|
Activity Number:
|
31
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #306388 |
|
Title:
|
A Micro-Level Latent Class Analysis of Underreporting on the Consumer Expenditure Survey
|
|
Author(s):
|
Brian Meekins*+ and Clyde Tucker and Paul Biemer
|
|
Companies:
|
Bureau of Labor Statistics and Bureau of Labor Statistics and RTI International
|
|
Address:
|
2 Massachusettes Ave., NE, Washington, DC, 20212,
|
|
Keywords:
|
latent class analysis ; measurement error ; consumer expenditures
|
|
Abstract:
|
This paper expands upon recent work by the authors using latent class analysis (LCA) to estimate the amount of underreporting on the BLS Consumer Expenditure Quarterly Survey (CEQ). Using micro-level or interview-specific indicators a latent construct is formed that indicates degree of underreporting of expenditures by those reporting a purchase of any item within that category. Previous models were applied to household total quarterly expenditures and a small number of commodity categories. This work expands the number of commodity categories examined using a variety of LC models. Best fitting models are determined from well-known statistical tests as well as substantive diagnostics developed by the authors. Data from the CEQ for the years 1996 to 2003 are used in the analysis. An evaluation of possible causes of underreporting is undertaken using multinomial logistic regression.
|
