Printer Friendly

Auditory feedback and writing: learning disabled and nondisabled students.

Auditory Feedback and Writing: Learning Disabled and Nondisabled Students

Instructional research with learning disabled (LD) students has traditionally focused on reading, mathematics, and spelling. Only recently have researchers begun to study written composition; a recent special issue of Exceptional Children, for example, was devoted to "Research and Instruction in Written Language" (Graham & Harris, 1988). The initial work has been descriptive; both processes (cognitive processes students use as they compose) and products (the outcomes of students' writing) have been studied. One approach common to each involves the comparison of novice and expert writers (Humes, 1983). This comparative approach is most relevant to the present study: Although learning disabled students have not been studied extensively (Newcomer, Nodine, & Barenbaum, 1988), the results of research with novice writers may offer insight into the difficulties that LD writers experience.

In one process study, Hayes and Flower (1983) found that experts had mastered the laws of written prose well enough to focus on the higher level processes of form, meaning, and voice. Novices, however, concentrated on sentence and word level processes to the exclusion of higher level processes. Sommers (1982) reported that experts stratified their revising, concentrating on higher level rhetorical concerns first, then moving through lower level, lexical problems. In contrast, novices concentrated on lexical concerns, to the exclusion of rhetorical issues. Faigley and Witte (1981) also found that expert writers made more revisions at a meaning level, whereas revisions of the novice occurred at a surface level.

The results of product research in which writing samples of LD and nondisabled writers were compared have fit well with process research. Myklebust (1973) found that LD writers were limited in their use of word order, word usage, word endings, and punctuation. Poteet (1978) found that LD students made more punctuation errors and omitted more words than did non-LD writers. Anderson (1982) reported that most errors by LD students were substitutions, additions, and omissions. These students also had difficulty with punctuation. Moreover, Poplin, Gray, Larsen, Bonikowski, and Mehring (1980) reported that differences between LD and non-LD writers increased with age from third through eighth grade.

Although product research has described differences between LD and nondisabled writers, it has not explained why these differences occur. Further investigation into the revision processes of LD and nondisabled writers could shed some light on the question. Errors in the final written products of LD writers could be due to either an absence of revision or to poor revision skills. If poor revision skills are the difficulty, differences could result fro LD students' inability to identify errors or to correct identified errors. Difficulties in locating and correcting errors may be manifestations of a general deficit in communication skills.

A study by MacArthur, Graham, and Skarvoed (1986) found that LD students were unable to correct identified errors: The proportion of errors did not change from initial to revised drafts, despite revisions in spelling, punctuation, and capitalization. The performance of LD students in this study, however, was not compared to that of nondisabled writers. It may be that nondisabled students at the same age or reading level would show the same pattern of performance.

Reading ability may have a significant effect on students' ability to locate and correct errors. Moran (1981) found no differences between the compositions of LD students and low achievers on measures of writing conventions (tense, number, possessive, subject-predicate agreement, and pronoun-referent number agreement) or writing mechanics (punctuation and capitalization). However, the reading levels of the two groups were not presented and implications about the effects of reading skills cannot be drawn directly from this study. Weiner (1980b) reported high correlations between reading and writing levels for both reading disabled (.81) and nondisabled (.91) students. Although these studies indicated that reading and writing are closely related, further research comparing LD students with nondisabled students at the same reading level is in order.

One purpose of the present study was to investigate the differences betweeen LD and nondisabled writers as they revise errors in written compositions. LD students were compared with two groups of nondisabled students: one matched for chronological age and one matched for reading level. Second, this study examined differences in performance under two different methods of revision: reading text and listening to it.

A number of authors have advocated a whole-language approach for improving the ability of writers at all skill levels, including both speaking before composing (Radcliffe, 1972; Zoellner, 1969), and listening to one's writing when revising (Faidman & Howard, 1979; Weiner, 1980a). Little empirical research, however, has been conducted to investigate the use of speaking or listening in writing. MacArthur et al. (1986) compared compositions of learning disabled writers under three different methods of composing: handwriting, word processing, and dictation. The found that stories composed by dictation were longer and of higher quality, and had fewer grammatical errors than those composed under the other two conditions. They attributed these results to differences in composing rate under the three conditions and to the fact that dictation enabled the writers to bypass the mechanical demands involved in handwriting and typing. Another factor possibly effecting these differences, however, was greater proficiency with spoken than written language.

