The Model Size Effect in SEM: Inflated Goodness-of-Fit Statistics Are Due to the Size of the Covariance Matrix

By Morten Moshagen. 2012

203 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Management Science and Operations Research, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/2d3ed5003548689b441d706782c46c28dbecb84f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

In Management Science and Operations Research

In Statistics and Probability

Citations in the corpus

Estimating the Maximum Likelihood Root Mean Square Error of Approximation (RMSEA) with Non-normal Data: A Monte-Carlo Study (2020, background): “However, when sample size is small, normality assumptions are violated, or a large number of variables is being modeled, p-values obtained in this fashion need not be accurate (Bentler & Bonett, 1980; Chou, Bentler, & Satorra, 1991; Curran, West, & Finch, 1996; Hoelter, 1983; Hu, Bentler, & Kano, 19”
Multiple-Group Analysis for Structural Equation Modeling With Dependent Samples (2015, methodology): “Methods to analyze all the variables collectively are needed, although most methods in SEM are unable to provide reliable analysis when the total number of variables is large or the sample size is small (Bentler & Yuan, 1999; Moshagen, 2012).”
Indicators per Factor in Confirmatory Factor Analysis: More is not Always Better (2020, background): “…recent studies have shown that minimum sample size requirements increase as number of indicators per factor increases due to the model size effect (Gagné & Hancock, 2006; Herzog, Boomsma, & Reinecke, 2007; Jackson, Voth, & Frey, 2013; Kenny & McCoach, 2003; Moshagen, 2012; Shi, Lee, & Terry, 2018).”
Revisiting the Model Size Effect in Structural Equation Modeling (2018, methodology): “For example, in Moshagen (2012), only two levels of q (q = 123 or q = 310) were examined when p = 60, whereas in our simulation, q ranged from 1 to 1,030 for the same value of p.”
Further specification of the leader political skill–leadership effectiveness relationships: Transformational and transactional leader behavior as mediators (2013, methodology): “…used these indicators in our analyses, since Political Skill and Leadership Effectiveness 27 research has shown that increases in the size of covariance matrices due to the increased number of variables lead to inflated goodness-of-fit statistics (Herzog, Boomsma, & Reinecke, 2007; Moshagen, 2012).”
A New Strategy for Testing Structural Equation Models (2016, background): “…when the H0 is true and noncentral χ2(df, λ) distributed when the H1 is true.3 Concerning normal theory estimation, several conditions might lead to violations of these assumptions, including multivariate nonnormality (e.g., Fouladi, 2000) and large covariance matrices (e.g., Moshagen, 2012).”
Enhancing Model Fit Evaluation in SEM: Practical Tips for Optimizing Chi-Square Tests (2023, background): “…theory, but a large number of variables 𝑝 and/or parameters 𝑞 , a small number of indicator loadings per factor, and small ratio of 𝑁 to 𝑑𝑓 also contribute to spurious goodness-of-fit model rejections (Arruda & Bentler, 2017; Moshagen, 2012; Yuan & Bentler, 1999; Zheng & Bentler, 2021, 2023).”
Tracing the Path from Personality — via Cooperativeness — to Conservation (2013, methodology): “Given that the w² model fit test statistics tends to be inflated for models involving many variables (Moshagen, 2012), model fit was evaluated by means of descriptive indices of fit (RMSEA and CFI).”
Examining Chi-Square Test Statistics Under Conditions of Large Model Size and Ordinal Data (2018, methodology): “Aiming to better understand the model size phenomenon, Moshagen (2012) conducted research to elucidate whether increasing number of indicator variables, number of free parameters, or model degrees of freedom resulted in an inflated chi-square statistic.”
Understanding the Model Size Effect on SEM Fit Indices (2018, methodology): “The model size effect on LR-based chi-square statistics has been investigated in many studies (Herzog et al., 2007; Moshagen, 2012; Shi, Lee, et al., 2015, 2017; Yuan et al., 2015).”
Personal Initiative and Job Performance Evaluations: Role of Political Skill in Opportunity Recognition and Capitalization (2017, methodology): “As recent research has shown that fit indices of structural equation models generally deteriorate with an increasing number of manifest variables (Moshagen, 2012), we built two indicator variables for each scale based on the odd- and even-numbered scale items to reduce the number of manifest variabl”