线性混合了效果与斜正常的错误和没有分发的随机的效果当模特儿被考虑。ANOVA 类型 F 测试被建议测试随机的意义固定效果上的效果和假设分别地感兴趣。测试被证明到是在这下面的准确 F 测试模型,和为兴趣的固定效果的准确信心间隔被导出。模拟结果被给学习 ANOVA 类型的力量测试。
Linear mixed effect models with skew-normal errors and distribution-free random effects are considered. The ANOVA-type F-tests are proposed to test the significance of random effects and the hypothesis on fixed effects of interest, respectively. Both tests are proved to be exact F-tests under this model, and the exact confidence interval for fixed effects of interest is derived. Simulation results are given to study the powers of ANOVA-type tests.