A Nonparametric Approach for Assessing Treatment Effects Using Ranked Set Sampling with Randomized Block Designs

• Activity Number: 494
• Type: Topic Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Marketing
• Author(s): Xinlei Wang and Ou Bai*+
• Companies: Southern Methodist University and Southern Methodist University
• Keywords: two-stage designs ; cost efficiency ; distribution-free ; mixed effects models ; imperfect judgement ranking ; order statistics

Abstract

This study examines the use of ranked set sampling (RSS) with Randomized Block Designs (RBDs), for potential improvement in estimation and detection of treatment or intervention effects. We first propose new sampling schemes that can be naturally incorporated into RBDs. Then under a general mixed effects model, we propose non-parametric methods for estimating and testing treatment effects and variance components. The performance of the new estimators and testing procedures under these RSS-structured RBDs will be thoroughly examined and compared with their simple random sampling counterparts. The impact of ranking error will be examined under two commonly-used imperfect ranking models.