While revising, writers must deal within the mechanical demand of reading their writing. Poor reading proficiency will limit ability to revise. Auditory feedback may enable poor readers to compensate for their reading difficulties by using a form of language with which they have greater proficiency. Existing evidence (Clark, 1983) supports this contention. Listening, in fact has been used successfully with poor readers to overcome difficulties in oral reading and learning content material from textbooks (Rose, 1984; Rose & Sherry, 1984; Sawyer & Kosoff, 1981) and has been proposed as a method for improving reading comprehension (Jenkins, Stein, & Osborn, 1981).

In conclusion, auditory feedback has been proposed as an effective aid in reviing--for both poor writers (who may use listening skills to overcome reading difficulties), and average or above-average writers (who may use listening skills to develop a sense of audience). This proposition, however, has yet to be directly teseted.

The present study was designed to test the following hypotheses:

1. Students who listen to written material containing errors of grammar and syntax will identify more errors than those who read the passage.

2. Learning disabled students will identify fewer errors than nonlearning-disabled students matched for age level, regardless of the mode of presentation.

3. Learning disabled students will identify the same number of errors as nondisabled students matched for reading level, regardless of the mode of presentation.



All 90 students attended rural schools in northcentral Pennsylvania. The 30 sixth- through eighth-grade LD students had been identified as learning disabled according to Pennsylvania Department of Education standards. The second group comprised 30 nondisabled students, matched with LD students for chronological age (CA). The 30 nondisabled students in the third group were matched with LD students for reading (RDG) ability. The LD students were randomly assigned to listening (LD-L) or reading (LD-R) groups, consisting of 15 subjects each. Students in the CA and RDG groups were assigned to L and R conditions based on the matching procedure.

The distributions of boys and girls and their mean grade, age, and reading scores are reported by group in Table 1. The listening and the reading groups did not differ significantly on the proportions of males and females, or on mean grade, age, and reading scores. Students in the LD groups did not differ significantly from students in the CA groups on age and from the students in RDG groups on reading achievement. Intelligence test scores were available for 11 subjects in each LD group. The mean IQ scores were 93.36 (SD = 5.78) and 91.91 (SD = 9.98) for the LD-L and LD-R groups.

LD-L students spent an average of 9.73 hours per week in special education (SD = 6.51); LD-R students averaged 11.93 hours per week (SD = 8.71). All LD subjects were identified by their teachers as needing remedial work in writing. Seven LD-L and nine LD-R students had goals regarding writing skills in their IEPs. All students in the CA and R groups were identified by teachers as average achievers, with grades of "C" or better in language arts.


One reading passage and one set of 25 sentences were used for this study. The 200-word fourth-grade level reading passage was selected from the Specific Skills Reading Series, Getting the Facts, (Boning, 1963). Twenty syntactical errors were placed in the reading passage. Errors conformed to the four categories measured by the syntax section of the Picture Story Language Test (PSLT; Myklebust, 1965): word usage, word order, word endings, and punctuation.

The sentences for the study were derived from the word usage section of the Test of Written Language (TOWL; Hammill & Larsen, 1983). The sentences in this section of the TOWL have one word missing, which students are required to fill in. For the purposes of this study, 20 sentences were completed with a syntactically incorrect word and 5 were randomly chosen to be completed correctly. The incorrect terms were chosen from the test manual list of common errors and were also classified according to the categories of the PSLT. The sentences were used in addition to the reading passage because they isolated errors and minimized the contextual cues available in the passages.

The reading passage and the sentences were recorded. Total reading time for the passage was 5 minutes. The 25 sentences were read 10 seconds apart; total time for this section was 6 minutes. Two pilot trials with LD students were the basis for revisions in error placement, reading, and timing.


Independent and Dependent Variables. The independent variable in this study was the method of presentation. The listening groups listened to the taped passages and sentences; the reading groups read them. The dependent variables for the passage and the sentence tasks were the number of errors identified, the percent of errors corrected, and the number of correct items or phrases marked as incorrect (false errors).

Listening Condition. Students were tested in groups. Before beginning, they were given examples of the types of errors they would find, and corrections for each error were demonstrated. Students listened first to the entire passage without pauses, then to the passage with pauses between sentences. During these pauses, the students circled errors and made corrections. After the passage had been completed, students were given 1 minute to check their work.

