Activity Number:
|
42
|
Type:
|
Contributed
|
Date/Time:
|
Sunday, August 11, 2002 : 4:00 PM to 5:50 PM
|
Sponsor:
|
Section on Survey Research Methods*
|
Abstract - #300993 |
Title:
|
Flexible Matching Imputation in the Manufactured Homes Survey
|
Author(s):
|
Bonnie Kegan*+ and Todd Williams
|
Affiliation(s):
|
U.S. Census Bureau and U.S. Census Bureau
|
Address:
|
FOB-4 Room 2136, Washington, District of Columbia, 20233,
|
Keywords:
|
Hot-deck imputation ; Flexible matching
|
Abstract:
|
The Manufactured Homes Survey (MHS) tracks a sample of shipments of homes from manufacturer to dealer to buyer, until they are placed or destroyed. Currently, missing data in the MHS are imputed using a hot-deck procedure that selects donors within groups formed by selection month and dealer state. This method does not consider relationships between the variables when selecting a donor. Flexible matching imputation (FMI) combines hot-deck imputation with model-based methodology. This method identifies and ranks matching variables for a given set of missing variables by fitting models to the set of observed data for the variables. It then uses the specified matching variables to impute missing data using hot-deck methodology. The purpose of our research is to show that, by using the FMI procedure for imputing missing data in the MHS, we are able to preserve the multivariable relationships found in the data. We will perform the current hot-deck method along with the FMI method on simulated missing data to compare the imputed data to the actual observed data. Our comparisons will also include imputed data derived directly from the fitted models.
|