NEW · Free logframe template with a worked example — enter your email to get it →
Complex methods

Experimental and quasi-experimental methods in M&E: when to use each

Nov 21 2026 · Complex methods

Randomised designs, comparison groups and when each is worth the cost.

Short answer

Experimental designs (randomised controlled trials) randomly assign participants to treatment and control groups, providing the strongest basis for attribution. Quasi-experimental designs use a comparison group without randomisation, such as matched groups, regression discontinuity, or difference-in-differences. Use experimental designs when randomisation is ethical and feasible, and when you need robust causal evidence. Use quasi-experimental designs when randomisation is not possible but a credible counterfactual can be constructed. Both require you to identify and address threats to validity, including selection bias, attrition, and spillover. Without a comparison group, you cannot attribute changes to your intervention. Choose based on the evaluation question, resources, and ethical constraints.

What is the difference between experimental and quasi-experimental designs?

Experimental designs, such as randomised controlled trials (RCTs), randomly assign participants to treatment and control groups. This randomisation ensures that known and unknown confounding factors are balanced, providing the strongest basis for causal attribution. Quasi-experimental designs also use a comparison group, but without randomisation. Instead, they rely on methods like matching, regression discontinuity, or difference-in-differences to construct a credible counterfactual. The key difference is that quasi-experimental designs cannot fully rule out selection bias, so they require careful justification and sensitivity analysis.

DesignComparisonStrength
RCTRandomisedStrongest
MatchedMatched groupStrong if matched well
Diff-in-diffBefore/after, two groupsGood with trends
Regression discontinuityA thresholdStrong near the cut
🔒Free · no card · 20 seconds

Keep reading — free.

Create an account to unlock the rest of this guide — the worked example, the tables and the checklists — and get The Bridge weekly. Already a member? Sign in.

You'll also get the free handbook by email. Unsubscribe anytime.

Written for M&E professionals. Not a guarantee of accuracy — always verify every claim against its source, and let a named human sign off. Get the free handbook.

Experimental and quasi-experimental methods in M&E: when to use each · Evaluation AI