Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 54 - Record Linkage, Data Integration, and Improving Survey Measurement
Type: Contributed
Date/Time: Sunday, August 8, 2021 : 3:30 PM to 5:20 PM
Sponsor: Survey Research Methods Section
Abstract #317852
Title: Relative Standard Error: A Misleading Indicator for Cell Suppression and Publishing Guidelines for Estimates of Proportions
Author(s): Sadeq R Chowdhury* and David K Kashihara
Companies: Agency for Healthcare Research and Quality and Agency for Healthcare Research and Quality
Keywords: RSE; Cell suppression; Publishing guidelines; MEPS
Abstract:

In publishing survey estimates, the relative standard error (RSE) of an estimate along with sample size is often used in developing cell suppression and publishing guidelines. Since the RSE is derived adjusting for the size of the estimate, it is useful to compare the relative precision of estimates with different sizes and hence convenient to use the same level of RSE as the suppressing/publishing threshold for all estimates. While it works well for estimates of continuous variables, it appears misleading for estimates of proportions. For small proportions close to zero, the RSE becomes very large even when the standard error (SE) or confidence interval (CI) is acceptably small. On the other hand, for large proportions the RSE becomes small even when the SE or CI is unacceptably large. Consequently, many small proportions with acceptable SEs/CIs get suppressed while many large proportions with unacceptable SEs/CIs get released. This poster will illustrate this misleading behavior of RSE and discuss how this problem was addressed in the context of revising the publishing guidelines for Medical Expenditure Panel Survey (MEPS) Insurance Component (IC).


Authors who are presenting talks have a * after their name.

Back to the full JSM 2021 program