Testing misspecifications in structural equation modeling

Model misspecifications include overparameterization and underparameterization. In this sense, modification indices provide useful information about additional specifications that could be done to improve model fit. However, it should be noted that re-specifications in the model may be not relevant....

Full description

Bibliographic Details
Main Authors: Dominguez-Lara, Sergio, Merino-Soto, César
Format: Online
Language:spa
Published: Universidad Nacional de Córdoba 2018
Subjects:
Online Access:https://revistas.unc.edu.ar/index.php/racc/article/view/19595
Description
Summary:Model misspecifications include overparameterization and underparameterization. In this sense, modification indices provide useful information about additional specifications that could be done to improve model fit. However, it should be noted that re-specifications in the model may be not relevant. Saris, Satorra and Van der Veld (SSV) proposed a method that provides information for judging the practical significance of model re-specification suggested by the modification indices, representing additional and different information than fit indices. Currently, there are available programs for testing misspecifications.  However, they require the output linked to a specific SEM program. The present study provides a SSV syntax for SPSS, which can expand the computational environment for their application.