Robustness with violations of heteroscedasticity assumption.
收藏NIAID Data Ecosystem2026-03-07 收录
数据链接：
官方服务：
资源简介：
Rates of type I errors in repeated-measures Anovas and UKS test for 3 representative experimental designs (lines) and the same 4 degrees of heteroscedasticity as in Figure 3A (columns). Rates are averages of designs with 5, 10 and 20 individuals. The rates of each design are equal to the percentages of 60000 random datasets found significant at the 0.05 threshold as the effect of factor was set to zero. Bold values indicate large excess of type I errors. UKS (and RM Anova) are globally robust to violations of heteroscedasticity.
创建时间：
2012-06-18
