This study examined the predictors of total nurse and registered nurse (RN) staffing hours per resident day separately in all free-standing California nursing homes (1,555), using staffing data from state cost reports in 1999. This study used a two-stage least squares model, taking into account nursing turnover rates, resident case mix levels, and other factors. As expected, total nurse and RN staffing hours were negatively associated with nurse staff turnover rates and positively associated with resident case mix. Facilities were resource dependent in that a high proportion of Medicare residents predicted higher staffing hours, and a higher proportion of Medicaid residents predicted lower staffing hours and higher turnover rates. Nursing assistant wages were positively associated with total nurse staffing hours. For-profit facilities and high-occupancy rate facilities had lower total nurse and RN staffing hours. Medicaid reimbursement rates and multifacility organizations were positively associated with RN staffing hours.