Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 1, 2001Insecta mundi

The Relative Performance of Full Information Maximum Likelihood Estimation for Missing Data in Structural Equation Models

View Full Paper
Ask AI
Bookmark
Share

Authors

CECraig K. EndersUniversity of California, Los AngelesDBDeborah L BandalosJames Madison University

Discussion

Loading...

Member takes

Implication

Simulation study shows superior estimation accuracy and lower convergence failures with full information maximum likelihood in structural equation models, highlighting its utility for missing data.

Key Points

  • Evaluate and compare the performance of full information maximum likelihood against three traditional missing data techniques across varying data conditions in structural equation models.
  • Conducted Monte Carlo simulations evaluating four missing data treatments: full information maximum likelihood (FIML), listwise deletion, pairwise deletion, and similar response pattern imputation.
  • Manipulated factor loading magnitude, sample size, and missing data rate under missing completely at random (MCAR) and missing at random (MAR) conditions.
  • Assessed outcomes using convergence failures, parameter estimate bias, parameter estimate efficiency, and model goodness-of-fit statistics.
  • FIML estimation demonstrated superior performance across all simulated conditions compared to deletion and imputation alternatives.
  • Under ignorable missing data conditions (MCAR and MAR), FIML produced unbiased and more efficient parameter estimates than competing methods.
  • FIML yielded the lowest proportion of model convergence failures and maintained near-optimal Type 1 error rates across both simulations.

Cite This Study

Enders et al. (2001) studied this question.

synapsesocial.com/papers/69d72c59a98988943d563cfehttps://doi.org/10.1207/s15328007sem0803_5
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Structural Equations with Latent Variables.1991 · 17,526 citations
  2. 2Pairwise deletion for missing data in structural equation models: Nonpositive definite matrices, parameter estimates, goodness of fit, and adjusted sample sizes1998 · 107 citations
  3. 3The Analysis of Incomplete Data1971 · 300 citations
  4. 4The Effect of Sampling Error on Convergence, Improper Solutions, and Goodness-of-Fit Indices for Maximum Likelihood Confirmatory Factor Analysis1984 · 1,852 citations