Quasi-experimental and factorial research designs
The use of appropriate controls
- Sound experimental research designs allow to eliminate threats to internal validity by :
• The use of random assignment
• The use of similar testing conditions.
- Not every design that looks experimental is experimental.
- The ability to run a t-test on data does not warrant a cause-and-effect explanation.
Three non-experimental design
Differential research The comparison of pre-existing groups
One-Group Pretest-Posttest design One group is measured before and after
Posttest-only nonequivalent control One group is measured after a treatment
group design while another is measured without having
received a treatment
- Differential research cannot really determine cause. Other things can come into play, and
so you can’t make a solid conclusion on it
- One-group... can draw cause from this design, cannot draw a certain conclusion***???
- Posttest-only... don’t know what happened before
Two classic non-experimental designs
- Cross-sectional design: Adevelopmental design comparing different groups of
individuals, each group representing a different age.
- Longitudinal design: Adevelopmental research design that examines development by
making a series of observations or measurements over time. Typically, a group of
individuals who are all the same age is measured at different points in time.
- Aresearch strategy that attempts to limit threats to internal validity and produce cause-
and-effect conclusions (like an experiment), but lacks one of the critical components—
either manipulation or control—that is necessary for a true experiment. - Typically compares groups or conditions that are defined with a nonmanipulated variable.
• School-based research
- You can draw a causal conclusion
- You have limited ability to do random assignment
An example of Quas