Balanced Random Imputation for Estimating Coefficients of Correlations in Surveys

• Activity Number: 483
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): David Haziza*+ and Guillaume Chauvet
• Companies: Université de Montréal and ENSAI
• Keywords: Coefficient of correlation ; Marginal imputation ; item nonresponse ; Random balanced imputation

Abstract

Marginal imputation that consists of imputing items separately, generally leads to biased estimators of population coefficients of correlation. To overcome this problem Shao and Wang (2002) proposed a joint random regression imputation method that succeeds in preserving the relationships between two variables. One drawback of the Shao-Wang method is that it introduces an additional amount of variability (called the imputation variance) due to the random selection of residuals. As a result, it could lead to inefficient estimators. Following Chauvet, Deville and Haziza (2009), we propose a balanced joint random regression imputation that preserves the coefficient of correlation between two variables, while virtually eliminating the imputation variance. Results of a simulation study will be presented.