This document is part of the ‘Better Evidence in Action’ toolkit.
Quasi-experimental methods are designed to explore the causal effects of an intervention, treatment or stimulus on a unit of study. Although these methods have many attributes associated with scientific experiments, they lack the benefits of the random assignment of treatments across a population that is often necessary for broad generalisability. Yet purposive sampling also has its benefits, especially when assessing small sub-groups that random sampling can miss. Researchers using these methods typically conduct tests in one of two ways: over time (pre-test, post-test) or over space (one-time comparisons), by establishing near-equivalence in factors that influence primary outcomes across treatment and control groups.