A Bayesian method for analyzing combinations of continuous, ordinal, and nominal categorical data with missing values

Factors Promoting Posttraumatic Growth in Sexual Minority Adults Following Adolescent Bullying Experiences

The minority stress model posits that stigmatized identities expose sexual minority individuals to chronic stressors that contribute to health disparities, but that individual-level resources may mitigate psychological distress. Sexual minority adolescents experience one such stressor, bullying victimization, at higher rates than heterosexual peers. Whereas negative consequences of sexual identity-based bullying are well documented, potential positive outcomes are not well understood. The present work examined hypothesized pathways to posttraumatic growth (PTG)-positive psychological changes stemming from trauma-in sexual minority adults following adolescent bullying experiences. We predicted that attributing bullying to one's sexual identity, as opposed to other factors (e.g., weight/appearance, personality), would exacerbate perceived bullying severity but, in turn, enhance PTG. We also predicted that outness about sexual identity would enhance social support and, in turn, facilitate PTG. The hypothesized conceptual model was tested in two samples of sexual minority adults who had experienced bullying during adolescence (Sample 1: Community Sample [N = 139]; Sample 2: National Online Sample [N = 298]), using structural equation modeling with Bayesian estimation. Mediation hypotheses were tested using the PROCESS v3.4 macro. Participants reported their adolescent experiences with bullying, attributions for bullying, outness, social support, and PTG as a result of adolescent bullying experiences, in addition to demographics. Supporting the hypothesized model, in both samples, attributions to sexual identity-based bullying directly and indirectly (via bullying severity) predicted greater PTG, and outness predicted greater PTG through proximal impact on social support. This research underscores the importance of supportive responses to individuals who disclose sexual minority identities and of (re)framing attributions about bullying to facilitate growth.

Bayesian Variable Selection for Gaussian Copula Regression Models

We develop a novel Bayesian method to select important predictors in regression models with multiple responses of diverse types. A sparse Gaussian copula regression model is used to account for the multivariate dependencies between any combination of discrete and/or continuous responses and their association with a set of predictors. We utilize the parameter expansion for data augmentation strategy to construct a Markov chain Monte Carlo algorithm for the estimation of the parameters and the latent variables of the model. Based on a centered parametrization of the Gaussian latent variables, we design a fixed-dimensional proposal distribution to update jointly the latent binary vectors of important predictors and the corresponding non-zero regression coefficients. For Gaussian responses and for outcomes that can be modeled as a dependent version of a Gaussian response, this proposal leads to a Metropolis-Hastings step that allows an efficient exploration of the predictors' model space. The proposed strategy is tested on simulated data and applied to real data sets in which the responses consist of low-intensity counts, binary, ordinal and continuous variables.