|
Activity Number:
|
465
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #307380 |
|
Title:
|
Standard Error Estimation for County-Level Radio Listening
|
|
Author(s):
|
Mandy Webb*+ and Richard Griffiths
|
|
Companies:
|
Arbitron Inc. and Arbitron Inc.
|
|
Address:
|
9705 Patuxent Woods Drive, Columbia, MD, 21046,
|
|
Keywords:
|
standard error ; generalized variance function
|
|
Abstract:
|
The County Coverage Service provided by Arbitron produces radio listening estimates at the county level. To retain MRC accreditation for this service, Arbitron seeks to find a robust methodology for estimating the precision of these radio listening estimates. Various methods for estimating standard errors are considered and compared. Arbitron started with the standard variance formula for a proportion and then included design effect adjustments that take into account the complex design and weighing procedures. Since the resulting standard error estimates will be provided to Arbitron's clients either in the County Coverage books themselves or on the Arbitron website, generalized variance functions (using the weighting, county and sample information) are also considered. Arbitron conducted an empirical study to compare the effectiveness of each of the various methods.
|
