예비유아교사가 지각하는 유아과학교육 필요성과 인공지능 지식, 과학교수태도 및 과학교수효능감 간의 구조적 관계

The Structural Relationship between Pre-Service Early Childhood Teachers' Perceived Need for Early Childhood Science Education and Knowledge of Artificial Intelligence, Attitudes Toward Science Instruction, and Science Teaching Self-Efficacy

본 연구는 예비유아교사가 지각하는 유아과학교육 필요성과 인공지능 지식, 과학교수태도 및 과학교수효능감 간의 구조적 관계를 알아보는데 목적이 있다. 이를 위해 서울, 경기도, 충청도, 경상도 지역의 유아교육과에 재학중인 224명의 데이터를 분석하였다. SPSS 25.0과 AMOS 23.0을 활용하여 연구변인들간의 관계는 피어슨의 상관분석하였고, 구조방정식으로 구조적 관계를 알아보았다. 주요 연구결과는 첫째, 연구변인의 경향성을 보면 모두 3.0점 이상으로 높게 나타났고, 정규성은 만족하는 것으로 분석되었다. 둘째, 내생변인인 과학교수효능감과 하위요인, 과학교수태도와 하위요인을 중심으로 모든 변인들 간의 통계적으로 유의미한 정적 상관관계가 나타났다. 셋째, 측정모형은 적합한 것으로 나타났다. 구조모형의 경로계수를 보면, 유아과학교육 필요성은 인공지능 지식과 과학교수태도, 과학교수효능감에 통계적으로 유의미한 영향을 주는 것으로 나타났다. 인공지능 지식은 과학교수태도와 과학교수효능감에 통계적으로 유의미한 영향을 주는 것으로 나타났다. 매개분석 결과, 유아과학교육 필요성이 과학교수태도와 과학교수효능감에 영향을 주는 과정에서 인공지능 지식이 부분매개하는 것으로 나타났다. 이러한 연구결과는 유아과학교육의 내실화 측면에서 인공지능에 대한 교육 적용 방안을 고려해야 함을 시사하는 것이다.

The purpose of this study was to investigate the structural relationship between pre-service early childhood teachers' perceived need for early childhood science education and their knowledge of artificial intelligence, attitudes toward science instruction, and self-efficacy in science teaching. The subjects of this study were 224 pre-service early childhood teachers in Seoul, Gyeonggi, Chungcheong, and Gyeongsang provinces, and their data were analyzed for the research. The relationships among research variables were analyzed using Pearson's correlation analysis, and the structural relationships were explored through structural equation modeling (SEM) using SPSS 25.0 and AMOS 23.0. These results were as followed. First, when examining the tendencies of research variables, all scored above 3.0 points, indicating a high level, and normality was considered satisfactory in the analysis. Second, statistically significant positive correlations were observed among all variables, with a particular focus on endogenous variables such as science teaching self-efficacy and its sub-factors, as well as attitudes toward science instruction and its sub-factors. Third, the measurement model was found to be appropriate. Examining the path coefficients of the structural model revealed that the perceived need for early childhood science education had statistically significant effects on artificial intelligence knowledge, attitudes toward science instruction, and science teaching self-efficacy. Statistically significant effects of artificial intelligence knowledge were found on attitudes toward science instruction and science teaching self-efficacy. Mediation analysis revealed that knowledge of artificial intelligence partially mediated the influence of perceived need for early childhood science education on attitudes toward science instruction and science teaching self-efficacy. These research findings suggest the need to consider incorporating educational approaches related to artificial intelligence to enhance the foundation of early childhood science education.