Kontakt
Kako nakupovati
Dostava
Svetovalec pri nakupu
Testing Latent Variable Interaction Effect
dealing with data nonnormality and model misspecification
Angleščina
142 b
30 dni za vračilo blaga
The book discusses the effects of data nonnormality, §model misspecification, sample size, and effect size §on testing latent variable interactions through an §inspection of the Jöreskog and Yang's (1996) model. §Mattson's (1997) method was used to generate §nonnormal latent variables in this Monte Carlo §study. One covariance parameter was deleted for §investigating the influence of misspecified models. §The simulation involved a balanced experimental §design, with 3 × 2 × 3 × 3 = 54 combinations. Data §analysis focused on bias of estimating parameters, §standard errors, model fit indexes. Variance §partition was conducted to further examine the §unique and combined influence of the factors (i.e., §data nonnormality, model specification, sample size, §effect size). Results indicated that data §nonnormality and model misspecification had large §effects on fit indexes (e.g., SRMR, RMSEA). Also, §severe nonnormality led to a large bias of §estimating the interaction effect. Implications of §and recommendations for testing latent variable §interactions are discussed.
O knjigi
Angleščina
