PsychStatsWiki

• 被试内设计，两组数据不存在组间差异
• Within subjects design has no difference between the two groups of data
• 匹配组t-test是有一个或若干个特征使得被试内两两建立联系，这种联系时实验前建立的，分析数据时已匹配成对
• The matching group t-test has one or more characteristics that make the participants establish a relationship between two pairs, and this connection is established before the experiment and matched in pairs when the data is analyzed.
• 匹配能大大减少个体间的误差，提高统计效力
• Matching can greatly reduce the error between individuals and improve the statistical efficiency.

—-

• 相关样本t统计量的计算基于样本分数的差异，每一对对应数据的差异D构成了一个差异样本
• The calculation of the t-statistic of the relevant samples is based on the difference in the sample scores, and the difference D of each pair of corresponding data constitutes a difference sample.
• 平均数D=X1-X2
• 方差

• 在每一种处理条件内，观察都彼此独立
• Within each treatment condition, the observations are independent of each other
• 差异分数的总体分布是正态的
• The overall distribution of difference scores is normal
• 不需要考虑方差同质性
• Variance homogeneity does not need to be considered

• 效应大小=差异样本均值/差异样本方差
• Effect size = difference sample mean/difference sample variance