|
Activity Number:
|
500
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #306023 |
|
Title:
|
Fractional Imputation for Missing Values in Linear Regression Models
|
|
Author(s):
|
Minhui Paik*+ and Michael D. Larsen and Shin-Soo Kang
|
|
Companies:
|
Iowa State University and Iowa State University and Iowa State University
|
|
Address:
|
1425 Coconino Road, 308, Ames, IA, 50014,
|
|
Keywords:
|
nonresponse ; variance estimation ; jackknife ; response probability ; survey sampling ; replication
|
|
Abstract:
|
Sample surveys typically gather information on a sample of units from a finite population. They frequently have missing values for some variables for some units. Fractional imputation using a regression imputation model selects multiple values from the set of observed residuals and produces multiple predicted values for each missing value by adding residuals to a regression prediction. The method assigns fractional survey weights to the imputed values. In this paper, we consider the situation in which some covariate information is missing. In particular, attention is paid to the performance of fractional imputation using a regression imputation model for estimating regression coefficients when missing covariate data is either missing completely at random or missing at random.
|