The sentences were presented next. Students were given a copy of the 25 sentences and instructed to listen to the sentences and determine if the underlined word was correct or incorrect. Students were instructed to circle and correct errors or, if the sentence were correct, to make no changes. Students were again given examples of possible errors and corrections. The sentences were then read to the students via tape recorder with 10-second pauses between them.

Throughout both passage and sentence sessions, students were encouraged to identify errors, even if they could not correct them. They were encouraged, however, to attempt all corrections.

Reading Condition. Students in the reading condition received the same instructions and examples as the listening group, althoughthey were instructed to read the passages. Students in the reading condition were given the same amount of time for each portion of the passage task: (a) 1 minute to read through the passage; (b) 5 minutes to identify and correct errors; and (c) 1 minute to check the work they had completed. Students were given 6 minutes to complete the sentence task.



Scoring for both the passage and sentence task entailed totaling the number of errors identified, the number corrected, and the number of false errors. An error was considered to be corrected if the change was gramatically and syntactically acceptable and fit into the context of the passage. Incorrect spellings were counted as errors only when they obscured the identity of the word.

To ensure reliable scoring, 18 passages and sentence sets (9 listening and 9 reading) were scored by an independent evaluator. Agreement for the scoring of the passage averaged 94% and ranged from 85% to 100%. Agreement for the scoring of the sentences averaged 98.5% and ranged from 88% to 100%.

Results of Experimental Manipulations

Means and standard deviations for number of errors identified are reported in Table 2. The analysis of variance of the passage task yielded significant main effects for both condition (F (1,84) = 18.54; p [is less than] .001), and group (F (2,84) = 7.58; p [is less than] .001). The condition by group interaction was nonsignificant. Students in the listening condition identified more errors than did students in the reading condition. Follow-up tests of group effect using Tukey's Wholly Significant Difference procedure (Kirk, 1968) indicated that students in the CA group identified more error (p [is less than] .01) than did students in either the LD or RDG groups, which did not differ. The analysis of variance for the sentence task also yielded a significant main effect for group (F (2,84) = 12.32; p [is less than] .001) and condition (F (1,84) = 4.87; p [is less than] .05). The group by condition interaction was nonsignificant. Results of the WSD follow-up test indicated that the CA group identified more errors than did either the LD or RDG groups (p [is less than] .05). Students in the listening condition identified more errors than those in the reading condition. The power of these analyses to detect an effect size of .33--one in which condition accounted for 10% of the total variance--was estimated to be .84 (Cohen, 1969). The power of these analyses to detect an effect size of .33 on the group means was .77.

The means and standard deviations for false errors are presented in Table 2. The analysis of variance for the passage task yielded a significant main effect for group (F (2,84) = 3.63; p [is less than] .05), but not condition. Follow-up tests indicated that the LD group made significantly more false errors than the CA group (p [is less than] .05). There were no significant differences between the LD and RDG group, or the CA and RDG group.

As a final analysis, percent correct was calculated. Results for both the passage and sentence tasks yielded few differences between condition. Students in the listening condition corrected 84% of the identified errors; students in the reading condition corrected 83%. The LD group corrected 77% of the errors it identified, the RDG group 85%, and the CA group, 87%. Analysis of the sentence task yielded similar results. Those in the listening condition corrected 92% of identified errors, while those in the reading condition corrected 95%. The Ld group corrected 91% of the identified errors, the RDG group 95%, and the CA group 94%. None of these differences was statistically significant.


These results support the experimental hypotheses. First, students who listened to the passage and sentences identified more errors than did those who read the same material. The difference was substantial on the passage task (13.89 vs. 10.62) but smaller on the sentence task (15.64 vs. 13.95). Second, LD students identified fewer errors than nondisabled students matched for age, but did not differ from the nondisabled students matched for reading level.

These results are consistent with previous process and product research. Process research has indicated that students master lower level skills before progressing to higher level skills. (Failey & Witte, 1981; Hayes & Flower, 1983; Sommers, 1982). The basic writers in this study--the LD and RDG groups --had greater difficulty than older normally achieving students identifying errors in word order, word usage, word endings, and punctuation. Because LD students have had many of the same learning experiences as their age-matched peers, yet still lag behind, they may need special, remedial help to become proficient with basic skills. The performance of RDG groups, on the other hand, may result from lack of exposure. A conversation with their language arts teacher revealed that students in this group had only begun to study the grammatical skills represented in the passage.

