Three Approaches to Using Lengthy Ordinal Scales in Structural Equation Models: Parceling, Latent Scoring, and Shortening Scales

摘要: Lengthy scales or testlets pose certain challenges for structural equation modeling (SEM) if all the items are included as indicators of a latent construct. Three general approaches (parceling, scoring, and shortening) to lengthy in SEM were reviewed evaluated. A hypothetical population model was simulated containing two exogenous constructs with 14 each an endogenous construct four indicators. The simulation generated data sets varying numbers response options, types distributions, factor loadings ranging from low high, sample sizes small moderate. varied incorporate one following: (1) single parcels, (2) various parcels constructs, (3) scores observed variables, (4) six individual constructs. dependent variables evaluated biases covariance partial parameters. Biases these parameters found be minimal under following conditions: when five options used constructs; at above 100 that relatively less skewed case dichotomous indicators; high diverse These findings provided guidelines resolving inconsistency applying empirical data.