Quasi-Evaluation: Assessing Interventions Without Randomized Control Groups

Quasi-evaluation (or quasi-experimental evaluation) assesses program effectiveness when random assignment to treatment and control groups is not feasible.

When Used

• Ethical/practical barriers to randomization.
• Program already implemented at scale.
• Policy changes affecting entire populations.

Common Designs

• Matching: Pair participants on covariates (age, income, etc.).
• Propensity Score Matching: Statistical model predicts treatment probability; match on scores.
• Regression Discontinuity: Compare units just above/below a cutoff.
• Interrupted Time Series: Track outcomes before/after intervention.
• Difference-in-Differences: Compare treatment/control trends over time.

Limitations

Less internal validity than RCTs; susceptible to unobserved confounding. Rigorous design and sensitivity analysis are crucial.