The Effect of Skewness and Kurtosis on Mean and Covariance Structure Analysis

The maximum likelihood (ML) method, based on the normal distribution assumption, is widely used in mean and covariance structure analysis. With typical nonnormal data, the ML method will lead to biased statistics and inappropriate scientific conclusions. This article develops a simple but informative case to show how ML results are influenced by skewness and kurtosis. Specifically, the authors discuss how skewness and kurtosis in a univariate distribution affect the standard errors of the ML estimators, the covariances between the estimators, and the likelihood ratio test of hypotheses on mean and variance parameters. They also describe corrections that have been developed to allow appropriate inference. Enough details are provided so that this material can be used in graduate instruction. For each result, the corresponding results in the higher dimensional case are pointed out, and references are provided.

[1]  A. Satorra,et al.  Power of the likelihood ratio test in covariance structure analysis , 1985 .

[2]  K. Yuan,et al.  Structural equation modeling with heavy tailed distributions , 2004 .

[3]  S. Breckler.  Applications of covariance structure modeling in psychology: cause for concern? , 1990, Psychological bulletin.

[4]  K. Bollen.  Latent variables in psychology and the social sciences. , 2002, Annual review of psychology.

[5]  K. Adachi.  OBLIQUE PROMAX ROTATION APPLIED TO THE SOLUTIONS IN MULTIPLE CORRESPONDENCE ANALYSIS , 2004 .

[6]  K. Yuan,et al.  Cross-validation by downweighting influential cases in structural equation modelling. , 2002, The British journal of mathematical and statistical psychology.

[7]  Ke-Hai Yuan,et al.  Robust mean and covariance structure analysis through iteratively reweighted least squares , 2000 .

[8]  Ke-Hai Yuan,et al.  Normal theory likelihood ratio statistic for mean and covariance structure analysis under alternative hypotheses , 2007 .

[9]  K. Yuan,et al.  Structural Equation Modeling with Small Samples: Test Statistics. , 1999, Multivariate behavioral research.

[10]  Alexander Shapiro,et al.  Asymptotic distribution theory in the analysis of covariance structures , 1983 .

[11]  Wai-Yin Poon,et al.  A distribution free approach for analysis of two-level structural equation model , 1994 .

[12]  Anthony G. Greenwald,et al.  Applications of Covariance Structure Modeling in Psychology : Cause for Concern ? , 2001 .

[13]  K. Yuan,et al.  9. Structural Equation Modeling with Robust Covariances , 1998 .

[14]  Ke-Hai Yuan,et al.  On asymptotic distributions of normal theory MLE in covariance structure analysis under some nonnormal distributions , 1999 .

[15]  A. Shapiro,et al.  Robustness of normal theory methods in the analysis of linear latent variate models. , 1988 .

[16]  Ke-Hai Yuan,et al.  F Tests for Mean and Covariance Structure Analysis , 1999 .

[17]  W. J. Langford.  Statistical Methods , 1959, Nature.

[18]  K. Mardia.  Measures of multivariate skewness and kurtosis with applications , 1970 .

[19]  T. Micceri.  The unicorn, the normal curve, and other improbable creatures. , 1989 .

[20]  B. Tabachnick,et al.  Using multivariate statistics, 5th ed. , 2007 .

[21]  G. Arminger,et al.  Specification and Estimation of Mean- and Covariance-Structure Models , 1995 .

[22]  Taeke Klaas Dijkstra.  Latent variables in linear stochastic models , 1981 .

[23]  J. Magnus,et al.  Matrix Differential Calculus with Applications in Statistics and Econometrics (Revised Edition) , 1999 .

[24]  Peter M. Bentler,et al.  Covariance structure analysis with heterogeneous kurtosis parameters , 1990 .

[25]  Ke-Hai Yuan,et al.  On normal theory based inference for multilevel models with distributional violations , 2002 .

[26]  A. Boomsma.  Reporting Analyses of Covariance Structures , 2000 .

[27]  T. W. Anderson,et al.  The asymptotic normal distribution of estimators in factor analysis under general conditions , 1988 .

[28]  B. Tabachnick,et al.  Using Multivariate Statistics , 1983 .

[29]  P M Bentler,et al.  Robust transformation with applications to structural equation modelling. , 2000, The British journal of mathematical and statistical psychology.

[30]  J. Magnus,et al.  Matrix Differential Calculus with Applications in Statistics and Econometrics , 1991 .

