체육학 연구에서 다층자료분석의 활용 = Application of multilevel data analysis to sports science research

다국어 초록 (Multilingual Abstract)

Correct and accurate inference depends on an appropriate experimental design and a proper statistical model. Most of data in many sports science studies multilevel structures, but the structures have been ignored in data analysis.
The purpose of this study was to introduce multilevel data analysis and to explore the possible applications to the study of physical education and sports science.
The multilevel data analysis have two important principles. The first is that observational and experimental units should be distinguished and taken into consideration in analyzing multilevel data. This can eliminate the problems of the classical statistical model like ANOVA or regression analysis. And the second is to build to an analysis based on the data structure. The major advantages of the multilevel data analysis in analyzing multilevel data include improving the estimation of individual performance, modeling cross-level effects, and partitioning variance-covariance components(Bryk & Raudenbush, 1992).
Furthermore, the multilevel data analysis can be used to assess the change, such as studying the structure of individual growth and estimating important statistical and psychometric properties of collections of growth trajectories, discovering correlates of initial states and change rates, and testing hypotheses about the effects of one or more experimental treatments on growth curves.
Through reviewing prior studies in related to multilevel data analysis, improving inference in small samples, testing cross-level effects, partitioning variance-covariance components, meta-analytical study, and assessing the change of physical fitness or human performance have been proposed as the possible applications for physical education and sports science.
In conclusion, for more valid inference in sports science study using the hierarchical data, multilevel data analysis and the possible applications should be considered.

목차 (Table of Contents)

• I.연구의 필요성
• II.다층자료분석의 이해
• 1.다층 자료의 개념
• 2.다층자료분석 통계모형의 발달
• 3.다층자료분석의 기본통계모형
• III.변화성장모형으로서의 다층자료분석
• 1.변화 평가 방법의 재개념화
• 2.변화성장모형의 기본 모형
• 3.변화성장모형의 분석 단계
• IV.체육학 연구에서의 활용방안
• 1.소집단 추정 효과의 개선
• 2.수준간 효과의 가설 검정
• 3.수준별 분산ㆍ공분산의 분할
• 4.메타분석적 연구
• 5.변화성장연구
• V.요약 및 결론