The Robustness of the Likelihood Ratio Chi-Square Test for Structural Equation Models: A Meta-Analysis

2001 ◽  
Vol 26 (1) ◽  
pp. 105-132 ◽  
Author(s):  
Douglas A. Powell ◽  
William D. Schafer
Keyword(s):  
Structural Equation ◽  
Structural Equation Models ◽  
Type I Error ◽  
Meta Analysis ◽  
Generalized Least Squares ◽  
Error Rates ◽  
Type I ◽  
Chi Square ◽  
Distribution Free ◽  
Projection Techniques

The robustness literature for the structural equation model was synthesized following the method of Harwell which employs meta-analysis as developed by Hedges and Vevea. The study focused on the explanation of empirical Type I error rates for six principal classes of estimators: two that assume multivariate normality (maximum likelihood and generalized least squares), elliptical estimators, two distribution-free estimators (asymptotic and others), and latent projection. Generally, the chi-square tests for overall model fit were found to be sensitive to non-normality and the size of the model for all estimators (with the possible exception of the elliptical estimators with respect to model size and the latent projection techniques with respect to non-normality). The asymptotic distribution-free (ADF) and latent projection techniques were also found to be sensitive to sample sizes. Distribution-free methods other than ADF showed, in general, much less sensitivity to all factors considered.

scholarly journals Cluster Wild Bootstrapping to Handle Dependent Effect Sizes in Meta-Analysis with a Small Number of Studies

2021 ◽  
Author(s):  
Megha Joshi ◽  
James E Pustejovsky ◽  
S. Natasha Beretvas
Keyword(s):  
Effect Size ◽  
Type I Error ◽  
Meta Analysis ◽  
Error Rates ◽  
Small Sample ◽  
Type I ◽  
Hypothesis Tests ◽  
Type I Error Rates ◽  
Meta Analyses ◽  
Small Sample Correction

The most common and well-known meta-regression models work under the assumption that there is only one effect size estimate per study and that the estimates are independent. However, meta-analytic reviews of social science research often include multiple effect size estimates per primary study, leading to dependence in the estimates. Some meta-analyses also include multiple studies conducted by the same lab or investigator, creating another potential source of dependence. An increasingly popular method to handle dependence is robust variance estimation (RVE), but this method can result in inflated Type I error rates when the number of studies is small. Small-sample correction methods for RVE have been shown to control Type I error rates adequately but may be overly conservative, especially for tests of multiple-contrast hypotheses. We evaluated an alternative method for handling dependence, cluster wild bootstrapping, which has been examined in the econometrics literature but not in the context of meta-analysis. Results from two simulation studies indicate that cluster wild bootstrapping maintains adequate Type I error rates and provides more power than extant small sample correction methods, particularly for multiple-contrast hypothesis tests. We recommend using cluster wild bootstrapping to conduct hypothesis tests for meta-analyses with a small number of studies. We have also created an R package that implements such tests.

scholarly journals The Use of Mixed Models for the Analysis of Mediated Data with Time-Dependent Predictors

2011 ◽  
Vol 2011 ◽  
pp. 1-12 ◽  
Author(s):  
Emily A. Blood ◽  
Debbie M. Cheng
Keyword(s):  
Longitudinal Data ◽  
Mixed Models ◽  
Structural Equation ◽  
Structural Equation Models ◽  
Type I Error ◽  
Time Dependent ◽  
Type I ◽  
Delayed Effects ◽  
Causal Pathways ◽  
Observational Cohort

Linear mixed models (LMMs) are frequently used to analyze longitudinal data. Although these models can be used to evaluate mediation, they do not directly model causal pathways. Structural equation models (SEMs) are an alternative technique that allows explicit modeling of mediation. The goal of this paper is to evaluate the performance of LMMs relative to SEMs in the analysis of mediated longitudinal data with time-dependent predictors and mediators. We simulated mediated longitudinal data from an SEM and specified delayed effects of the predictor. A variety of model specifications were assessed, and the LMMs and SEMs were evaluated with respect to bias, coverage probability, power, and Type I error. Models evaluated in the simulation were also applied to data from an observational cohort of HIV-infected individuals. We found that when carefully constructed, the LMM adequately models mediated exposure effects that change over time in the presence of mediation, even when the data arise from an SEM.

scholarly journals Mendelian Randomization Analysis Using Multiple Biomarkers of an Underlying Common Exposure

2021 ◽  
Author(s):  
Jin Jin ◽  
Guanghao Qi ◽  
Zhi Yu ◽  
Nilanjan Chatterjee
Keyword(s):  
Structural Equation ◽  
Type I Error ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Association Studies ◽  
Error Rates ◽  
Type I ◽  
Genome Wide Association Studies ◽  
Increased Risk ◽  
Multiple Biomarkers

