BackHome
Refresh: OffView Personal Schedule
Session Summary

Composite/Factor Scores Analyses With Bias Correction: Good Alternatives to Full Structural Equation Modeling?

Fri, April 9, 4:10 to 5:40pm EDT (4:10 to 5:40pm EDT), SIG Sessions, SIG-Structural Equation Modeling Paper and Symposium Sessions

Session Type: Structured Poster Session

Abstract

While structural equation modeling has been a preferred approach to simultaneously model relations among latent variables and between latent variables and indicators, challenges in computations and specifications arise as researchers work with increasingly complex models. Alternative, two-step methods using composite/factor scores with bias corrections for measurement errors have gained attention as research has shown their numerical stability in small samples and flexibility in model specifications. In this session, presenters will share new development, applications, and findings on composite/factor score methods handling multilevel data, latent interactions, and ordinal indicators. These methods include composite variables with reliability adjustment, factor scores with Croon's method or with definition variables, and a fully Bayesian factor score regression approach.

Sub Unit

Chair

Papers

Discussant