영재 판별 평정 척도의 타당화 연구

Abstract

본 연구는 다차원적인 관점에서 영재의 개념과 특성에 대한 이론을 형성하여 영재 의 개념모형과 측정모형을 측정하고 이를 근거로 영재의 행동관찰 평정척도를 개발하여 경험적으로 도구의 타당성을 확인하는 것이 목적이다. 이를 위하여 영재의 개념과특성에 관련된 선행 연구들을 분석하여 개념모형을 설정하였으며, 개념모형을 근거로하여 영재의 행동특성과 판별법에 관한 측정도구를 개발하여 그 타당성을 확인하였다. 행동관찰 평정척도의 타당화를 위하여 신뢰도와 타당도를 통한 타당화 과정과 t검증, 외적타당도 검증을 한 결과 본 도구는 신뢰롭고 타당하다는 결론을 얻었다. 방법적으로는 내용타당도와 중다특성-중다방법 분석을 통한 타당도 검증 과정에서 수렴타 당도와 변별타당도를 확인할 수 있었다. 이상의 연구 결과로 본 연구의 목적을 달성하였다. 후속 연구로는 표집의 광역화, 영재의 행동특성에 관한 세분화를 통하여 영역 확대 등이 요구된다.

The purpose of this study are to 1) analyze various theories and preceding studies regarding gifted children and thereby establish some theoretical frameworks defining the concepts and characteristics of gifted children, 2) thereupon design a measurement model and instrument to identify gifted children and 3)verify the rating scale of their behavioral characteristics. For this purpose, characteristics of gifted children were grouped in large into academic, affective and aesthetic areas from multidimensional views ; The first area covers language, mathematics, spacial perception and learning, while the second area embraces creativity, motivation and leadership. And the third area covers visual and performing arts and psychomotor ability. The rating was done by children themselves, parents, teachers and peer group. The result of test can be summarized as follows; 1) In order to test the reliability, those items which may increase the reliability were removed in consideration of the results from self-ratings. As a consequence, the level of reliability for rating of gifted children was .7102～.9226, while that for rating of ordinary ones was .6122～.9397. 2) Based on the results from self-ratings each item was examined for its validity by factor analysis and content validity. Furthermore, each item was checked if it has a significant independence by using such methods as varimax and oblique rotation. 3) In multitrait-multimethod, the convergent validity was tested comparing the coefficients of same traits which was measured by different methods. Meanwhile, the discriminant validity was positively tested in case the correlation between same constructs is higher than that between different constructs. 4) In order to test the difference between gifted and ordinary children, T-tests were done using mean and SD values. As a result, every sub-areas of four rating scales showed significant differences at the level of P<.001, which demonstrates the validity ofthe scales. 5) The correlation between behavioral characteristics and IQ was significant. In conclusion, the rating scale for gifted children tested positively for their validity.