Evaluation of Confidence Interval Estimation Methods for the National Compensation Survey

Glenn Springer, Alan H. Dorfman, Steven P. Paben, and Martha Walker


The National Compensation Survey (NCS) is a Bureau of Labor Statistics (BLS) program which provides data on occupational wages. A theoretical investigation presented by Casady, Dorfman, and Wang (1994 , 1996) suggested that 95% confidence intervals for estimates, when based on the standard normal distribution and standard methods of variance estimation, tend to yield less than the actual 95% coverage. They presented nonstandard methods of constructing confidence intervals, which give more accurate coverage. These intervals tend to be longer than the standard intervals and depend mainly on the use of a t-statistic having degrees of freedom dependent on the available domain data. We modified this method to make it suitable to the multi-stage design of the NCS. Using NCS data, an artificial sampling frame was created and simulated samples were selected. The standard normal confidence intervals were compared to confidence intervals using the t-distribution with weighted degrees of freedom for estimates of means, totals and quantiles. Coverage properties for confidence intervals using the non-standard approach were found to be superior to the standard normal approach.