University of South Carolina Scholar Commons Faculty Publications Statistics, Department of 4-2019 Median Confidence Regions in a Nonparametric Model Edsel A. Pena Taeho Kim Follow this and additional works at: https://scholarcommons.sc.edu/stat_facpub Part of the Statistics and Probability Commons Electronic Journal of Statistics Vol. 13 (2019) 2348–2390 ISSN: 1935-7524 https://doi.org/10.1214/19-EJS1577 Median confidence regions in a nonparametric model EdselA.Pe˜na∗ and Taeho Kim† Department of Statistics University of South Carolina Columbia, SC 29208 USA e-mail:
[email protected];
[email protected] Abstract: The nonparametric measurement error model (NMEM) postu- lates that Xi =Δ+i,i =1, 2,...,n;Δ ∈with i,i =1, 2,...,n,IID from F (·) ∈ Fc,0,whereFc,0 is the class of all continuous distributions with median 0, so Δ is the median parameter of X. This paper deals with the problem of constructing a confidence region (CR) for Δ under the NMEM. Aside from the NMEM, the problem setting also arises in a variety of situ- ations, including inference about the median lifetime of a complex system arising in engineering, reliability, biomedical, and public health settings, as well as in the economic arena such as when dealing with household income. Current methods of constructing CRs for Δ are discussed, including the T -statistic based CR and the Wilcoxon signed-rank statistic based CR, ar- guably the two default methods in applied work when a confidence interval about the center of a distribution is desired. A ‘bottom-to-top’ approach for constructing CRs is implemented, which starts by imposing reasonable invariance or equivariance conditions on the desired CRs, and then op- timizing with respect to their mean contents on subclasses of Fc,0.This contrasts with the usual approach of using a pivotal quantity constructed from test statistics and/or estimators and then ‘pivoting’ to obtain the CR.