Data comes in many forms, from the quantitative realm of questionnaire scores, biomarkers and financial metrics, to the qualitative world of narrative and observation. At GiveDirectly, we strive to be rigorous, honest and respectful in how we use all types of data to describe the impact of cash transfers to the extreme poor. Different types of data tell us different things about impact, so we publish a range of research and analysis about our programs. Starting this week we are also publishing an unfiltered stream of qualitative data direct from the recipients of cash transfers through GDLive. After all, stories are not the opposite of data — they are just another type of data.

GiveDirectly’s cash transfers are based on evidence from dozens of randomized controlled trials (RCTs) — highly valued empirical, scientific studies — from around the world that show cash transfers work. Our own RCT in Kenya told us that GiveDirectly’s particular model of cash transfers have positive impacts on range of outcomes including food security, mental health, domestic violence, earnings and assets. It’s a careful quantitative analysis of what happened to the 503 households who received cash transfers, and how their lives were different a year later to the 505 households who didn’t. The rigor provided RCTs is vital for policy-makers (as well as some individual donors) in making decisions about whether channel international aid through cash transfers. Without this sort of evidence, we are shooting in the dark.

But in the distillation of so many lives, we lose a lot. We lose the human stories of how people used their cash transfers, and what difference it made to them individually. After all, part of the point of giving cash rather than specific goods is that poor people are experts in their own situation and can spend cash in the ways that will benefit them most, whether that’s a durable roof, livestock or school fees. We want to present these human stories honestly and respectfully. We also want to adhere to a common principle across both qualitative and quantitative data: that the data is representative of the group, not cherry picked. That’s why the stories you can see on GDLive are unfiltered and unedited — not hand-selected, curated, or tweaked in any way — and why we’ve been careful to ensure that none of our recipients feel compelled to participate.

RCTs and GDLive show our determination to communicate the impact of GiveDirectly’s programs fully and fairly. Please tell us how you think we’re doing, and where we should go next. We’d love to hear your ideas.


Michael Cooke is GiveDirectly’s Research Director. Prior to joining GiveDirectly, he spent six years at UK healthcare charity Marie Curie in impact evaluation, data analysis, and strategic development. Michael holds a B.A. in experimental psychology from Oxford University and a Ph.D. in neuroscience from King’s College London. He is also a Clore Social Leadership Fellow.

Back to List