Sawyer-Morris, Ginnie2023-01-122023-01-122022https://udspace.udel.edu/handle/19716/32050Recovery housing is a promising community-based treatment modality for the 21.2 million individuals living with substance use disorders (SUDs) in the United States. However, women and men face unique barriers in their recovery, and little is known about whether and how such barriers persist over time in recovery housing contexts. The current study sought to address this gap by identifying key, gender-specific predictors of women’s and men’s recovery status (i.e., stable versus unstable recovery) using a latent growth modeling and machine learning approach. Through secondary analysis of a community-based sample of Delaware sober living home residents, multiple-group latent growth modeling was used to capture gender-specific trajectories of women’s and men’s recovery capital. These trajectories were then used in a series of gender-specific random forest predictions to identify variables strongly associated with women’s and men’s recovery status. Findings suggest that while social support was the strongest predictor of both women’s and men’s recovery status, women presented with more trauma and co-occurring mental health disorders, made less money, and reported greater financial strain, stress, and depressive symptomatology compared to men. Given the gender-specific barriers women face in recovery, sober living homes represent an ideal context for the implementation of gender-responsive programming.Gender differencesGender-responsive programmingRandom forestRecovery capitalSubstance use disorderThesis1358406761https://doi.org/10.58088/s9ea-xb912022-08-10en