In an experimental study to learn the effects of a new allergy medication, researchers can select a specific subset of the population that they want to study in addition to deciding who will receive the allergy medication and who will not. Studying the effect of a social issue, such as the health effects of tanning bed use in teen mothers, presents several problems to an experimental design that the quasi-experimental design corrects. In this example, researchers utilizing an experimental design will randomly select a subset of teens to both impregnate and provide access to tanning beds in order to study the effects of tanning on the health of a teen mother. A quasi-experimental approach to this research topic will select participants who are already pregnant and compare the health effects of those who report using tanning beds versus those who do not.
Statistics generated from quasi-experimental studies do not accurately reflect how the study relates to the general public due to the lack of random sampling. Quasi-experiments also do not pre-screen applicants for experience or knowledge. Dividing a college classroom in half to test a student's ability to solve complex math problems does not take into account the level of mathematical ability each student has, their age, gender and any other variable that impacts the reliability of the statistics learned from this experiment.
Quasi-experiments are useful in case studies due to the lack of a larger population to select from. Case studies give researchers the opportunity to perform experiments on individuals with rare medical or mental conditions that do not have a large enough sample size within the general population to perform a traditional experiment.
Internal validity refers to the amount of influence a researcher has over the results of the experiment. Quasi-experimental designs are at risk for having low internal validity due to the lack of screening in participants and researcher bias.The lack of pre-screening and randomization allows for a researcher to deliberately distribute the experimental materials in a manner that will yield desired results. A teacher who gives the highest performing students in his class an experimental test because his yearly raise depends on the performance of his students is an example of researcher bias.
External validity refers to the influence of variables outside of the researchers control on the outcome of the experiment. The temperature of the room in which an experiment is run, noise from construction in a nearby construction site and the weather are all elements that impact the results of a an experiment. Because quasi-experiments are less controlled than traditional experiments, external validity in quasi-experiments is high and reflects conditions in the real world.