Trends in Demographic and Health Survey data quality: an analysis of age heaping over time in 34 countries in Sub Saharan Africa between 1987 and 2015

• Main Authors: Mark Lyons-Amos, Tara Stones
• Format: Article
• Language: English
• Published: BMC 2017-12-01
• Series: BMC Research Notes
• Subjects: Data quality; Demographic and Health Survey
• Online Access: http://link.springer.com/article/10.1186/s13104-017-3091-x

Abstract Objective This paper evaluates one aspect of data quality within DHS surveys, the accuracy of age reporting as measured by age heaping. Other literature has explored this phenomenon, and this analysis build on previous work, expanding the analysis of the extent of age heaping across multiple countries, and across time. Results This paper makes a comparison of the magnitude of Whipple’s index of age heaping across all Demographic and Health Surveys from 1986 to 2015 in Sub-Saharan Africa. A random slope multilevel model is used to evaluate the trend in the proportion of respondents within each survey rounding their age to the nearest age with terminal digit 0 or 5. The trend in the proportion of misreported ages has remained flat, in the region of 5% of respondents misreporting their age. We find that Nigeria and Ghana have demonstrated considerable improvements in age reporting quality, but that a number of countries have considerable increases in the proportion of age misreported, most notably Mali and Ethiopia with demonstrate increases in excess of 10% points.