Activity Number:
|
284
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Survey Research Methods
|
Abstract - #304933 |
Title:
|
Sequential Modeling for Contact and Cooperation Propensity for the United States Military Health Survey
|
Author(s):
|
Darryl V. Creel*+ and Vincent G. Iannacchione
|
Companies:
|
RTI International and RTI International
|
Address:
|
6110 Executive Blvd., Rockville, MD, 20852,
|
Keywords:
|
Nonresponse ; Sequential Weight Adjustments ; United States Military Health Survey
|
Abstract:
|
Deconstructing a response rate into contact and cooperation rates can provide insight into the nature of nonresponse. For example, the contact rate for the 14,818 Gulf War era veterans selected for the United States Military Health Survey was 70.9 percent. Among those who were contacted however, the cooperation rate was 82.4 percent. Combining these produces a 58.4 percent response rate which is more influenced by contact propensity than cooperation propensity. We used sequential weight adjustments for contact and cooperation based on the propensity of a success at each stage. This paper (1) describes the modeling methodology, (2) investigates the relationship between the contact propensity and cooperation propensity, and (3) compares estimates using weights from our two-stage methodology with estimates from weights constructed from the design weights and a single-stage adjustment.
|