Related Experiment Videos
The evaluation and selection of adequate causal models: a compensatory education example
Evaluation and Program Planning
|December 12, 1981
Summary
This study presents methods for evaluating structural equation models, using goodness of fit statistics and coefficients of determination to assess model adequacy. It demonstrates how parameter estimates are sensitive to different model specifications.
Area of Science:
- Statistics
- Psychometrics
- Social Sciences
Background:
- Assessing the adequacy of latent variable structural relation models is crucial for valid statistical inference.
- Existing methods for model evaluation require careful consideration of estimation techniques and fit indices.
Purpose of the Study:
- To discuss procedures for determining the relative adequacy of latent variable structural relation models.
- To demonstrate the numerical sensitivity of parameter estimates under alternative model specifications.
Main Methods:
- Utilizing the chi-square goodness-of-fit statistic.
- Employing incremental fit indices for covariance structure models.
- Calculating latent variable coefficients of determination.
Main Results:
- Model adequacy can be assessed through various statistical measures, including chi-square fit, incremental fit indices, and coefficients of determination.
- Parameter estimates exhibit sensitivity to alternative model specifications, impacting interpretive validity.
Conclusions:
- The discussed procedures provide a framework for evaluating model adequacy in structural equation modeling.
- Understanding parameter sensitivity is essential for accurate interpretation of results in evaluation research.