Quasi-experiment
| Part of a series on |
| Research |
|---|
| Philosophy portal |
A quasi-experiment is a research design used to estimate the causal impact of an intervention. This research design is aimed at assessing the difference between outcomes (e.g., reading knowledge, depressive symptoms) in a group that experienced an intervention and a group that did not. The intervention is broadly construed such that it could be designed by researchers (e.g., a reading program) or it could be an event affecting a group of people such as disaster (e.g., an earthquake). Quasi-experiments share similarities with experiments and randomized controlled trials, but specifically lack random assignment to intervention and control conditions. Instead, quasi-experimental designs typically compare groups that are either preexisting (e.g., whether someone was exposed to COVID-19) or groups that were created without random assignment (e.g., students attending schools with different reading programs).
The causal analysis of quasi-experiments depends on assumptions that render non-randomness irrelevant, for example, the parallel trends assumption for difference-in-differences (DiD) method. Quasi-experiments are subject to concerns about internal validity. In other words, it may be difficult to convincingly demonstrate a causal link between the treatment and control conditions on the observed outcomes in quasi-experimental designs. This is particularly true if there are confounding variables that cannot be controlled or accounted for. A major impediment to drawing causal conclusions in quasi-experiments is the possibility that unmeasured variables could have contributed to treatment–control group differences in the outcomes, which is less a problem in true experiments because randomization largely equates the intervention and control groups on both measured and unmeasured background factors, particularly when there are large numbers of experimental units to randomize. The experimental units most commonly used in quasi-experimental research are humans although sometimes organizational units such as schools and divisions in an organization are employed.