Objective To describe the Green House (GH) model of nursing home (NH) care, and examine how GH homes vary from the model, one another, and their founding (or legacy) NH. Data Sources/Study Setting Data include primary quantitative and qualitative data and secondary quantitative data, derived from 12 GH/legacy NH organizations February 2012—September 2014. Study Design This mixed methods, cross‐sectional study used structured interviews to obtain information about presence of, and variation in, GH‐relevant structures and processes of care. Qualitative questions explored reasons for variation in model implementation. Data Collection/Extraction Methods Interview data were analyzed using related‐sample tests, and qualitative data were iteratively analyzed using a directed content approach. Principal Findings GH homes showed substantial variation in practices to support resident choice and decision making; neither GH nor legacy homes provided complete choice, and all GH homes excluded residents from some key decisions. GH homes were most consistent with the model and one another in elements to create a real home, such as private rooms and baths and open kitchens, and in staff‐related elements, such as self‐managed work teams and consistent, universal workers. Conclusions Although variation in model implementation complicates evaluation, if expansion is to continue, it is essential to examine GH elements and their outcomes.