Abstract
Bayesian estimation of multilevel structural equation models (MLSEMs) offers advantages in terms of sample size requirements and computational feasibility, but does require careful specification of the prior distribution especially for the random effects variance parameters. The traditional “non-informative” conjugate choice of an inverse Gamma prior with small hyperparameters has been shown time and again to be problematic. In this paper, we investigate alternative, more robust prior distributions. In contrast to multilevel models without latent variables, MLSEMs have multiple random effects variance parameters, both for the multilevel structure and for the latent variable structure. It is therefore even more important to construct reasonable priors for these parameters. We find that, although the robust priors outperform the traditional inverse-Gamma prior, their hyperparameters do require careful consideration.
Original language | English |
---|---|
Pages (from-to) | 875-893 |
Number of pages | 19 |
Journal | Structural Equation Modelling: A Multidisciplinary Journal |
Volume | 28 |
Issue number | 6 |
Early online date | 21 Jun 2021 |
DOIs | |
Publication status | Published - 2021 |
Bibliographical note
Funding Information:This research was supported by a Research Talent Grant [406-15-264] from the Netherlands Organisation for Scientific Research. We would like to thank Irene Klugkist, Duco Veen, and Mari?lle Zondervan-Zwijnenburg for helpful comments on an earlier version of this manuscript.
Publisher Copyright:
© 2021 The Author(s). Published with license by Taylor & Francis Group, LLC.
Research Groups and Themes
- SoE Centre for Multilevel Modelling
Keywords
- Bayesian
- robust priors
- multilevel models