What is Quasi-experiment?
A quasi-experiment is a study that resembles a randomized experiment in structure but lacks random assignment to conditions. It uses naturally occurring groups, policy changes, or pre-existing differences to approximate an experimental comparison.
How it works
Quasi-experimental designs include pre-post comparisons, non-equivalent control groups, interrupted time series, and regression discontinuity. They are used when randomization is impractical (you cannot randomly assign people to poverty), unethical (you cannot randomly assign exposure to pollution), or impossible (the policy has already been implemented). The main threat is selection bias: differences between groups may exist before the treatment and confound the results. Techniques like propensity score matching and difference-in-differences help mitigate this threat.
Applied example
A city implements a sugar tax and compares sugar beverage sales before and after the tax (interrupted time series), using a neighboring city without the tax as a comparison. This quasi-experiment cannot prove the tax caused the decline as definitively as a randomized trial, but it provides useful causal evidence given the impossibility of randomly taxing some residents and not others.
Why it matters
Quasi-experiments provide causal evidence for real-world interventions that cannot be randomized, filling the space between observational correlations and idealized experiments.