Results from the present study also support previous product research (Anderson, 1982; Myklebust, 1973; Poplin et al., 1980; Poteet, 1978): LD students identified fewer errors in word order, word usage, word ending, and punctuation than did nondisabled students of the same chronological age. Unlike previous research, the present study offers plausible explanations for the difficulties.

First, since the revision task and the amount of time spent in revision were held constant for all groups, results should reflect differences in skill, not in amount of time spent revising. It is possible, of course, that motivation or time on task could have affected the outcomes; however, during all sessions, students appeared to work steadily to task completion, and to work diligently. Thus, differences would seem to be the result of varying skill abilities.

The analysis of false errors also suggests a skill deficit. Identifying a large number of false errors suggests an inability to distinguish between correct and incorrect syntax and grammar, and the LD students identified significantly more false errors than did their nonhandicapped age-mates. This difference also sheds some light on the MacArthur et al. (1986) findings. In addition to failing to recognize errors, the LD students in MacArthur et al. may have falsely revised correct constructions. The task performed by the LD-R group in this study was most similar to the task performed by the students in MacArthur et al. On the average, students in the present study identified 9 of 20 errors (45%) and correctly revised 7 of them (74%), leaving 13 errors uncorrected. In addition, they averaged two false errors for a total, after revision, of 15 errors. When compared to the 20 errors in the original passage, these 15 suggest little improvement through revision--just as MacArthur et al. reported.

The results of the study also support the efficacy of auditory feedback for purposes of revision. Writers at various levels of expertise--elementary age, secondary age, and learning disabled--benefited from auditory input when making revisions in punctuation, grammar, and syntax. The present study does not offer explanations for the increased performance under the listening condition; however, speculations can be made based on other research.

The benefits of listening may derive from increased comprehension of the written material. In previous research (Oakan, Weiner, & Cromer, 1971; Sawyer & Kosoff, 1981; Wiseman, Hartwell, & Hannafin, 1980), comprehension of text material was improved for poor readers through listening. The LD and RDG groups may also have benefited from the use of a mode of language with which they were more skilled: reception of spoken language. As proposed by Leigh (1980) and Giordano (1983), these students may have been able to use their proficiency with the reception of spoken language to make written revisions.

These results must be considered in light of the inherent limitations of the study. First, the passages and sentences were not written by the students. Rather, they were constructed to ensure that each group was exposed to equal numbers and types of errors. That listening will prove an effective technique for students in revising their own writing can only be inferred from the results of this study. This inference must be tested directly in future research, and should shed light on the discrepancies between these findings and those of MacArthur et al. (1986). Second, time was held constant across tasks to prevent repeated readings of the passage or sentences by students in the reading condition. The time limit prevented some subjects in the reading groups from completing the passage and sentence tasks. Future research examining the effects of listening when time is not held constant is in order, although the problem of rereading must be addressed.


Anderson, P L. (1982). A preliminary study of syntax in the written expression of learning disabled children. Journal of Learning Disabilities, 15, 359-362.

Boning, R. (1963). Getting the facts: Specific skills reading series. Baldwin, NY: Barnell Loft.

Clark, I. L. (1983), March). Listening to writing: Implications for evaluation and pedagogy. (ERIC Document Reproduction Service: ED 236 625). Paper presented at the Annual meeting of the Conference on College Composition and Communication, Detroit.

Cohen, J. (1969). Statistical power analysis for the behavioral sciences. New York: Academic Press.

Faidman, C. A., & Howard, T. (1979). Empty pages. Belmont, CA: Fearon Pitman.

Faigley, L., & Witte, S. (1981). Analyzing revision. College Composition and Communication, 32, 400-414.

Giordano, G. (1983). The pivotal role of grammar in correcting writing disabilities. Journal of Special Education, 17,473-496.

Graham, S., & Harris, K. R. (Eds). (1988). Research and instruction in written language [Special issue]. Exceptional Children, 54(6).

Hammill, D. D., & Larsen, S. C. (1983). Test of written language, Austin, TX: Pro-Ed.

Hayes, J. R., & Flower, S. (1983). A cognitive model of the writing process in adults. Final report. (ERIC Document Reproduction Service No. ED 240 608). Pittsburgh, PA: Carnegie-Mellon University.

Humes, A. (1983). Research on the composing process. Review of Educational Research, 53, 201-216.

