|
Activity Number:
|
288
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #309531 |
|
Title:
|
Effects of Calculating Standard Errors Using Sampling versus Variance Weights for 2004 National Nursing Home Survey
|
|
Author(s):
|
Rong Cai*+ and Sarah Gousen and Iris Shimizu and Robin Remsburg
|
|
Companies:
|
National Center for Health Statistics and National Center for Health Statistics and National Center for Health Statistics and National Center for Health Statistics
|
|
Address:
|
3311 Toledo Road, Hyattsville, MD, 20782,
|
|
Keywords:
|
standard errors ; sampling weights ; variance weights ; ratio adjustment ; NNHS
|
|
Abstract:
|
The NNHS (National Nursing Home Survey) includes two weights. A sampling weight is used to calculate estimates and a variance weight is used to determine sampling errors of the estimates. Sampling weights are the product of ratio adjustment factors and variance weights. Software restrictions in programs such as SUDAAN, require two separate analyses to get both the estimates and their standard errors, which often confuses analysts and can be a source of error. Therefore, the differences between sampling errors based on the two sets of weights are investigated. The effects of using sampling weights instead of variance weights to compute the sampling errors in analysis are evaluated and corrections are offered to minimize erroneous analytic results.
|
