In industrialized countries, response rates in household surveys are often low and have been steadily declining for at least two decades. Sample calibration techniques are therefore implemented, taking advantage of ancillary information to adjust the sample weights.In developing countries, response rates are typically high (usually above 90 percent). Complex adjustments of sample weights for non-response are thus not seen as critical and are rarely implemented. But there are signs that the situation is changing in some of these countries, which may affect the reliability of estimates of socio-economic indicators obtained from sample surveys.
The Sample Calibration project (jointly funded by the IHSN Trust Fund and the Knowledge for Change Program of the World Bank) aimed to assess the relevance of sample calibration techniques in the context of middle-income countries. Three countries, where “anomalies” were detected in the age distribution of the population in household surveys, have been selected for a pilot work. Sample calibration techniques were applied to the corresponding datasets.
The objectives of the project were to (i) assess if and how sample calibration techniques improve the quality of socio-economic indicators (in particular poverty and inequality measures), (ii) produce guidelines and training materials for the implementation of the techniques, and (iii) pilot a training course on sample calibration for official statisticians.
The project started with a sample calibration exercise, and with the preparation of related guidelines and training materials. This was followed by a regional hands-on training course. The workshop, organized by the World Bank was hosted by the Asian Development Bank in Manila. The workshop was attended by official statisticians from the Philippines, Thailand and Vietnam
All materials of the project, with the exception of the country survey datasets, are made available below.
|Project status:||Closed: The project was implemented between June 2015 and August 2016|
|Sponsor(s):||World Bank Knowledge for Change Program (KCP), Grant No TF0A1077 and UK Department for International Development (DFID, Trust Fund No TF011722 administered by the World Bank Development Data Group)|
|Implemented by:||The project is implemented by the World Bank Development Data Group, with support of an expert from the Italian Statistics Office. The Asian Development Bank also contributed by hosting the training workshop in Manila (May 2016)|
|Type of output:||Feasibility study report, and training materials|
|Tools||The project makes use of ReGenesees, an open source R package developed by the Italian Statistics Office|
Feasibility study report
Training program and presentations
Exercises and related datasets