Cited by 8,975Psychometric Methodologies and TestingStatistical Methods and Bayesian InferenceStatistical Distribution Estimation and Applications

Evaluating the Fit of Structural Equation Models: Tests of Significance and Descriptive Goodness-of-Fit Measures

Open access
Schermelleh-Engel, Karin, Moosbrugger, Helfried, Müller, Hans·2003·Department of Psychology - University of Koblenz-Landau

Abstract

For structural equation models, a huge variety of fit indices has been developed. These indices, however, can point to conflicting conclusions about the extent to which a model actually matches the observed data. The present article provides some guidelines that should help applied researchers to evaluate the adequacy of a given structural equation model. First, as goodness-of-fit measures depend on the method used for parameter estimation, maximum likelihood (ML) and weighted least squares (WLS) methods are introduced in the context of structural equation modeling. Then, the most common goodness-of-fit indices are discussed and some recommendations for practitioners given. Finally, we generated an artificial data set according to a "true" model and analyzed two misspecified and two correctly specified models as examples of poor model fit, adequate fit, and good fit.

Integrity signals

Registry

DataCite · 10.23668/psycharchives.12784

Cites

71 of 71 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Additional fit indices GFI = 0.92, AGFI = 0.89, and RMS = 0.06 indicated an acceptable fit of the data to the MIS-PL model.

    “Furthermore, additional fit indices such as GFI = 0.92, AGFI = 0.89 and RMS = 0.06 also indicated an acceptable fit of the data to the model (Schermelleh-Engel et al., 2003).”

    From Psychometric Evaluation of the Polish Version of the Mindfulness Inventory for Sport (MIS-PL) (2024) · cites this paper as Schermelleh-Engel et al., 2003

Cites

Flagged first, then by citations

2 of the 73 references couldn’t be matched to a known paper.

Cited by

Sorted by citations

Authors
Schermelleh-Engel, K., Moosbrugger, H., & Müller, H.
Affiliations
  • Goethe University Frankfurt (DE)
  • University of Erfurt (DE)
Venue
Department of Psychology - University of Koblenz-Landau
Publisher
Department of Psychology - University of Koblenz-Landau
Published
2003
Volume
8
Pages
23-74
OpenAlex
W1627191216
Cited by
8,975
References
73
Citations by year
2026: 128 · 2025: 749 · 2024: 891 · 2023: 928 · 2022: 910 · 2021: 879 · 2020: 837 · 2019: 538 · 2018: 489 · 2017: 420 · 2016: 392 · 2015: 395 · 2014: 348 · 2013: 284 · 2012: 212
FWCI
19.63
Type
journal article
Language
en
Open access
Locations
Psychology Archives
Topics
  • Social Sciences › Decision Sciences › Management Science and Operations Research › Psychometric Methodologies and Testing
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Distribution Estimation and Applications
Keywords
structural equation modeling, model fit assessment, goodness-of-fit indices, significance tests, maximum likelihood estimation, weighted least squares estimation, model misspecification

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review