Power Comparison of EGLS Test Statistic for Fixed Effects with Arbitrary Distributions



Lee, Jang-Taek

  • 발행 : 2003.04.01


Quite often normality assumptions are not satisfied in practical applications. In this paper, an estimated generalized least squares(EGLS) analysis are considered in two way mixed linear models with arbitrary types of distributions for random effects. We investigate the power performance of EGLS analysis based on Henderson's method III, ML, REML and MINQUE(1). The power performances depend on the imbalance of design, on the actual values of ratio of variance components, and on the skewness and kurtosis parameters of the underlying distributions slightly. Results of our limited simulation study suggest that the EGLS F-statistics using four estimators and arbitrary distributions produce similar type I error rates and power performance.


mixed models;estimated generalized least squares


  1. Biometrical Journal v.29 pp.383-396 Measures of Imbalance for Unbalanced Models Khuri, A. I. https://doi.org/10.1002/bimj.4710290402
  2. Journal of Quality Technology v.33 no.3 pp.265-278 Examples of Designed Experiments With Nonnormal Responses Lewis, S. L.;Montgomery, D. C.;Myers, R. H.
  3. 응용통계연구 v.10 no.1 pp.177-187 이원혼합모형에서 고정효과 유의성검정에 대한 검정력 분석 이장택
  4. Estimation of Variance Components and Applications Rao, C. R. ;Kleffe, J.
  5. Linear Models for Unbalanced Data Searle, S. R.