How M&E data can hide development impact

While Brendan Brown was dispatched to Ghana to conduct fieldwork for the Food and Agriculture Organization in 2012, he noticed an odd disconnect. The official statistics on conservation agriculture didn’t seem to match up with what he was finding on the ground.

The data FAO had reported was binary: it categorized farmers as either an adopter of conservation practices, or not. In the field, Brown saw this as too simplistic. “Adoption is not an outcome, it is a process where a farmer learns, assesses, experiments and will decide how they are going to use that technology,” he explained. “If it is adopted, it might be in modified form, it might be at a lower threshold than what is totally beneficial, it might be on a few fields or they might embrace it totally.”

Now a PhD student at the University of Adelaide, Brown was determined to capture a more nuanced picture. Using data collected by the Australian Centre for International Agricultural Research between 2010 and 2016, Brown was able to re-analyze a large dataset of over 6,500 households from 1,601 villages in Ethiopia, Kenya, Malawi, Mozambique and Tanzania to determine how applying categories of adoption would impact numbers.