Sample Size and Number of Parameter Estimates in Maximum Likelihood Confirmatory Factor Analysis: A Monte Carlo Investigation

By Dennis L. Jackson. 2001

289 citations recorded by the source corpus.

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

View Paper: https://www.semanticscholar.org/paper/b742abb8be45a111e11a752ec2fc5edde35c1796

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

In Sociology and Political Science

Citations in the corpus

Factor Analysis in Counseling Psychology Research, Training, and Practice (2006, background): “Whereas early thinking was that a fixed number of cases per estimated parameter was necessary (e.g., Bentler, 1990), recent research shows that the number of parameters does not appear to be as important as overall sample size (Jackson, 2001).”
A Multifactor Lesbian, Gay, and Bisexual Positive Identity Measure (LGB-PIM) (2014, unknown): “We selected a smaller sample size for the EFA than the CFA, given evidence that the higher sample size would most benefit the CFA in terms of minimizing bias and parameter estimate variance (Jackson, 2001; MacCallum, Widaman, Zhang, & Hong, 1999).”
The impact of anti-corruption strategies on corruption free performance in public construction projects (2012, methodology): “Based on the stated criteria and the suggestions made by various researchers (Bentler and Yuan, 1999; Jackson, 2003) four parameters are selected for validating the hypothesized relationship as shown in Table 3.”
My Influence on the World (of Others): Goal Efficacy Beliefs and Efficacy Affect Predict Private, Public, and Activist Pro‐environmental Behavior (2020, result): “However, our sample size was larger than 200 participants (Jackson, 2001) and we exceeded a 5:1 ratio in Study 2 (Tanaka, 1987).”
Workload Manageability Among Novice Special and General Educators: Relationships With Emotional Exhaustion and Career Intentions (2017, methodology): “For both measurement models and path models, at least five observations are required per parameter estimated (Jackson, 2001, 2003); all analyses meets these criteria.”
A Note on Sample Size and Solution Propriety for Confirmatory Factor Analytic Models (2013, background): “An early conditional rule expressed sample size (n) requirements in terms of the number of parameters estimated (q), denoted n/q (see Jackson, 2001, 2003, for more discussion on this rule).”
Legitimacy, Visibility, and the Antecedents of Corporate Social Performance: An Investigation of the Instrumental Perspective (2011, methodology): “Please note that although it might have been preferable to complete a confirmatory factor analysis (CFA) on our CSP indicators, we did not have a sufficient sample size (Jackson, 2005) to generate a confirmatory structure in which we could have confidence.”
Small-Sample Robust Estimators of Noncentrality-Based and Incremental Model Fit (2009, unknown): “As mentioned earlier, however, although rules of thumb, like N:t ratios, might provide sensible guidelines for applied researchers in many situations, they should always be employed with abundant vigilance and leniency (e.g., Brown, 2006; Jackson, 2001, 2003, 2007).”
Sample Size Requirements for Structural Equation Models: An Evaluation of Power, Bias, and Solution Propriety (2013, background): “These characteristics of a model have also been shown to influence other important aspects of model output that are not under consideration in this study, such as the performance and stability of fit indices (see Jackson, 2001; Kim, 2005; Velicer & Fava, 1998).”
LATENT GROWTH MIXTURE MODELING: A SIMULATION STUDY (2007, background): “This efficiency, as the requirement of observation per estimated parameters, could be very small (Jackson, 2001).”
Applied Psychometrics: Sample Size and Sample Power Considerations in Factor Analysis (EFA, CFA) and SEM in General (2018, unknown): “ML with multivariate normal data (continuous) 1) 100 2) 200-400 3) N:q = 5:1 4) N:q = 10:1 5) 30-460 1) Anderson & Gerbing (1984) 2) Jackson (2001) 3) Tanaka (1987) 4) Bentler & Chou (1987) 5) Wolf et al. (2013)”
Revisiting Sample Size and Number of Parameter Estimates: Some Support for the N:q Hypothesis (2003, methodology): “Though this seems like a reasonable assertion, and limitations of the previous work (Jackson, 2001) not withstanding, the assertion does not appear to have empirical support.”
A Brief Guide to Structural Equation Modeling (2006, background): “For example, Jackson (2001, 2003) found only a small effect of sample size on model fit when he tested the hypothesis that an inadequate sample would result in poor-fitting models.”
Sport Tourists’ Involvement with a Destination (2013, background): “However, the RMSEA is sensitive to the number of estimated parameters in the model as well as to the sample size (Byrne, 2001; Jackson, 2001).”
The Model Size Effect in SEM: Inflated Goodness-of-Fit Statistics Are Due to the Size of the Covariance Matrix (2012, unknown): “In fact, Jackson (2001) also used a less strong variation of the number of free parameters and, in turn, did not find an effect of this manipulation on T ML .”
In Search of Underlying Dimensions: The Use (and Abuse) of Factor Analysis in Personality and Social Psychology Bulletin (2002, background): “…Bentler (personal communication, August 5, 2001), one must be cautious in drawing inferences about the effects of variations in number of parameters on model fit from the results of Jackson (2001) because he did not investigate models that varied widely in the number of parameters being estimated.”
A Comparison of Latent Growth Models for Constructs Measured by Multiple Items (2007, result): “These results agree with the finding of Jackson (2001) that for a correctly specified measurement model with 20 indicators, the”