|
Activity Number:
|
428
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #310078 |
|
Title:
|
The Use of Propensity Score Matching Methods in Mixed Mode Studies
|
|
Author(s):
|
John Bremer*+ and Kerri Miller
|
|
Companies:
|
Harris Interactive and Harris Interactive
|
|
Address:
|
161 Sixth Avenue, New York, NY, 10013,
|
|
Keywords:
|
statistical matching methods ; multi-mode surveys ; survey research ; propensity score ; MSE ; bias
|
|
Abstract:
|
The use of multiple data collection modes in a single survey instrument is gaining in popularity. Different modes allow for more flexible control of the sample and for more advanced methodologies to be used to investigate the problem at hand. Little has been written with regard to how to combine data from different data collection modes. Often the data from different modes are combined in a simple manner by simple stacking. This technique can produce bias in the results. This paper proposes the use of an application of propensity score matching in order to more accurately combine data from multiple sources. There is a tradeoff between bias and variability in incorporating this technique but generically the MSE of using a propensity score approach is lower than the MSE of using a stacking method. This paper will examine the conditions under which this is specifically the case.
|
