RCT vs Quasi-Experimental Design

When to use a randomized controlled trial vs a quasi-experimental design.

Last updated

An RCT randomly assigns who receives a program to create a clean control group, giving the strongest causal evidence. A quasi-experimental design approximates that comparison statistically when randomization is not possible, producing credible impact evidence at lower cost and greater flexibility.

At a Glance

FactorRCTQuasi-Experimental
Causal evidenceVery strong (gold standard)Moderate to strong
Requires randomizationYesNo
Comparison groupRandomly assigned controlMatched or naturally occurring
Typical cost$100K-500K+$30K-150K
Timeline2-5 years1-3 years
Statistical expertiseHighHigh
Best forStandardized, simple interventionsPrograms where randomization was not possible
Handles complexityPoorlyBetter (more flexible designs)
Donor acceptanceUniversally acceptedWidely accepted for impact evidence

Both designs try to answer the same question: did the program cause the change? The difference is how they construct the counterfactual. An RCT creates it through random assignment. A quasi-experimental design approximates it through statistical methods and naturally occurring comparison groups.

Quasi-RCT meaning: people use "quasi-RCT" in two ways. Loosely, it means any quasi-experimental design: a program group compared with a comparison group that was not randomly assigned. Strictly, it means a trial that assigns people by a rule that is not random, such as alternating or birth date. On this page, "quasi-experimental" covers the first meaning.

Frequently Asked Questions

When should I use an RCT instead of a quasi-experimental design?

Use an RCT when you can randomize (phased rollout, limited resources), the intervention is standardized, your budget supports it ($100K-500K+), and ethical review approves withholding the program from a control group. If any of these conditions are missing, a quasi-experimental design is likely more appropriate and will still produce credible causal evidence.

What is the most common quasi-experimental design in development evaluation?

Difference-in-differences (DID). It compares the change over time in program areas versus comparison areas. Most development programs collect baseline and endline data in both program and non-program sites, which is exactly what DID requires. The main assumption is that both groups would have followed the same trend without the program.