지능측정모형의 확인적 연구 - 청소년기를 중심으로

The fittness-test on the hierarchical models and stability of intellectual structure in adolescent and adulthood

본 연구는 Guilford의 지능이론에 근거한 16개의 측정변인을 3단계와 4단계 위계적 지능구조 모형의 적합성을 비교하고 타당한 모형의 발달적 안정성을 확인하였다 연구의 표집은 5556명을 대상으로 하였으며 자료분석을 위해 LISREL프로그램을 사용하였다. 연구결과는 다음과 같다. 위계적 지능구조에 대한 확인적 요인분석 결과는 g를 정점으로하는 지능의 3단계구조모형이 통계적 타당성을 갖는 것으로 나타났다. 또한 청소년기의 지능구조의 발달적 특징을 검증하기 위한 분석결과 연령에 따른 지능구조의 변화에는 의의있는 차이점를 발견하지 못했다.

The porpose of this study is to verify the appropriateness of hierarchical strucure of intelligence with 16 intellectual factors which arc suggested by Song et al.(l997d). In this study I hypochesize 2 competing models based on Gf-Gc theory and Guilford's SOl. Then I explore the best fittness model of 2 hierarchical models. After that, I compares the developmental trats of the hierarchical structure model of the intelligence with the age group of the adolesent and the adulthood. The sample sizes are 5556 including the highschool students and the university student in the area of Seoul, In-cheon, Pusan, Kwang-joo and Kong-joo. The LISREL8 program is used to analyze the data. The results are follows: The confirmatory factor analytic investigation shows the 3rd-order factor model are more suitable than the 4th~order factor model which combined by the SOl. This study supports the third-order factor model putting g-factor on the apex and the second-order factors are Gf-Gc-Gcr. In the hierarchical model 1.3.1 verified as valid excludes the possibility that 16 measurement variables can be related with higher-concept overlappingly. But Carroll(1993), Keith et al(I996) suggested the relaxed- model or other higher-concepts between g and second-order factors. So, model 1.3.1 can be more confirmatory test and modification. In methodological aspect, with using the structure equation model connected factor analysis with path analysis it is possible to measure the error variance within the measurement model. And the latent variables which are the higher-order concept from the 16 intellectual factors and the caual relationships between the latent variables let the structure model lead to the factor structure of intelligence with g at the top of the hierarchical model. The result using the equality constraint among 17, 20, 23 age groups with hierarchical model could not find the significant differences in the three age groups. And this result supports the developmental stability of the intellectual structure. This implies the practical value that the focus on the study of the intelligence after the adolescent needs to investigate the possibilities of the changes in the individual abilities as following the time, the cultural, and the educational factors.