Publications Summary

Document Type
Methodological Reports
Publication Topic(s)
Anthropometry/Biomarkers, Survey Methodology, Data Quality, Analysis
Angola, Benin, Burkina Faso, Burundi, Cameroon, Chad, Congo Democratic Republic, Cote d'Ivoire, Gabon, Gambia, Guinea, Kenya, Lesotho, Liberia, Malawi, Mali, Mozambique, Namibia, Niger, Nigeria, Senegal, Sierra Leone, South Africa, Tanzania, Togo, Uganda, Zambia, Zimbabwe, Albania, Armenia, Egypt, Jordan, Yemen, Kyrgyz Republic, Tajikistan, Bangladesh, Cambodia, India, Maldives, Pakistan, Timor-Leste, Papua New Guinea, Dominican Republic, Haiti, Honduras, Peru
Recommended Citation
Benedict, Rukundo, Sorrel Namaste, and Trevor Croft. 2022. Evaluation of Implausible Anthropometric Values by Data Collection Team in Demographic and Health Surveys 2010–2020. DHS Methodological Reports No. 33. Rockville, Maryland, USA: ICF
Download Citation
RIS format / Text format / Endnote format
Publication Date
September 2022
Publication ID

There is no printed copy available to order.


High quality anthropometry data is required to inform country and global decisions on nutrition policies and programs in low and middle-income countries. In this report, we examine and compare data quality between data collection teams that measure anthropometric data in Demographic and Health Surveys (DHS) surveys. The goal is to inform continuous quality improvement efforts at The DHS Program. All available DHS data from 2010–20 with height and weight measurements for children age 0–59 months of age were included. The percentage of implausible anthropometric data per collection team was examined within the same survey and compared across surveys. Implausible anthropometric data were defined as height-for-age z scores (HAZ) below –6 SD or above +6 SD, weight-for-height z scores (WHZ) below –5 SD or above +5 SD, and weight- for-age z scores (WAZ) below –6 SD or above +5 SD. The acceptable level of implausible anthropometric data was defined as less than 1% according to World Health Organization recommendations. A total of 90 DHS surveys with HAZ, WHZ, and WAZ were identified in the persons recode data files. This resulted in a total sample of 871,629; 871,069; and 875,277 children, respectively. By survey, the number of surveys with a percent implausible less than 1% for HAZ, WHZ, and WAZ was 41, 32, and 80 of 90 surveys, respectively. The median of the mean percentage of implausible HAZ, WHZ and WAZ across surveys was 1.4%, 1.3% and 0.2%, respectively, while the median of the inter-quartile range of implausible HAZ, WHZ, and WAZ values was 1.1%, 1.3% and 0.3%, respectively. By teams, the median percentage of teams with implausible HAZ, WHZ, and WAZ values above the 1% threshold was 42%, 50%, and 1.4%, respectively. Team variability was common across all regions, but decreased over time. The DHS Program supports surveys in areas with hard-to-reach populations and insecure environments and this may partially contribute to variations in data quality between surveys and teams. Our findings suggest that poor performance by teams contributes to overall survey quality. In many surveys, there was team variability that declined in more recent surveys. This reinforces the importance of training and data quality measures that improve the quality of anthropometric data.


Browse for Publications by:

Browse for Journal Articles based on DHS data by: