Quasi-Experimental design
Summary
Quasi-experimental design is a research-methods term from psychology and sociology, not physics. It compares groups exposed to different conditions without random assignment, so the groups may differ beforehand. It supports weaker causal claims than a randomized experiment.
Detail
Meaning: A quasi-experimental design compares outcomes between groups that differ in an intervention or exposure, but participants are not randomly assigned. The groups are often pre-existing, such as two classrooms, clinics, or towns, or assignment is set by circumstance or ethics. Without random assignment, the groups may already differ in age, motivation, or income, and these differences are possible confounding variables.
On the MCAT: A passage describes one school that adopts a new program while a neighboring school does not, then compares test scores. You may be asked to name the design, identify a confound or selection bias, or choose the conclusion the data justify. The usual answer is that the results show an association consistent with an effect, but they cannot establish causation as firmly as a randomized controlled trial. A pretest, a comparison group, or statistical control makes the design stronger.
Common confusion: Students call any study with a comparison group a true experiment. Random assignment settles it. With random assignment, a manipulated independent variable and a control, it is a true experiment. Without random assignment but with an intervention or exposure being compared, it is quasi-experimental. It is still useful when randomizing would be unethical or impractical.
Sources
- Psychology (Myers)
- Research Methods in Psychology
Written for the MCAT and checked before it was published. For study reference only. Spotted an error? Let us know.