Quasi-experiment refers to study designs in which the participants are not randomised to conditions, although experimental procedures are still employed. In such studies, researchers do not have complete control of independent variables because the intervention is already in process, or because it is impossible or unethical to manipulate the variable (e.g. when measuring the effects of smoking on people's health, it would be unethical to randomise people to either a smoking or none smoking group). Therefore, researchers have to rely on existing populations, such as people already smoking versus those who do not smoke. The problem with such designs is that any differences between two groups are harder to control for, giving less certainty of the cause and effect relationship, because researchers do not have complete control over independent variables. However, for many studies, this may be the only design option available. There are three approaches:

Pre-test post-test with non-equivalent groups

The experimental group is tested before an independent variable is provided (e.g. a new service) and then once again after this event. Measures are also taken for the control group pre and post intervention, although they will not have received the independent variable (they may have had nothing, or a standard and/or alternative service). If, at the outset, as far as possible, the two groups are similar, it is impossible to say that this variable has caused the outcome, but it can be reasonably assumed, increasing arguments for an additional, follow up randomised controlled trial (RCT).

Interrupted time series

Involves just one group of participants assessed on repeated occasions before an independent variable has been provided and once again afterwards to see whether any significant differences are exhibited at this later point from when they were originally tested.

Regression discontinuity

A number of participants are tested before an independent variable is applied. Those attaining a pre-defined score receive this variable, whilst those failing to do so do not. A two-group experiment then ensues, although the rationale for allocation is based on pre-test scores, rather than on randomisation.