Evaluating Small Sample Approaches for Model Test Statistics in Structural Equation Modeling

By J. Nevitt, G. R. Hancock. 2004

232 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/ec938f4d2baeaa707237a43e2bf95ab6be66cd01

How researchers see this paper

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

In Computer Networks and Communications

In Management Science and Operations Research

Citations in the corpus

Small Sample Statistics for Incomplete Nonnormal Data: Extensions of Complete Data Formulae and a Monte Carlo Comparison (2010, methodology): “The second study by Nevitt and Hancock (2004) had the goal of studying larger models than those considered by Fouladi (2000).”
Are Robust Standard Errors the Best Approach for Interval Estimation With Nonnormal Data in Structural Equation Modeling? (2018, background): “That scaled chi-square tests for model fit sometimes perform poorly in small samples (Bentler & Yuan, 1999; Nevitt & Hancock, 2004) and do not meet the assumptions for all types of nonnormal data (Yuan & Bentler, 1998, 1999) could be taken as an indication that robust LCIs have important…”
Indicators per Factor in Confirmatory Factor Analysis: More is not Always Better (2020, methodology): “05 2.575(.05 ∗ .95/1000)^½ = [0.0323, 0.0677]; Nevitt & Hancock, 2004).”
Small-Sample Robust Estimators of Noncentrality-Based and Incremental Model Fit (2009, unknown): “However, T MLb more closely follows a chi–square distribution when N is small (Fouladi, 2000; Herzog et al., 2007; Nevitt & Hancock, 2004).”
The Structure of Identity Consolidation: Multiple Correlated Constructs or One Superordinate Construct? (2007, background): “It should be noted that splitting the current sample in half resulted in two fairly small half-samples, and that small sample sizes may cause instability in model parameters ( Bentler & Yuan, 1999 ; Nevitt & Hancock, 2004 ) and may increase the likelihood that the null hypothesis of invariance will”
Structural Equation Modeling in Marketing: Some Practical Reminders (2008, background): “Long-standing issues in SEM, such as appropriate sample sizes (Lee and Song 2004b; Nevitt and Hancock 2004) and missing data (Allison 2003; Lee and Song 2004a; Nevitt and Hancock 2004) are being addressed.”
Relative Performance of Categorical Diagonally Weighted Least Squares and Robust Maximum Likelihood Estimation (2014, background): “The mean- and mean- and variance-adjusted estimators for ML have been investigated in several studies (Finch, West, & MacKinnon, 1997; Fouladi, 2000; Green et al., 1997; Herzog, Boomsma, & Reinecke, 2007; Lei, 2009; Nevitt & Hancock, 2004; Rhemtulla, Brosseau-Liard, & Savalei, 2012).”
The Construct of Creativity: Structural Model for Self‐Reported Creativity Ratings (2009, methodology): “…(Byrne, 2001) that Mardia’s coefficient (1970) is significant, indicating marked multivariate kurtosis, we decided to implement Bollen-Stine (1992) bootstrapping based on the recommendations of Nevitt and Hancock (2004) Each of the two development models was refined until it had sufficient fit.”
The Model-Size Effect on Traditional and Modified Tests of Covariance Structures (2007, methodology): “In addition, the studies by Fouladi (2000) and Nevitt and Hancock (2004) provided empirical evidence that, relative to T ML , these corrections might also improve small-sample performance even when the normality assumption is not violated at all.”
Collective Dynamics of Citizenship Behaviour: What Group Characteristics Promote Group-Level Helping? (2009, unknown): “Adapting items from Oldham and Cummings (1996), unit managers’ supportive leadership was measured by five items (a = 0.”
The Role of Personality and Career Decision-Making Self-Efficacy in the Career Choice Commitment of College Students (2006, methodology): “Following the guidelines of Bentler (1989), a sample size-to-estimated parameter ratio of 5:1 is recommended, and ratios of 10:1 or better have been shown to be even better, often with sample sizes of fewer than 100 (Marsh, Balla, & Hau, 1996; Nevitt & Hancock, 2004).”
Robust Two-Stage Approach Outperforms Robust Full Information Maximum Likelihood With Incomplete Nonnormal Data (2014, methodology): “Adjustments to the residual-based statistic that better control Type I error rates are available (Bentler & Yuan, 1999; Yuan & Bentler, 1998), but have been shown to have relatively low power (Nevitt & Hancock, 2004; Savalei, 2010b).”
Multiple-Group Analysis for Structural Equation Modeling With Dependent Samples (2015, background): “However, simulation results in Nevitt and Hancock (2004) and Savalei (2010) suggest that TAML tend to have smaller Type I errors than nominal levels.”
The Relationship Between Root Mean Square Error of Approximation and Model Misspecification in Confirmatory Factor Analysis Models (2012, background): “Some simulation studies report the RMSEA for the created misspecifications to communicate their severity (Nevitt & Hancock, 2004; Savalei, 2010).”
Small Sample Size Solutions (2020, methodology): “Many authors have documented that when the sample size is small, the chi-square test leads to inflated Type I errors even under ideal circumstances (i.e., correctly specified model, normal data; see Nevitt & Hancock, 2004 and references therein).”
A Note on Sample Size and Solution Propriety for Confirmatory Factor Analytic Models (2013, methodology): “…are presented for each of the combinations of f, p/f, and loading size that would be expected to yield either ML or MLS ¦ 2 values that fall within the 99% confidence interval of ’ D :05, where the confidence interval was constructed based on Nevitt and Hancock (2004): CI is Œ:05 ˙ 2:575.”