Two simple approximations to the distributions of quadratic forms
- 1 May 2010
- journal article
- research article
- Published by Wiley in British Journal of Mathematical and Statistical Psychology
- Vol. 63 (2), 273-291
- https://doi.org/10.1348/000711009x449771
Abstract
Many test statistics are asymptotically equivalent to quadratic forms of normal variables, which are further equivalent to with zi being independent and following N(0,1). Two approximations to the distribution of T have been implemented in popular software and are widely used in evaluating various models. It is important to know how accurate these approximations are when compared to each other and to the exact distribution of T. The paper systematically studies the quality of the two approximations and examines the effect of the λi and the degrees of freedom d by analysis and Monte Carlo. The results imply that the adjusted distribution for T can be as good as knowing its exact distribution. When the coefficient of variation of the λi is small, the rescaled statistic is also adequate for practical model inference. But comparing TR against will inflate type I errors when substantial differences exist among the λi, especially, when d is also large.Keywords
This publication has 25 references indexed in Scilit:
- Sufficient Dimension Reduction via Inverse RegressionJournal of the American Statistical Association, 2005
- Performance of Modified Test Statistics in Covariance and Correlation Structure Analysis Under Conditions of Multivariate NonnormalityStructural Equation Modeling: A Multidisciplinary Journal, 2000
- Mean and Covariance Structure Analysis: Theoretical and Practical ImprovementsJournal of the American Statistical Association, 1997
- On Principal Hessian Directions for Data Visualization and Dimension Reduction: Another Application of Stein's LemmaJournal of the American Statistical Association, 1992
- Can test statistics in covariance structure analysis be trusted?Psychological Bulletin, 1992
- Sliced Inverse Regression for Dimension ReductionJournal of the American Statistical Association, 1991
- Algorithm AS 256: The Distribution of a Quadratic Form in Normal VariablesJournal of the Royal Statistical Society Series C: Applied Statistics, 1990
- Likelihood Ratio Tests for Model Selection and Non-Nested HypothesesEconometrica, 1989
- On Chi-Squared Tests for Multiway Contingency Tables with Cell Proportions Estimated from Survey DataThe Annals of Statistics, 1984
- The Significance of the Difference Between Two Means when the Population Variances are UnequalBiometrika, 1938