Innovation Advisory · Framework 12
6 questionsThe Baseline Design Test
Six questions to check whether your baseline survey can carry the endline it is meant to feed. For programme designers, MEL teams, and funders about to sign a baseline TOR.
What comparison will the endline make against this baseline?
Baselines exist to make endlines legible. Write the specific comparison you plan to make — before/after, treatment/control, cohort/cohort — on the first page of the design. If the comparison is fuzzy, the baseline is measuring things you will not need and missing things you will.
Is the sample large enough to detect the effect you are hoping to see?
A baseline that under-samples cannot be rescued at endline by adding respondents; the pre-post link is broken. Do the power calculation for the smallest effect that would matter in practice, not the largest one that would look impressive. Report the calculation with the design.
Will the sample frame you draw today still describe the population at endline?
In-migration, out-migration, deaths, new construction, seasonal households, urban expansion — all of these move the frame. If the endline will be two years out and the population is mobile, you need either a rolling frame refresh or a design that survives frame drift. Choose now.
Will you use the same instrument at endline — literally, item for item?
If the endline instrument evolves — new items added, old items dropped, wording revised because the pilot exposed problems — the pre-post comparison degrades in ways the report will find hard to defend. Freeze the core items at baseline. Add exploratory items separately. Document the freeze.
What does the baseline already know that you have not asked?
Every population has knowledge about itself that a survey cannot manufacture. Before spending money on a household questionnaire, run a listening exercise — small, qualitative, deliberately unstructured. What you hear will change the survey design, and often will change the programme design too.
How many separate comparisons will this baseline eventually feed — and are you correcting for them?
Baselines with lots of variables invite multiple-comparison fishing at endline. Decide up front how many bivariate tests you plan to run and pre-register the correction (Bonferroni is a floor, not a ceiling). If the baseline design cannot support the correction, either shrink the analysis or increase the sample.
What the six questions are doing
Read together, the six locate the six ways a baseline commonly fails to survive to endline. The comparison forces the design to name the specific evaluative question. The power ensures the sample is large enough for that question. The frame drift catches the population change that decouples baseline from endline. The instrument continuity protects the pre-post match. The already-knowns keeps the design from paying for information the community could have given for free. The corrections handles the multiple-comparison inflation that turns baselines into fishing grounds.
Run this before the survey firm is procured. The design decisions you defer to the firm end up shaping the finding you can defend.