ways in order to compare the effectiveness of its features. Students with learning disabilities were assigned to one of four groups: the traditional group, which wrote with paper and pencil, and three technology groups. The treatments were implemented for 18 weeks and students with LD received an average of 38 hours of writing instruction during that period. Students in Technology Group A spent the first intervention period (9 weeks in length) using Write This Way with no checkers activated; in the second intervention period (also 9 weeks in length), all checkers were activated. Group B used the spelling and capitalization checkers during the first intervention period; like Group A, in the second intervention period, all checkers were activated. Group C, the speech synthesis group, began with spelling and capitalization checkers and speech; in the second intervention period, all checkers were activated as well as speech synthesis.


STUDENTS WITH LEARNING DISABILITIES

At the end of the study, there were 106 students with LD. The most common reason for attrition was that students' families had moved. Of the 106 students, 68 percent were males; 73 percent received special education services in resource room settings and the rest in special class settings. The sample was primarily white (70 percent) and Hispanic (19 percent). The average grade placement of students with LD was grade 6.2 and their average age was 12 years, 0 months. All students were identified as fluent English speakers.

As a group, the students with LD were characterized by average intellectual performance (average overall IQ 96) and poor writing achievement (average writing standard score 74). On their current IEPs, the most common goals in the area of writing were those related to spelling (74% of students), capitalization and punctuation (57%), grammar/usage (36%), and production (e.g., write a paragraph, 34%; write more than one paragraph, 32%).


THE MEASURES

Measures included a scale to assess attitude toward writing, a measure of writing speed, and the collection of writing samples to allow analysis of writing quality and accuracy. In addition, students and their teachers were interviewed at the end of the study to learn their views about the usefulness of the technologies.

Writing samples, both first and second drafts, were word processed before analysis. The quality measure was a holistic rating scale adapted with permission from the Test of Written English. Two raters read and evaluated each sample, compared ratings, and reconciled ratings if necessary.

The accuracy measure was a checklist listing several types of errors in spelling, capitalization, punctuation, and syntax. Two major types of spelling errors were considered: errors which were real words (e.g., substituting a homonym such as "there" for "their") and nonreal word errors (e.g., writing "werk" for "work"). Syntax errors were divided into two categories. Type I syntax errors included mistakes related to sentence elements such as nouns, verbs, pronouns, adverbs, adjectives, articles, and negatives. Type II syntax errors related to more meaningful aspects of written language: sentence formation (e.g., run-on's fragments), missing or extra words, word choice, word order, and lack of intelligibility.

In addition, changes students made from first to second draft of the writing sample were studied with a change analysis checklist. This checklist was organized using the same categories as the accuracy checklist. Each change was tallied according to type and rated according to whether the change was an improvement, a decrease in accuracy, or neutral.


ACCURACY RESULTS

At pretest and at posttest, the largest number of errors made by students with learning disabilities were spelling errors, followed by syntax type II errors. At pretest, students in the traditional group made an average of 18.9 spelling errors per 100 words written and students in the technology groups made an average of 22.4 spelling errors per 100 words. At the end of the study, students in the technology groups had decreased their errors to 14.1 per 100 words. The error rate for students in the traditional group remained approximately the same, 18.5 per 100 words. Syntax type II errors decreased over time for both traditional and technology group students.

In the area of spelling, the most typical errors were nonreal word errors. These decreased over time for students in the technology groups but not for traditional group students. This decrease is likely attributable to the use of the spelling checker.

This finding becomes clearer when results of the change analysis are considered. A subsample of students was selected for this analysis: 15 students with LD and 13 general education matches. Every change these students made from first to second draft was examined for their pretest and posttest writing samples. At pretest, general education students and students with LD in the traditional and technology groups made approximately the same numbers of changes and the effectiveness of those changes was similar. For example, general education students made an average of 24 changes, of which 40% were positive while technology group students made an average of 21 changes, of which 37% were positive.

At posttest, the number of changes made by technology group students with LD fell in relationship to those of other groups. However, the effectiveness of their changes increased dramatically. At posttest, 60% of the changes made by technology group students were positive. This contrasts with a rate of 48% for general education students and 42% for traditional group students with LD.

This change was most impressive in the area of spelling. At posttest, 69% of the spelling changes made by technology group students were positive. For general education students, 36% of their changes were positive; for traditional group students with LD, only 16% were positive.


OTHER RESULTS

All groups of students with LD showed improvements in quality of writing over time but a slight decrease in attitude toward writing. At both pre and post, general education students were superior to students with LD in writing accuracy, quality, and attitude toward writing. Only in the area of writing accuracy did the gap appear to narrow over the course of the study.


CONSUMERS' VIEWS

Teachers were generally positive about Write This Way. Seventy-eight percent of teachers said that they would recommend this program to others; 70% said that their students could use the program independently. Advantages of the program were its speech, spelling and grammar checkers, ease of use, and proofing. However, teachers named as disadvantages its slow speed, lack of more advanced word processing features, and difficulties with the spelling and grammar checkers. Teachers felt that the help provided by the spelling checker was not always adequate; the dictionary did not include all the words their students required, correct words were not always suggested, and they did not like having to request help for each individual word. The grammar checker received much more severe criticism. One of the drawbacks of this feature is that its algorithms rely upon text with correctly spelled words, which clearly presents problems for students with learning disabilities. According to teachers, help for grammar is often confusing, often not helpful (e.g., too specific or too difficult), and often not available (the program says "ask your teacher").

At the end of the project, teachers told us stories about how the interventions had affected their students. Teachers talked about improved attitude and increased motivation, skill improvement, increases in fluency, and improved form. For example, one teacher said, "Student X was off the charts in improvement."

Another teacher reported, "Student X at the beginning of the year when given a written assignment would shake, even cry, and would have tremendous difficulty with spelling errors. He now writes, is comfortable at the computer, and can do it."


RECOMMENDATIONS

Consider these recommendations in choosing word processing tools for the editing and revising stages of writing.

  • 1. Teach writing skills AND teach students how to word process.
  • 2. Consider using a spelling checker. To see if a specific program will work in your classroom, try it out using samples of your students' writing.
  • 3. Remember that spelling checkers are not perfect. When possible, add "personal" words (such as students' names, the name of your school and community) to the program's dictionary. Be alert for real word errors; spelling checkers will not flag a word as incorrect if it's a real word. When the correct word isn't suggested by the program, offer students other strategies. Have students try another spelling, consult a personal dictionary, ask a peer or teacher for help. If students can't read the words suggested by the spelling checker, choose a program with speech synthesis where the options suggested by the program are read aloud. Finally, a spelling checker is not a substitute for instruction; continue to teach spelling.
  • 4. Be wary of grammar checkers.
  • 5. Consider choosing a program that offers speech synthesis.

NOTE. The contents of this report were developed under a grant from the U. S. Department of Education. However, those contents do not necessarily represent the policy of the Department of Education, and you should not assume endorsement by the Federal Government.