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。

**F统计值（F-statistic）**
----
  *F=处理间均方/处理内均方
  *F = Mean Square Between / Mean Square Within
  *F分布（F-distribution）{{:f分布.png?400|}}
  *当F统计量的观测值小于临界值时，说明数据的总变异中，大部分是由实验误差和个体差异所致，不能认为实验处理有效；
  *When the observed value of the F statistic is smaller than the critical value, it indicates that most of the total variation in the data is due to experimental error and individual differences, and thus the effectiveness of the experimental treatment cannot be assumed.
  *当F统计量的观测值大于临界值时，说明实验数据的变异的确由不同的实验处理所造成，即不同处理之间存在差异。
  *When the observed value of the F statistic is greater than the critical value, it indicates that the variation in the experimental data is indeed caused by different experimental treatments, implying differences between the treatments.
  *使用F统计量的前提：1.总体正态；2.变异的同质性；3.独立性。
  *Premises for using the F statistic: 1. Normality of the population; 2. Homogeneity of variance; 3. Independence.