교육적 변화의 실험ㆍ통계적 분석방법에 관한 논의 = A Discussion on the Experimental and Statistical Methods for The Analysis of Educational Chage

다국어 초록 (Multilingual Abstract)

Some methodological alternatives for the analysis of educational change, specifically for eliminating the influence of pre-treatment differences on post test scores were disussed.
Experimental design consider action comparing past-test scores with assuming independent and equivalent samples, analysis model of difference scores comparing scores of subtraction of a pre-test score from each post-test score, analysis model of residual gain scores comparing scores of deviation of the post score from the post -test/pre- test regression line, repeated measures analysis of variance model including the pre- treatment differences as one dimension of analysis, and analysis of covariance model using the pre-test differences as a covariate can be used for the analysis of educational change.
But there are problems such as the difficulty of true randomization including independent and equivalent sampling in the experimental design consideration in educational settings, such as the risk of being primarily composed of error variance in the statistics for the statistical analysis in the perspective of measurement theory. Also, there are problems that the measurement and meaurement level of education variables are not suited well to the general as sumptions of statistical models.
Therefore, of course, above- mentioned methods for the analysis of educational change can be used, but the interpretation of those analysis results must be done in the context of their limitations.

목차 (Table of Contents)

• Ⅰ.문제의 진술
• Ⅱ.변화 분석모형과 그 논의
• 1.실험설계적 고려
• 2.차이점수에 의한 분석
• 3.잔여획득점수에 의한 분석
• 4.반복측정의 변량분석
• 5.공변량분석
• Ⅲ.맺는말