反復測定에 의한 三元 變量分析 프로그램 開發硏究 = Statisitical Program of 2-Factor Analysis of Variance with Repeated Measures

다국어 초록 (Multilingual Abstract)

This study is attempted to develop a statistical program of 2-way ANOVA with repeated measures. Factorial experiments in which the same experimental subject is observed under more than one treatment condition are referred to as designs with repeated measures. These designs represent a reasonable proportion of the research conducted in the natural and social sciences.
In this program, the in-depth analyses of interaction effects are undertaken to identify the locus of the significant interaction as well as the analysis of main effects. Additionally, the emphasis of this program is placed on making the results into tables for the convenience of reserarchers.
The program is written in the IBM personal computer BASIC(version A2.10,1983) and has been implemented on SUPER 5(compatible with IBM-XT). The sequence of the program is as follows: (1)introduction of the program, (2) data-Input, (3) statistical computation of the data, (4) analysis of main effects, (5) analyses of simple main effects, (6) multiple comparison between marginal and/or cell means.
The user is first asked to respond to a series of questions concerning the factorial design, such as number of levels of variable A and R, and cell size. Then, raw data points can be entered into the computer by two methods: typed directly through the key board or read from the disk. After data-input, the user has a menu for various options, such as data-displaying for verification and correction, data-printing, data-saving into disk and data-processing for statistical computation.
After the computation has been performed, the user has another selection menu for printing or video-display of the results, such as: (1) a summary table of basic statistics(n's, means, and SD's),(2) an ANOVA table with sources of variance, SS's, df's, MS's, F-ratios, significance levels, and probabilities, (3) summary tables of simple main effects. In addition, the user has the option of (4) Newman-Keul's tests for multiple comparison between marginal and/or cell means, if one or both of the variables have more than three levels.
CONCLUSION: This program is started just with an attempt to demonstrate a way of applying the personal computers to developing statistical programs. It is hopefully expected that lots of subsequent efforts would be continued and add up this program, and finally a complete package of ANOVA programs could be performed.

목차 (Table of Contents)

• Ⅰ. 緖 論
• Ⅱ. 硏究道具 및 節次
• Ⅲ. 硏究結果
• Ⅳ. 結 論
• 參考文獻
• 附 錄