Quasi-experimental design
resemble experimental designs, but use a quasi-independent variable.
Time series design
Collecting several observations of behaviour across time.
- Compare pretreatment and posttreatment performance.
- There is no untreated control group, thus you cannot assess whether performance would have changed regardless of treatment introduction.
Interrupted time series design: you chart changes in behaviour as a function of some naturally occurring event, rather than manipulating an independent variable.
Equivalent time samples design: administer and withdraw treatment repeatedly, most appropriate when effects of the treatment are temporary.
+ and - of quasi experiments
+ evaluate impact of quasi-independent variable under naturally occurring conditions.
+ May be able to establish clear causal relationships among variables.
- no control over confounding variables
Solution: include appropriate control groups in quasi experiments.
Non-equivalent control group design:
include a time series component along with a control group that is not exposed to the treatment.
- make comparisons that you may usually not be able to make, but validity is compromised if the two groups differ on some important variable before the begin of the study.
- To minimise this issue, groups should be matched as closely as possible and unbiased.
Pretest-posttest design
includes a pretest of participants on a dependent measure before the introduction of a treatment, followed by a posttest after the introduction of treatment.
Solomon four-group design
two-group pretest-posttest design, to which 2 additional groups are added in order to test for sensitization effects of the pretest.
Eliminating pretest
transforms the design into a simple 2-group experiment. The decision whether to eliminate the pretest depends on the research question that needs to be answered.
