Our approach
We start from your decision, not from a methods list. Four questions decide how we approach a study, before we ever propose a method, and every stage after that is checked by someone other than the person who did the work.
Four questions we ask before we propose a method
These come before the sample size, the instrument, or the technique. If we cannot answer them, we say so, rather than proposing a study anyway.
What decision needs this evidence, and by when?
A study with no decision behind it is research for its own sake. We ask what will actually change once you have the answer, and whether our timeline can meet your deadline.
What would a good answer need to prove, and what would change your mind?
If no result could change the decision, the study is not worth running. We ask this before we ask what method to use.
Who has already answered part of this?
We check the published evidence before proposing new fieldwork. Where the answer already exists, we say so instead of billing you to find it again.
What could this study not tell you, even if it goes perfectly?
Every method has a limit. We state it before you commit, not in a footnote after you have paid for the report.
Then we choose the method, not the other way round
The four questions decide which of these fits, not a preference for one technique.
Surveys and measurement
How many people are affected, and how does it differ by group?
Qualitative research
Why is something happening, in people's own words?
Evaluation
Did it work, for whom, and at what cost?
Monitoring and verification
Did the payment, delivery or service actually arrive?
Economics and financing
What would this cost, and who benefits?
Policy and systems analysis
Why isn't a sound policy reaching people?
Evidence synthesis
What does the evidence already say?
Data systems and M&E
Can you trust the data you already collect?
How a study runs, once we start
Four stages, the same ones described in full on our About page: design, collect, analyse, translate. Each stage ends with a sign-off by someone other than the person who did the work.
01
Design
02
Collect
03
Analyse
04
Translate
What holds every stage to the same standard
“Show your working” is not a slogan on this site. It is a specific, named check on every kind of study we run, and each one lives on the page for that service.
A written protocol before fieldwork
DesignA counterfactual stated, or its absence admitted
EvaluationA search strategy fixed before results come in
Evidence synthesisA second person screens, codes or reviews before a finding is reported
Qualitative research, Evidence synthesisBack-checks on a sample of interviews
Surveys and measurement, Monitoring and verificationA certainty or confidence rating on every headline figure
Surveys and measurement, Evidence synthesisThe assumptions behind a number written out in the report itself
Economics and financingA fix-first list, not only a diagnosis
Data systems and M&EA stated rule against taking a position on the policy itself
Policy and systems analysisWhere this way of working comes from
It is not unique to us. IDinsight, a research firm founded in 2011 that works across Africa and Asia, states its approach rests on rigour tempered by cost-effectiveness, and says plainly: if a client's funds could be better used another way, it says so, even when that means less work for itself. We hold ourselves to the same discipline: a method we cannot properly deliver, or a study that would not answer your real question, is a study we will say no to.
Ready to test this against a real decision?
Tell us what you need to decide. We will apply these four questions to it and say, honestly, whether a study is the right next step.