[31]  K. Yuan,et al.  Robust mean and covariance structure analysis. , 1998, The British journal of mathematical and statistical psychology.

[32]  Peter M. Bentler,et al.  Improving parameter tests in covariance structure analysis , 1997 .

[33]  A. Satorra.  Alternative test criteria in covariance structure analysis: A unified approach , 1989 .

[34]  Stephen E. Fienberg,et al.  Discrete Multivariate Analysis: Theory and Practice , 1976 .

[35]  K. Yuan.  Fit Indices Versus Test Statistics , 2005, Multivariate behavioral research.

[36]  A. Shapiro,et al.  On the multivariate asymptotic distribution of sequential Chi-square statistics , 1985 .

[37]  Ke-Hai Yuan,et al.  Eight test statistics for multilevel structural equation models , 2003, Comput. Stat. Data Anal..

[38]  T. Ferguson.  A Course in Large Sample Theory , 1996 .

[39]  Douglas G. Bonett,et al.  Estimating p-values for Mardia's coefficients of multivariate skewness and kurtosis , 2002, Comput. Stat..

[40]  K. Yuan,et al.  Bootstrap approach to inference and power analysis based on three test statistics for covariance structure models. , 2003, The British journal of mathematical and statistical psychology.

[41]  Albert Satorra,et al.  Model Conditions for Asymptotic Robustness in the Analysis of Linear Relations , 1990 .

[42]  A. Shapiro,et al.  Analysis of Covariance Structures under Elliptical Distributions , 1987 .

[43]  Ke-Hai Yuan,et al.  On robusiness of the normal-theory based asymptotic distributions of three reliability coefficient estimates , 2002 .

[44]  Ke-Hai Yuan,et al.  Mean Comparison: Manifest Variable Versus Latent Variable , 2006 .

[45]  K. Yuan,et al.  A NEW MEASURE OF MISFIT FOR COVARIANCE STRUCTURE MODELS , 2004 .

[46]  P M Bentler,et al.  Normal theory based test statistics in structural equation modelling. , 1998, The British journal of mathematical and statistical psychology.

[47]  T. K. Dijkstra.  Latent variables in linear stochastic models : refletions on "maximum likelihood" and "partial least squares" methods , 1985 .

[48]  P. Bentler.  Some contributions to efficient statistics in structural models: Specification and estimation of moment structures , 1983 .

[49]  P M Bentler,et al.  Effect of outliers on estimators and tests in covariance structure analysis. , 2001, The British journal of mathematical and statistical psychology.

[50]  Peter M. Bentler,et al.  Statistical Inference Based on Pseudo-Maximum Likelihood Estimators in Elliptical Populations , 1993 .

[51]  Peter M. Bentler,et al.  EQS : structural equations program manual , 1989 .

[52]  Michael E. Sobel,et al.  Pseudo-Maximum Likelihood Estimation of Mean and Covariance Structures with Missing Data , 1990 .

[53]  P. Bentler,et al.  Robustness of normal theory statistics in structural equation models , 1991 .

[54]  T. W. Anderson,et al.  Asymptotic Chi-Square Tests for a Large Class of Factor Analysis Models , 1990 .

[55]  K. Yuan,et al.  5. Three Likelihood-Based Methods for Mean and Covariance Structure Analysis with Nonnormal Missing Data , 2000 .

[56]  Ke-Hai Yuan,et al.  Inferences on Correlation Coefficients in Some Classes of Nonnormal Distributions , 2000 .

[57]  Y Kano,et al.  Can test statistics in covariance structure analysis be trusted? , 1992, Psychological bulletin.

[58]  A. Satorra,et al.  Corrections to test statistics and standard errors in covariance structure analysis. , 1994 .

[59]  Ke-Hai Yuan,et al.  ON NORMAL THEORY AND ASSOCIATED TEST STATISTICS IN COVARIANCE STRUCTURE ANALYSIS UNDER TWO CLASSES OF NONNORMAL DISTRIBUTIONS , 1999 .

[60]  Ronald Schoenberg,et al.  Pseudo maximum likelihood estimation and a test for misspecification in mean and covariance structure models , 1989 .

[61]  M. Browne.  Asymptotically distribution-free methods for the analysis of covariance structures. , 1984, The British journal of mathematical and statistical psychology.

[62]  Ke-Hai Yuan,et al.  Mean and Covariance Structure Analysis: Theoretical and Practical Improvements , 1997 .

[63]  R. MacCallum,et al.  Applications of structural equation modeling in psychological research. , 2000, Annual review of psychology.