Date of Degree


Document Type


Degree Name

Doctor of Philosophy (PhD)




Ashley S. Love


Chunling Niu


Lynn T. Downs


Introduction: Retaining employees is crucial to the success of organizations. While significant effort has been made to model employee retention, models have been unable to significantly predict retention. The purpose of this quantitative, non-experimental, cross-sectional study was to develop an improved employee retention model which could be used as a practical guide to improve employee retention. Materials and Methods: The study analyzed a survey created, implemented, and collected by the Statistical Cell at the U.S. Army Center of Excellence. In September of 2020, all 864 active-duty PAs were sent a link to the survey and 290 completed the survey, for a participation rate of 33.6%. The study used a combination of population-based regression analysis and subgroup-based latent class analysis. To guide the study a conceptual model combined a population-based retention model (Causal Model of Turnover) with a subgroup-based person-centered model (Proximal Withdrawal State Theory). IRB approval was obtained thru the U.S. Army Medical Center of Excellence and University of the Incarnate Word. Results: Multivariate logistical regression found belief career goals were achievable in Army medicine (adjusted odds ratio 0.74, 95% CI 0.57, 0.98), age (adjusted odds ratio 0.63, 95% CI 0.41, 0.98), and perception DHA control of Army medicine affects retention (adjusted odds ratio 0.54, 95% CI 0.39, 0.75) as factors able to predict the intent to serve 20 years in the Army. Latent Class Analysis (LCA) divided the population into three groups, the sensitive stayers, moderates, and indifferent leavers. The sensitive stayers had the highest retention (0.93) and were most likely to be concerned about variables affecting retention. The indifferent leavers had the lowest retention (0.68) and the least concern for variables affecting retention. The moderates fell in between the other two groups. Belief career goals were achievable in Army Medicine was the only variable able to predict class memberships in the three classes of LCA (p < .001). A revised conceptual model was improved by adding perception career goals were achievable in the organization to better unify the model. Conclusions: The extent goals are perceived to be achievable in the current job create job satisfaction and organizational commitment which strengthens intent to stay. Career goals vary based on the individual and accounting for these individual preferences provides a bridge to cohesively unite population-based retention models with person centered retention models. Strengths of this study include the incorporation of two leading retention theories into one cohesive model improved by the study using results triangulated with multiple statistical techniques and past criticisms of retention theories. Weaknesses include the use of secondary cross-sectional data not designed to support the conceptual framework and the population was limited to only U.S. Army PAs. Future studies should utilize experimental longitudinal studies to consider how the perception of career goals being achievable in an organization can be improved to increase retention.