The Development of Hot-deck Corrected Item Mean (HDD-CIM) for Estimating Missing Data

บทคัดย่อ

Abstract

The purpose of this study was to develop the hot-deck corrected item mean (HDD-CIM) for missing data estimation and compare its under missing complete at random (MCAR) and simple random sampling with two methods, namely; Corrected item mean (CIM) and hot-deck (HDD). The secondary data from a survey of public information about the prevalence of drugs, 2011, a survey by the Bureau of Statistics Mahasarakham Province were used and the comparisons were made with three sample sizes (100, 200 and 500) and four levels of percentage of missing data (5%, 10%,15% and 20%). It appears that the HDD-CIM method has most efficiency in estimating missing data.

Keywords: Missing data, MCAR, Simple random sampling, Corrected item mean (CIM) , hot-deck (HDD)