Jenkins, J. R., Stein, M. L., & Osborn, J. P. (1981). What next after decoding? Instruction and research in reading comprehension. Exceptional Educational Quarterly, 2(1), 27-38.

Kirk, R. E. (1968). Experimental design procedures for the behavioral sciences. Belmont: Brooks-Cole.

Leigh, J. E. (1980). Whole language approaches: Remises and possibilities. Learning Disability Quarterly, 3(4), 62-69.

MacArthur, C. A., Graham, S., & Skarvoed, J. (1986). Learning disabled students' composing with three methods: Handwriting, dictation, and word processing. (Tech. Rep. No. 109). College Park, MD: Institute for the Study of Exceptional Children and Youth.

Moran, M. R. (1981). Performance of learning disabled and low achieving secondary students on formal features of a paragraph writing task. Learning Disability Quarterly, 4, 271-279.

Myklebust, H. R. (1965). Developmental and disorders of written language. Vol. 1. Picture Story Language Test. New York: Grune & Stratton.

Myklebust, H. R. (1973). Developmental and disorders of written language. Vol. 2. Studies of normal and exceptional children. New York: Grune & Stratton.

Newcomer, P., Nodine, B., & Barenbaum, E. (1988). Teaching writing to exceptional children: Reaction and recommendations. Exceptional Children, 54, 559-564.

Oakan, R., Weiner, M., & Cromer, W. (1971). Identification, organization, and reading comprehension for good and poor readers. Journal of Educational Psychology, 62, 71-78.

Poplin, M. S., Gray, R., Larsen, S., Bonikowski, A., & Mehring, T. (1980). A comparison of components of written expression abilities in learning disabled and non-learning disabled students at three grade levels. Learning Disability Quarterly, 3(4), 46-53.

Poteet, J. A. (1978). Characteristics of written expression of learning disabled and non-learning disabled elementary school students. Diagnostique, 4, 60-74.

Radcliffe, T. (1972). Talk-write composition: A theoretical model proposing the use of speech to improve writing. Research in the Teaching of English, 6, 187-199.

Rose, T. L. (1984). The effects of two prepractice procedures on oral reading. Journal of Learning Disabilities, 17, 544-548.

Rose, T. L. & Sherry, L. (1984). Relative effects of two previewing procedures on LD adolescents oral reading performance. Learning Disabilities Quarterly, 7(1), 39-44.

Sawyer, D. J. & Kosoff, T. O. (1981). Accomodating the learning needs of reading disabled adolescents: A language-processing issue. Learning Disability Quarterly, 4(1), 61-67.

Sommers, N. (1982). Revision strategies of student writers and experiences adult writers (ERIC Document Reproduction Service No. ED 220 839). (Contract No. NIE-P-0029). Washington, DC: National Institution of Education.

Weiner, E. S. (1980a). Disagnostic evaluation of writing skills. Journal of Learning Disabilities, 13, 43-48.

Weiner, E. S. (1980b). The diagnostic evaluation of writing skills (DEWS): Application of DEWS criteria to writing samples. Learning Disability Quarterly, 3(2), 54-59.

Wiseman, D. E., Hartwell, L. K., & Hannafin, M. J. (1980). Exploring the reading and listening skills of secondary mildly handicapped students. Learning Disability Quarterly, 3(3), 56-61.

Zoellner, R. (1969). Talk-write: A behavioral pedagogy for composition. College English, 30, 267-320.

CHRISTINE A. ESPIN is Doctoral Student, Educational Psychology, University of Minnesota. PAUL T. SINDELAR is Chairperson, Department of Special Education, University of Florida, Gainesville.
COPYRIGHT 1988 Council for Exceptional Children
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1988 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Espin, Christine A.; Sindelar, Paul T.
Publication:Exceptional Children
Date:Sep 1, 1988
Previous Article:Measuring progress on IEPs: a comparison of graphing approaches.
Next Article:Classwide peer tutoring with mildly handicapped high school students.

Related Articles
Process-oriented writing: instructional implications for mildly handicapped students.
Story composition skills of middle-grade students with learning disabilites.
Helping college bound clients with learning disabilities.
Writing issues in college students with learning disabilities: a synthesis of the literature from 1990 to 2000.
Feedback facilitates the acquisition and retention of numerical fact series by elementary school students with mathematics learning disabilities.
What if young children were grouped for reading with learning-style responsive approaches?
Oral feedback in the context of written language.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters