|
Activity Number:
|
455
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #308041 |
|
Title:
|
Combining RDD and Web Panel Surveys
|
|
Author(s):
|
Karol Krotki*+
|
|
Companies:
|
RTI International
|
|
Address:
|
701 13th St NW, Washington, DC, 20005,
|
|
Keywords:
|
Coverage ; Nonresponse bias ; Survey mode
|
|
Abstract:
|
In light of decreasing response rates, it is important to research alternative modes. One such approach is based on web panel surveys that are relatively inexpensive and may become more viable as the proportion of Internet households increases beyond the current 70%. We discuss a 2005 smoking survey in Ohio that was implemented both via RDD and web panel. We first describe the weighting adjustments that were used to bring the data sets in line with population benchmarks. We then identify differences with respect to basic underlying demographics as well as substantive variables related to smoking. We explore models to identify differences and propose methods to adjusting for these differences. We conclude that web panel data can be made to resemble corresponding RDD survey data and that there are robust methods for combining data from these two survey modes.
|