AbstractMendelian Randomization (MR) analysis is increasingly popular for testing the causal effect of exposures on disease outcomes using data from genome-wide association studies. In some settings, the underlying exposure, such as systematic inflammation, may not be directly observable, but measurements can be available on multiple biomarkers, or other types of traits, that are co-regulated by the exposure. We propose method MRLE, which tests the significance for, and the direction of, the effect of a latent exposure by leveraging information from multiple related traits. The method is developed by constructing a set of estimating functions based on the second-order moments of summary association statistics, under a structural equation model where genetic variants are assumed to have indirect effects through the latent exposure and potentially direct effects on the traits. Simulation studies showed that MRLE has well-controlled type I error rates and increased power compared to single-trait MR tests under various types of pleiotropy. Applications of MRLE using genetic association statistics across five inflammatory biomarkers (CRP, IL-6, IL-8, TNF-α and MCP-1) provided evidence for potential causal effects of inflammation on increased risk of coronary artery disease, colorectal cancer and rheumatoid arthritis, while standard MR analysis for individual biomarkers often failed to detect consistent evidence for such effects.

scholarly journals The Effect of Publication Bias on the Assessment of Heterogeneity

2017 ◽  
Author(s):  
Hilde Augusteijn ◽  
Robbie Cornelis Maria van Aert ◽  
Marcel A. L. M. van Assen
Keyword(s):  
Publication Bias ◽  
Effect Size ◽  
Type I Error ◽  
Meta Analysis ◽  
Error Rates ◽  
Population Heterogeneity ◽  
Type I ◽  
Monte Carlo Simulation Study ◽  
True Effect Size ◽  
True Effect

One of the main goals of meta-analysis is to test and estimate the heterogeneity of effect size. We examined the effect of publication bias on the Q-test and assessments of heterogeneity, as a function of true heterogeneity, publication bias, true effect size, number of studies, and variation of sample sizes. The expected values of heterogeneity measures H2 and I2 were analytically derived, and the power and the type I error rate of the Q-test were examined in a Monte-Carlo simulation study. Our results show that the effect of publication bias on the Q-test and assessment of heterogeneity is large, complex, and non-linear. Publication bias can both dramatically decrease and increase heterogeneity. Extreme homogeneity can occur even when the population heterogeneity is large. Particularly if the number of studies is large and population effect size is small, publication bias can cause both extreme type I error rates and power of the Q-test close to 0 or 1. We therefore conclude that the Q-test of homogeneity and heterogeneity measures H2 and I2 are generally not valid in assessing and testing heterogeneity when publication bias is present, especially when the true effect size is small and the number of studies is large. We introduce a web application, Q-sense, which can be used to assess the sensitivity of the Q-test to publication bias, and we apply it to two published meta-analysis. Meta-analytic methods should be enhanced in order to be able to deal with publication bias in their assessment and tests of heterogeneity.

scholarly journals Analysis of Variance Models with Stochastic Group Weights

2019 ◽  
Author(s):  
Axel Mayer ◽  
Felix Thoemmes
Keyword(s):  
Analysis Of Variance ◽  
Error Rate ◽  
Latent Variables ◽  
Structural Equation Models ◽  
Type I Error ◽  
Error Rates ◽  
Type I ◽  
Type I Error Rate ◽  
Inflated Type ◽  
Group Weights

The analysis of variance (ANOVA) is still one of the most widely used statistical methods in the social sciences. This paper is about stochastic group weights in ANOVA models – a neglected aspect in the literature. Stochastic group weights are present whenever the experimenter does not determine the exact group sizes before conducting the experiment. We show that classic ANOVA tests based on estimated marginal means can have an inflated type I error rate when stochastic group weights are not taken into account, even in randomized experiments. We propose two new ways to incorporate stochastic group weights in the tests of average effects - one based on the general linear model and one based on multigroup structural equation models (SEMs). We show in simulation studies that our methods have nominal type I error rates in experiments with stochastic group weights while classic approaches show an inflated type I error rate. The SEM approach can additionally deal with heteroscedastic residual variances and latent variables. An easy-to-use software package with graphical user interface is provided.

