[Home ] [Archive]   [ فارسی ]  
:: Main :: About :: Current Issue :: Archive :: Search :: Submit :: Contact ::
:: Volume 19, Issue 1 (9-2008) ::
مجله‌ی بررسی‌ها 2008, 19(1): 95-118 Back to browse issues page
Comparison of the Two Imputation Methods “Little-Su” and “Nearest Neighbor” in Panel Surveys by a Simulation Study
Elham Bakhshandeh *, Reza Navvabpour
Abstract:   (3475 Views)

Abstract. َA type of survey called “repeated survey” or “sampling over time” which is repeated continuously during time periods is common in many statistical systems. Such surveys are often performed in one of the forms of repeated cross-sectional, rotation, or panel surveys. Panel survey is a method in which a “fixed sample” is contacted during different waves. In panel survey, there is another type of non-response in addition to item non-response called “wave non-response”. Wave non-response occurs when one or more waves of the panel data are missing for a unit for which information has been provided for at least one wave.

When repeating observations on units over time, estimate of parameters change between waves are very important. Therefore, neither can we introduce the variability of the estimates of change by imputing solely based on cross-sectional information, nor would we like to decrease the variability of the change estimates by imputing using information of one sample unit. Clearly, the imputation techniques used in the subsequent waves of a panel survey are needed to incorporate the information collected in previous waves, among which “Little-Su imputation” can be named as a useful method. In this paper, types of panel surveys and “wave non-response” are studied, and “basic Little-Su” and “nearest neighbor” imputation methods are introduced. Finally, the effects of application of Little-Su imputation on efficiency and absolute bias of the statistic of interest are assessed by simulating a three-wave panel survey, and the method is compared to the nearest neighbor imputation (with respect to sample size effect, wave non-response rate, and correlations between waves). The results of this simulation study show that most of the times the “Little-Su” method gives more reliable estimates than the nearest neighbor imputation method.

Keywords: Basic Little-Su imputation, Efficiency, Nearest neighbor, Panel survey, Simulation, Wave non-response.
Full-Text [PDF 406 kb]   (666 Downloads)    
Type of Study: Research | Subject: General
Received: 2009/04/6 | Accepted: 2009/07/4 | Published: 2016/03/1
Send email to the article author

Add your comments about this article
Your username or Email:


XML   Persian Abstract   Print

Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:

Bakhshandeh E, Navvabpour R. Comparison of the Two Imputation Methods “Little-Su” and “Nearest Neighbor” in Panel Surveys by a Simulation Study. مجله‌ی بررسی‌ها 2008; 19 (1) :95-118
URL: http://ijoss.srtc.ac.ir/article-1-114-en.html

Rights and permissions
Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Volume 19, Issue 1 (9-2008) Back to browse issues page
مجله‌ی بررسی‌های آمار رسمی ایران Ijoss Iranian Journal of Official Statistics Studies
Persian site map - English site map - Created in 0.05 seconds with 30 queries by YEKTAWEB 4540