Bayesian Estimation

I am interested in Bayesian approaches to overcome current estimation challenges.

For example:

• For mixture models (e.g. growth mixture models, latent class analysis), a fully Bayesian estimation method allows researchers to examine the distribution of the number of classes.

• Sparse item endorsement arises in psychological data for a number of reasons and can lead to estimation difficulties. I have developed a set of recommended, general priors (which I call “moderately informative priors”) for models with categorical indicators which can improve convergence and reliability of estimates and increase statistical power.

• I am currently working on applications of Bayesian variable selection methods, in particular stochastic search variable selection (SSVS) for research questions in psychology.

Latent Variable Modeling

I’m also interested in leveraging structural equation models to understand processes in clinical and developmental psychology.

For example:

• The Latent Curve Model with Structured Residuals (LCM-SR) was designed to disentangle and disattenuate between- and within-person effects in longitudinal SEMs. I have also done comparative work to understand how different multivariate longitudinal models of change capture inferences about within-person change.

• I am interested in the integrative data analysis (IDA) framework, which can be leveraged to better understand normative and non-normative developmental processes. IDA can be used to pool raw data from multiple studies to examine how the measurement of key constructs changes over time, enable researchers to study underrepresented groups, rare behaviors, and longer developmental periods.​