Summary/Abstract |
By means of the latest seven national household surveys of the Democratic Republic of the Congo, this article uncovers the very volatile sampling frame used underneath all survey designs. As a result, the reliability of much associated survey information as well as any corresponding temporal analysis are seriously jeopardized. Relying on recent vaccination, school enrolment and election data, the article proposes a post-stratification technique to retroactively control for these erratic variations in sampling frame in an attempt to identify real socio-economic trends. Although the proposed technique did not restore full comparability of survey data in all respects, it has been able to eliminate an essential part of the spuriousness as illustrated by assessing trends in asset ownership under both the biased and stabilized sampling frames.
|