|
Activity Number:
|
253
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #305730 |
|
Title:
|
Using Income as an Auxiliary Variable To Improve the Design of Household Expenditure Surveys
|
|
Author(s):
|
Charles Mitchell*+ and Christian Nadeau
|
|
Companies:
|
Statistics Canada and Statistics Canada
|
|
Address:
|
890 Cahill Drive W., Unit 48, Ottawa, ON, K1V 9A4, Canada
|
|
Keywords:
|
household surveys ; stratification ; allocation ; spending
|
|
Abstract:
|
When conducting household surveys, household spending habits often are linked to household income. However, for some expenditure variables, the relationship with household income is not linear. For example, the probability of reporting a nonzero value seems to increase with household income for spending on furniture and household renovations. This variety of relationships provides a challenge in constructing an efficient sampling design. The aim of this paper is to compare alternate sampling designs with the goal of reducing the variance of estimates that show different types of relationships with household income. We will consider different two-stage sampling designs using an area frame. The variance for different estimates will be compared for alternative allocation methods that use income as an auxiliary variable.
|
