PsychStatsWiki

本页面只读。您可以查看源文件，但不能更改它。如果您觉得这是系统错误，请联系管理员。
==== 相关样本的意义 The significance of the relevant test====
  * 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检验的逻辑 The basic logic of revelent sample t-test ====
  * 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{{::方差.png?200|}}


----
==== 相关样本t的计算 The calculation of relevent sample t-statistics ====
  * {{::计算相关样本t.png?200|}}


----
==== 相关样本t检验的统计前提 Statistical premise of the t-test of relevant samples====
  * 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

----
==== 相关样本t检验的效应大小 Effect size of relevent sample t-test====
  * 效应大小=差异样本均值/差异样本方差
  * Effect size = difference sample mean/difference sample variance