Experimental vs. Quasi-Experimental Designs

Solution Preview

There are numerous features of true experiments. For one, true experiments are more easily conducted in laboratory rather than natural settings. This is because laboratories allow for more control over the experiment particularly via random assignment to the conditions of the experiment. Thus, control is one feature of true experiments; it increases the internal validity of the experiment (but not the external validity), and it leads to unambiguous outcomes regarding what caused the results of the experiment. For instance, when the conditions of the experiment are controlled, any difference in the dependent variable (DV) can logically be attributed to the IV (e.g. treatment and comparison conditions), and it helps to rule out errors or alternative explanations for a researchers results. A true experiment then, involves the manipulation of an IV, and it determines the IV's effect on the DV. The final feature of a true experiment is that there are usually two similar groups that are treated exactly the same except for the factor of interest (i.e. the IV), here again to rule out alternative plausible explanations for the results of our experiment. Random assignment to the conditions of the experiment, however; "is often seen as the most critical defining characteristic of the true experiment" (Shaughnessy et al., 2006, p. 359).

As for threats to internal validity in true experiments, there are several of these, and here again, they are seen as plausible alternative causes of a researchers results. They are usually called "confounds", and there are eight of them according to Shaughnessy et al. (2006, p. 363). History can be one confound, and this occurs when the two groups in the experiment (the experimental and control group) are exposed to different historical factors during the time of the experiment that may have an effect on the experiment's outcome. For example, if we are testing the two groups knowledge about algebra and some of the people in one group spend time studying this subject during the time of the experiment while others in the experiment do not, this will lead to an increase in that group or part of that groups knowledge of the subject outside of the experiment, and this in turn will skew the results. To avoid the history effect, both groups need to be exposed to the same conditions of the experiment both inside and outside the laboratory.

Testing is another confound or threat to the internal validity of a true experiment. For instance, if group A is given a pretest and then a posttest on algebra knowledge and group B is not given the pretest, then group A has not been treated exactly the same as group B, and will clearly have an advantage over group B in this case. Taking the pretest will cause group A to have better results ...