Education and Training
- PhD Quantitative Psychology, University of Southern California
- MA Social Psychology, University of Southern California
- BFA Jazz Piano Performance, The New School University
Tim Hayes is an Assistant Professor of Quantitative Psychology at Florida International University. His quantitative research primarily focuses on using statistical simulations to evaluate novel methods for addressing missing data in longitudinal research. Missing data are a ubiquitous problem in a variety of applied research settings. While researchers often have clear a priori hypotheses concerning key variables in their substantive models, they rarely have well-developed a priori hypotheses concerning the factors that might lead to missing data. As a result, many existing methods of assessing the relationship between predictor variables and missing data are used by researchers in an inherently exploratory manner.
A family of recent missing data methods takes advantage of this reality by using exploratory data mining techniques based on Classification and Regression Tree (CART) analysis to address missing data. Whereas many popular missing data techniques assume multivariate normality and linearity, CART and its extensions model complex nonlinearities and interactions among predictors without making the same parametric assumptions. One approach addresses missing data by using CART and random forests analyses for binary outcomes to model the probability of dropout and create inverse probability weights (Hayes & McArdle, 2017; McArdle, 2013) A second approach addresses missing data by using CART and random forest analyses for continuous and categorical outcomes to generate multiple imputations (Doove, Van Buuren, & Dusseldorp, 2014). Tim’s main program of missing data research has focused on evaluating and comparing these CART-based missing data methods, with a particular focus on their performance in small sample research settings, such as randomized longitudinal clinical trials.
Hayes, T. & McArdle, J. J. (2017). Should we impute or should we weight? Examining the performance of two CART-based techniques for addressing missing data in small sample research with nonnormal variables. Computational Statistics and Data Analysis, 115, 35-52. https:/doi.org10.1016/j.csda.2017.05.006
Wilcox, R. R. & Hayes, T. (2016). Within groups ANOVA when using a robust multivariate measure of location. Journal of Modern and Applied Statistical Methods, 15 (2), 41-52. Doi: 10.22237/jmasm/1478001840.
Hayes, T., Usami, S., Jacobucci, R, & McArdle, J. J. (2015). Using Classification and Regression Trees (CART) and Random Forests to Analyze Attrition: Results from Two Simulations. Psychology and Aging 30(4), 911-929. http:/dx.doi.org10.1037/pag0000046.
Usami, S., Hayes, T., & McArdle, J. J. (2015a). On the mathematical relationship between latent change score and autoregressive cross-lagged factor approaches: Cautions for assessing causal precedence between variables. Multivariate Behavioral Research,50(6), 676-687. http:/dx.doi.org10.1080/00273171.2015.1079696.
Usami, S., Hayes, T., & McArdle, J. J. (2015b). Inferring Longitudinal Relationships Between Variables: Model Selection Between the Latent Change Score and Autoregressive Cross-Lagged Factor Models. Structural Equation Modeling 23(3), 331-342. http:/dx.doi.org10.1080/10705511.2015.1066680.