Robust tests for multivariate factorial designs under heteroscedasticity

Abstract

The question of how to analyze several multivariate normal mean vectors when normality and covariance homogeneity assumptions are violated is considered in this article. For the two-way MANOVA layout, we address this problem adapting results presented by Brunner, Dette, and Munk (BDM; 1997) and Vallejo and Ato (modified Brown-Forsythe [MBF]; 2006) in the context of univariate factorial and split-plot designs and a multivariate version of the linear model (MLM) to accommodate heterogeneous data. Furthermore, we compare these procedures with the Welch-James (WJ) approximate degrees of freedom multivariate statistics based on ordinary least squares via Monte Carlo simulation. Our numerical studies show that of the methods evaluated, only the modified versions of the BDM and MBF procedures were robust to violations of underlying assumptions. The MLM approach was only occasionally liberal, and then by only a small amount, whereas the WJ procedure was often liberal if the interactive effects were involved in the design, particularly when the number of dependent variables increased and total sample size was small. On the other hand, it was also found that the MLM procedure was uniformly more powerful than its most direct competitors. The overall success rate was 22.4% for the BDM, 36.3% for the MBF, and 45.0% for the MLM.

Analysis of Psychometric Properties of the Quality and Satisfaction Questionnaire Focused on Sustainability in Higher Education

“…Although the distribution was not normal (Table 1), the literature supports that with very large samples, such as this case, the analysis of variance can be used without restriction [48][49][50][51]. However, for further verification, the Box's M and Wilks' Lambda tests [52][53][54][55] were conducted, resulting in rejection of the null hypothesis (<0.05), which confirms the use of a parametric analysis [56,57]. In order to study the relationships between the variables as proposed in the second of our research objectives, a multilevel analysis was conducted to make several comparisons.…”

The effects of teachers' homework follow-up practices on students' EFL performance: a randomized-group design

“…The hypotheses referring to nullity and parallelism of the regression slopes were tested using SAS PROC MIXED with the solution proposed by Kenward and Roger ( 2009 ). PROC MIXED allows the use of a linear model that relaxes the assumption of constant variance (for details, see Vallejo et al, 2010 ; Vallejo and Ato, 2012 ). The post-hoc contrasts were done using the ESTIMATE expression in SAS PROC MIXED and the BH/FDR option in SAS PROC MULTITEST.…”

Parametric and nonparametric bootstrap methods for general MANOVA

“…In an extensive simulation study involving heterogeneous covariance matrices and nonnormal distributions, Vallejo and Ato [33] investigated the behavior of different small-sample approximations for multivariate tests.…”