The Single Indicator Approach when Measures do not follow a Common Factor Model, and when Tau-Equivalence is Violated

• In Event: Division D Graduate Student in Progress Research Gala

Abstract

When latent variable models cannot be used the single indicator (SI) approach is an alternative method to account for measurement error when only one measured variable is available. Literature shows that the SI approach outperforms the latent variable SEM in small sample sizes, although this work has primarily been performed assuming a common factor model with equal factor loadings. The current proposal explores conditions when SI approach may be superior to latent variable techniques when data do not assume a common factor model with equal factor loadings. A simulation study is proposed to examine the how the SI approach compares to full LV-SEM techniques. Anticipated results and implications are discussed.

Author

• Daniel Peche Gonzalez, The University of Texas at Austin