Do international surveys and censuses exhibit ‘Dry Season’ bias?

This study aims to assess whether Demographic and Health Surveys (DHS) and population censuses in developing countries are predominantly conducted during drier months and are therefore subject to seasonal bias. We examined field implementation dates of 176 DHS conducted in 64 countries and enumeration dates of the latest population censuses in 59 of those countries. We compared dry versus wet season implementation rates, distinguishing seasons using mean precipitation estimates and Koppen climate classes for each country. For countries with seasonal precipitation climates, 23% of 39 censuses were conducted in the wet season compared with an expected 47%. In the same countries, 43.3% of DHS fieldwork took place in the wet season compared with an expected 45.4%. There was thus a slight tendency for DHS and a greater, statistically significant tendency for censuses to be implemented in drier months. Thus, the recent censuses analysed exhibit greater 'dry season' bias than DHS when viewed internationally. Because past studies have identified seasonal patterns in diarrhoea incidence, water source, and malarial bednet use, care should be taken when comparing census and survey results for these variables internationally. © 2011 John Wiley & Sons, Ltd.