Random-effects meta-analysis: the number of studies matters

2015 ◽  
Vol 26 (3) ◽  
pp. 1500-1518 ◽  
Author(s):  
Annamaria Guolo ◽  
Cristiano Varin
Keyword(s):  
Random Effects ◽  
Type I Error ◽  
Meta Analysis ◽  
Practical Interest ◽  
Error Rates ◽  
Type I ◽  
Model Framework ◽  
Random Effects Models ◽  
Meta Analyses ◽  
The Impact

This paper investigates the impact of the number of studies on meta-analysis and meta-regression within the random-effects model framework. It is frequently neglected that inference in random-effects models requires a substantial number of studies included in meta-analysis to guarantee reliable conclusions. Several authors warn about the risk of inaccurate results of the traditional DerSimonian and Laird approach especially in the common case of meta-analysis involving a limited number of studies. This paper presents a selection of likelihood and non-likelihood methods for inference in meta-analysis proposed to overcome the limitations of the DerSimonian and Laird procedure, with a focus on the effect of the number of studies. The applicability and the performance of the methods are investigated in terms of Type I error rates and empirical power to detect effects, according to scenarios of practical interest. Simulation studies and applications to real meta-analyses highlight that it is not possible to identify an approach uniformly superior to alternatives. The overall recommendation is to avoid the DerSimonian and Laird method when the number of meta-analysis studies is modest and prefer a more comprehensive procedure that compares alternative inferential approaches. R code for meta-analysis according to all of the inferential methods examined in the paper is provided.

scholarly journals Robust Chi-Square in Extreme and Boundary Conditions: Comments on Jak et al. (2021)

Psych ◽  
2021 ◽  
Vol 3 (3) ◽  
pp. 542-551
Author(s):  
Tihomir Asparouhov ◽  
Bengt Muthén
Keyword(s):  
Boundary Conditions ◽  
Type I Error ◽  
Error Rates ◽  
Type I ◽  
Chi Square ◽  
Type I Error Rates ◽  
Chi Square Test

In this article we describe a modification of the robust chi-square test of fit that yields more accurate type I error rates when the estimated model is at the boundary of the admissible space.

scholarly journals Evaluating Equivalence Testing Methods for Measurement Invariance

2020 ◽  
Author(s):  
Alyssa Counsell ◽  
Rob Cribbie
Keyword(s):  
Measurement Invariance ◽  
Goodness Of Fit ◽  
Type I Error ◽  
Error Rates ◽  
Type I ◽  
Equivalence Testing ◽  
Fit Indices ◽  
Equivalence Test ◽  
Chi Square ◽  
Difference Test

Measurement Invariance (MI) is often concluded from a nonsignificant chi-square difference test. Researchers have also proposed using change in goodness-of-fit indices (ΔGOFs) instead. Both of these commonly used methods for testing MI have important limitations. To combat these issues, To combat these issues, it was proposed using an equivalence test (EQ) to replace the chi-square difference test commonly used to test MI. Due to concerns with the EQ's power, and adjusted version (EQ-A) was created, but provides little evaluation of either procedure. The current study evaluated the Type I error and power of both the EQ and EQ-A, and compared their performance to that of the traditional chi-square difference test and ΔGOFs. The EQ was the only procedure that maintained empirical error rates below the nominal alpha level. Results also highlight that the EQ requires larger sample sizes than traditional difference-based approaches or using equivalence bounds based on larger than conventional RMSEA values (e.g., > .05) to ensure adequate power rates. We do not recommend the proposed adjustment (EQ-A) over the EQ.

ModL: exploring and restoring regularity when testing for positive selection

2018 ◽  
Vol 35 (15) ◽  
pp. 2545-2554 ◽  
Author(s):  
Joseph Mingrone ◽  
Edward Susko ◽  
Joseph P Bielawski
Keyword(s):  
Positive Selection ◽  
Likelihood Ratio ◽  
Type I Error ◽  
Error Rates ◽  
Ratio Test ◽  
Type I ◽  
Chi Square ◽  
Type I Error Rates ◽  
Modified Likelihood Ratio Test ◽  
Modified Likelihood

Abstract Motivation Likelihood ratio tests are commonly used to test for positive selection acting on proteins. They are usually applied with thresholds for declaring a protein under positive selection determined from a chi-square or mixture of chi-square distributions. Although it is known that such distributions are not strictly justified due to the statistical irregularity of the problem, the hope has been that the resulting tests are conservative and do not lose much power in comparison with the same test using the unknown, correct threshold. We show that commonly used thresholds need not yield conservative tests, but instead give larger than expected Type I error rates. Statistical regularity can be restored by using a modified likelihood ratio test. Results We give theoretical results to prove that, if the number of sites is not too small, the modified likelihood ratio test gives approximately correct Type I error probabilities regardless of the parameter settings of the underlying null hypothesis. Simulations show that modification gives Type I error rates closer to those stated without a loss of power. The simulations also show that parameter estimation for mixture models of codon evolution can be challenging in certain data-generation settings with very different mixing distributions giving nearly identical site pattern distributions unless the number of taxa and tree length are large. Because mixture models are widely used for a variety of problems in molecular evolution, the challenges and general approaches to solving them presented here are applicable in a broader context. Availability and implementation https://github.com/jehops/codeml_modl Supplementary information Supplementary data are available at Bioinformatics online.

Sign in / Sign up

Export Citation Format

Share Document