search for




 

Data presentation standards for proportions based on confidence intervals from complex sample surveys
Journal of the Korean Data & Information Science Society 2025;36:127-37
Published online January 31, 2025;  https://doi.org/10.7465/jkdi.2025.36.1.127
© 2025 Korean Data and Information Science Society.

Inho Park1
Correspondence to: This work was supported by the Pukyong National University Research Fund in 2021.
1 Professor, Major of Statistics and Data Science, Pukyong National University, Busan 48513, Korea. E-mail: ipark@pknu.ac.kr
Received December 30, 2024; Revised January 15, 2025; Accepted January 15, 2025.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
The assessment of the precision of a proportion estimate and whether it should be published is primarily based on its relative standard error. However, by definition, for very small or large estimates, their relative standard error may be too large or small, thus being an in-appropriate criterion for the reliabilithy measure. The U.S. National Center for Health Statistics recently developed new criteria based on the width and relative width of confidence intervals and recommended using the Clopper-Pearson confidence interval, which reflects the elements of a complex sampling design, instead of the standard Wald confidence interval. In this study, we carried out a simulation study to compare the widths and relative widths of confidence intervals based on several confidence interval methods against well-accepted RSE based reference values. We found that the Clopper-Pearson method covers an appropriate range of proportion values, centered around its value for which has traditional 30% relative standard error, for proportion values that are not too extremely small or too unstable. The Clopper-Pearson confidence interval method was found to have a conservative coverage probability, as well as less extreme and more or less conservative confidence interval widths and relative widths compared to other methods.
Keywords : Complex sampling design, proportion estimation, confidence interval, data presentation standard