The Effect of Scale Referent on the Testing of Mean Structure Parameters

• In Event: Roundtable Session 101
  In Roundtable Session: Extensions and Issues in Growth Modeling

Abstract

Structured means analysis is a very useful approach for testing hypotheses about population means on latent constructs. In such models, a z-test is most commonly used for testing the statistical significance of the relevant parameter estimates or of the differences between parameter estimates, where the z value is computed based on the asymptotic standard error estimates associated with the parameter(s) of interest. In the current paper, a series of population analyses demonstrate that the z-test, or more accurately the standard error estimates upon which z-tests are based, are not invariant to how factors’ reference indicators are chosen. This problem is illustrated in the context of multisample latent means models, within-group latent means models, and second-order latent growth models.

Authors

• Yoonjeong Kang, University of Maryland - College Park

• Gregory R. Hancock, University of Maryland