Online Program Home
My Program

Abstract Details

Activity Number: 453
Type: Contributed
Date/Time: Tuesday, August 2, 2016 : 3:05 PM to 3:50 PM
Sponsor: Section on Statistical Computing
Abstract #321784
Title: Optimal Stratification of Univariate Populations via StratifyR Package
Author(s): Karuna Garan Reddy* and Mohammed G. M. Khan
Companies: University of the South Pacific and University of the South Pacific
Keywords: Optimum stratification ; Mathematical programming problem ; Dynamic programming technique ; Stratified random sampling ; Optimum sample sizes ; Univariate populations

Stratifi cation reduces the variance of sample estimates for population parameters by creating homogeneous strata. Often, surveyors stratify the population using the most convenient variables such as age, sex, region, etc. Such convenient methods often do not produce internally homogeneous strata, hence, the precision of the estimates of the variables of interest could be further improved. This paper introduces an R-package called 'stratifyR' whereby it proposes a method for optimal strati cation of survey populations for a univariate study variable that follows a particular distribution estimated from a data set that is available to the surveyor. The strati cation problem is formulated as a mathematical programming problem and solved by using a dynamic programming technique. Methods for several distributions such as uniform, weibull, gamma, normal, lognormal, exponential, right-triangular, cauchy and pareto are presented. The package is able to construct optimal stratifi cation boundaries (OSB) and calculate optimal sample sizes (OSS) under Neyman allocation. Several examples, using simulated data, are presented to illustrate the stratifi ed designs that can be constructed with the proposed methodology. Results reveal that the proposed method computes OSB that are precise and comparable to the established methods. All the calculations presented in this paper were carried out using the stratifyR package that will be made available on CRAN.

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

Back to the full JSM 2016 program

Copyright © American Statistical Association