Abstract:
|
This paper provides a two-step approach to estimate the standard errors of parameter estimates in a trend analysis. This analysis used single-year American Community Survey (ACS) Public Use Microdata Sample (PUMS) data from years 2000 to 2015 to examine the trend of the race/ethnicity diversity of Registered Nurses (RNs). In estimating the standard errors of parameter estimates in logistic regression, the existing methods provided by U.S. Census Bureau doesn't apply. A two-step method was developed to estimate variance for the data across 16 years. This method estimated adjustment factors based on data years with replicate weights available (2005-2015), and then apply the adjustment factor to the estimates based on the data across all years (2000-2015). Using the idea of the design factor approach, this two-step approach takes advantage of all the available data.
|