학습장애아 판별방법간 비교 연구 = A Comparative Study of Methods for Identifying Learning Disabled Children

다국어 초록 (Multilingual Abstract)

Despite the fact that criteria for the identification of students with learning disabilities(LD) have generated substantial controversy, a severe discrepancy between expected and actual achievement levels has emerged as the predominant element to identify LD students in this field. The purpose of the present study is to examine the differential impact of the use of four methods for determining a severe discrepancy on the identification of students with possible LD.
For this purpose, 100 subjects with IQ above 70 were randomly selected from 4 to 6 grades in special classes of public elementary schools in Seoul. The date were collected utilizing the KEDI-WISC and the reading and arithmetic subtests of the individual Achievement Test. The four methods applied to determine a severe discrepancy were those based on (1) deviation from grade level (M1), (2) grade expectancy formula (M2), (3) standard score comparison (M3), and (4) regression model (M4). To verify the differences among the four methods, this study employed percentage of students identified as LD and the X²-test. The degree of agreement of the identified subjects was examined by Cohens Kappa statistic and Pearsons correlation coefficient across different methods.
The results obtained in this study are summarized as follows :
1. The M3 approach identified the lowest number of students (64%) as LD, and the M4 and M1 procedures identified the most LD students (92% & 93%) to present significant differences in the percentages of subjects identified between pairs of the four methods except the pair of M1 and M4.
2. The degree of agreement measured by Kappa statistic and Pearsons was fair between M2 and M3 and between M2 and M4, with the rest of the comparisons being relatively low.
3. The M4 procedure was able to identify students with high IQ and severe academic problems as well as those with low IQ and low achievement. While the M3 approach resulted in the under-identification of students with IQs 70-84 and over-identification of students with IQs above 85, vice versa resulted in the M1 approach. The M2 approach showed the tendency which appeared in the M3 approach.
4. In both reading and arithmetic areas, the LD students who met the severe discrepancy criterion using a regression method failed to meet the criterion when using the standard score method and the grade expectancy formula method.
5. Almost all of the LD students with the high IQs (IQ 100-111) exhibited severe discrepancies of more than 20 or 30 points between expected and actual achievement levels.
In conclusion, the present study recommended that if severe discrepancy methods are to be used to identify students with LD, the method taking regression into account be employed as the most appropriate procedure that results in proportional identification of students with severe discrepancies regardless of IQ level.

목차 (Table of Contents)

• Ⅰ. 서론
• A. 연구의 필요성 및 목적
• B. 연구 문제
• Ⅱ. 이론적 배경
• A. 학습장애 판별기준
• B. 기대치-성취도간 격차 산출 방법
• C. 선행 비교 연구
• Ⅲ. 연구 방법
• A. 연구 대상
• B. 연구 도구
• C. 자료 수집 절차
• D. 자료 처리 방법
• Ⅳ. 결과
• Ⅴ. 논의 및 결론
• A. 논의
• B. 결론
• 참고문헌
• 〈부록1〉백분위점수에 따른 표준점수 환산표
• 〈부록2〉회귀현상을 고려하기 전후의 기대 성취 수준
• Abstract