Purpose: We compared the rates of specialized care for residents with Alzheimer's disease or dementia in special care units (SCUs) and other nursing home (NH) units and examined the associations of SCU residence with process of care and resident outcomes. Design and Methods: Data came from the 2004 National Nursing Home Survey. The indicators of process of care included physical restraints, continence management, feeding tubes, and influenza and pneumococcal vaccinations. Resident outcomes included pressure ulcers, hospitalization, emergency room visits, weight loss, and falls. Analyses were conducted by using Stata SE version 10. Results: Multivariate logistic regression analyses show that SCU residents were more likely to have received specialized dementia care and specialized behavioral problem management. They were less likely to have bed rails (adjusted odds ratio [AOR] = 0.39, AOR = 0.35, ps less than 0.01), use catheters (AOR = 0.33, AOR = 0.33, ps less than 0.01), and yet more likely to have toilet plans/bladder training for incontinence control (AOR = 1.90, AOR = 1.62, ps less than 0.01) than those in regular units and those in NHs without an SCU. Moreover, SCU residents were less likely to have pressure ulcers, hospitalization than those in regular units, and less likely to have experienced weight loss than those in NHs without an SCU. However, they were more likely to have falls (AOR = 1.32, AOR = 1.36, ps less than 0.05) than those in regular units and those in NHs without an SCU. Implications: Our study shows that SCU residents had, in general, better process of care than those in regular units and in NHs without an SCU. Further studies are needed to assess specific outcome changes among SCU residents and to evaluate the cost-effectiveness of having such units.