|
Abstract:
|
The analysis of survey data includes estimating variance that provides insights into the reliability of the estimate. Publications from the National Center for Health Statistics (NCHS) commonly present estimated proportions of disease prevalence and other health-related topics that are subject to the NCHS Data Presentation Standards for Proportions. These standards contain multiple criteria to identify cases that are subject to statistical review or suppression. For estimates from complex sample surveys the criteria include the sample size, effective sample size, and the absolute and relative widths of Clopper-Pearson confidence intervals as adapted by Korn and Graubard (KG CI). In some circumstances, such as age-standardized estimates and estimates for subgroups not represented in all primary sampling units, the specific survey design may affect the calculation of the KG CI. This presentation describes the criteria needed to meet the presentation standards of reliability, a macro created to automate the calculation of KG CIs, and the application of the macro for presenting results. The macro may assist analysts of complex survey data when assessing the reliability of estimates.
|
