Activity Number:
|
572
- Measuring Household Wealth in Europe: The Household Finance and Consumption Survey
|
Type:
|
Topic Contributed
|
Date/Time:
|
Wednesday, August 1, 2018 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Survey Research Methods Section
|
Abstract #327034
|
Presentation
|
Title:
|
Comparing Non-Response Adjustment Methods in the Panel on Household Finances
|
Author(s):
|
Panagiota Tzamourani* and Julian Sengewald
|
Companies:
|
Deutsche Bundesbank and University of Bamberg
|
Keywords:
|
PHF;
HFCS;
weights;
non-response;
bias;
random forest
|
Abstract:
|
This paper computes different non-response adjustment methods in the Panel on Household Finances and compares the resulting weights with respect to their effect on reducing non-response bias of key survey estimates. The Panel on Household Finances is a household survey conducted by the Deutsche Bundesbank and it forms the German component of the euro area Household Finance and Consumption Survey. The non-response adjustment methods are based on logistic response propensity models and a random forest algorithm. The estimated response propensities are used either directly to form non-response adjustments or are stratified to form response homogeneity groups. Within the latter, different non-response adjustment factors are used. The choice of the auxiliary variables used in the estimation of response propensities and the use of weights are further discussed. The study focuses on the non-response adjustment for the panel component, for which a rich set of variables from the previous wave is known for both respondents and non-respondents. These are also used to measure the reduction in bias between the resulting non-response adjusted weights.
|
Authors who are presenting talks have a * after their name.