|
Activity Number:
|
388
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #309925 |
|
Title:
|
Sampling and Estimation Issues in the Redesign of the Canadian Survey of Household Spending
|
|
Author(s):
|
Christian Nadeau*+
|
|
Companies:
|
Statistics Canada
|
|
Address:
|
RH Coats Building Floor 16, Ottawa, ON, K1A 0T6, Canada
|
|
Keywords:
|
Expenditure survey ; Diary ; Recall interview ; Weighting
|
|
Abstract:
|
The Survey of Household Spending is conducted by Statistics Canada at the beginning of each year through recall interviews. A major redesign has been undertaken in order to spread the collection of data over the year, to adapt the reference periods to the recall capacity of the respondents and to integrate the Food Expenditure Survey. A two-week diary will allow the collection of frequent expenditure data while less frequent data will be collected through an interview for recall periods varying from one to twelve months. The implementation of this approach raises a number of sampling and estimation issues. Among these are the combination of the monthly samples in order to produce coherent annual estimates from the interview and the diary data, and the adjustment of the estimates to the calendar year. Different options to be tested on pilot survey data will be discussed in this paper.
|
