The Bean and Metzner Nontraditional Undergraduate Student Attrition Model appeared to be a good model to address community college student attrition because it takes into account the environmental variables that colleges must consider to enhance persistence. A study was conducted to determine how well the Bean and Metzner Model fit with community college student data and to make theoretically consistent modifications to the model if the fit were weak. Survey data collected from 597 respondents enrolled at an urban community college were analyzed. Variables included age, gender, ethnicity/race, enrollment status, educational goals, high school rank, academic variables (e.g., study habits and absenteeism), environmental variables (e.g., finances, employment, and family responsibilities), grade point average, psychological outcomes (e.g., satisfaction and stress), intent to leave (e.g., plan to re-enroll and transfer difficulty), and dropout. The method used to assess the Bean and Metzner model with these data was structural equation modeling, and statistical procedures were carried out using LISREL and LISCOMP. When the model was found to be a weak fit with community college data, exploratory factor analysis was employed and resulted in identifying a new model, the Community College Retention Model, which was tested with the community college data and further tested on two student subgroups. The Community College Retention Model proved to be a plausible model. (Author/AC)