The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Online Program Home
Abstract Details
Activity Number:
|
626
|
Type:
|
Contributed
|
Date/Time:
|
Thursday, August 2, 2012 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Survey Research Methods
|
Abstract - #304912 |
Title:
|
NORCSuite_Impute: Two-Way Search Hot-Deck Imputation Macro Using SAS IML
|
Author(s):
|
Fang Wang*+ and Steven Pedlow and Yongyi Wang
|
Companies:
|
NORC and NORC and NORC
|
Address:
|
1400 E 55th PL Apt 804S, Chicago, IL, 60637, United States
|
Keywords:
|
hot-deck imputation ;
two-way search ;
donor use limit ;
number of donors
|
Abstract:
|
Traditional hot-deck imputation macros utilize SAS arrays and data steps. They keep only one nearest donor above the recipient in the sort order in the memory without limiting the number of times that a donor can be used. The new NORCSuite Impute macro using SAS IML has the following new functionalities: 1) a two-way search (above and below) algorithm that selects the global nearest donor without violating the user-specified donor use limit to control imputation bias; and 2) the number of donors kept in memory is raised automatically to avoid violating the donor use limit, and to minimize the underestimation of data variance inherent in hot-deck. The new macro also supports serpentine sorting using multiple sort variables with different variable types, imputing a set of variables together, and applying unequal donor use limits to different subsets of the data. In this paper, we will compare our new and previous hot deck imputation macros in terms of mean, variance (the amount of underestimation) and imputation bias.
|
The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.
Back to the full JSM 2012 program
|
2012 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Continuing Education program, please contact the Education Department.