Activity Number:
|
138
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, August 12, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Government Statistics*
|
Abstract - #300064 |
Title:
|
Data Quality in Federal Statistics
|
Author(s):
|
Amrut Champaneri*+ and David Banks
|
Affiliation(s):
|
U.S. Department of Transportation and U.S. Department of Transportation
|
Address:
|
400 7th Street, SW, Washington, District of Columbia, 20590,
|
Keywords:
|
data quality ; S-estimators ; total survey error ; government
|
Abstract:
|
Federal agencies collect enormous amounts of information. Much of it is of dubious quality, yet these data underpin decisions that shape national policy. This talk describes methods that can be used to measure the quality of federal databases and inferential strategies (such as robust S-estimators) that can redeem the necessary use of suspect data. These concerns connect in a natural way with the problem of estimating total survey error. Examples of problems and solutions are drawn from experience with several well-known databases.
|