“[A] quasi-experimental design is not the method of choice, but rather a fallback strategy” (Hendrick, Bickman & Rog, 1993). 

Research can be classified a nonexperimental when it is focused on a single variable, as the example of Milgram's Obedience Study, than on a statistical relationship between two variables with ambiguous results of cause-and-effect.

   

The quasi-experiment can be seen as a derivation of nonexperimental research.  It is an alternative strategy to produce research when the conditions available to collect the data are not so favorable to the researcher. Even when it resembles an experiment,  and the quasi-experimental research strategy is designed to address cause-and-effect about the relationship between two variables, it will eventually contain some flaws that would prevent underlining the cause of a specific behavior.  Generally, this strategy will involve confounding variables that will compromise the internal validity as observed in nonexperimental research design. 

However, the major distinction between both is that in nonexperimental research strategy, the design makes less or no attempt at all to minimize threats to the internal validity. The quasi-experimental research strategy has a little more rigor to control the extraneous variables in its attempt to reach possible findings. Moreover, the quasi-experimental design cannot establish an unambiguous situation, while in the nonexperimental research results are based on the variance of scores between groups.

A significant issue faced by both strategies is related to nonequivalent group design, due to the lack of randomization pertinent to these strategies practice. It is important to acknowledge that the lack of randomization will limit the generalizability of the results to a larger population, representing a critical threat to external validity even though reducing the internal validity too.  In both cases, the researcher usually analyzes pre-existing groups, missing the chance to establish similarities among participants for avoiding assignment bias. This bias is enhanced by the inequalities of the groups and generates unambiguity in a cause-effect-explanation.  The different types of nonequivalent groups are critical to control and ensure that this kind of bias does not threaten these strategies. 

It is necessary to consider that some factors and other influences are not taken into account because the variables are less controlled in quasi-experimental research. For example, when examining the impact of drugs abuse by homeless young women, there may be other factors such as health status, level of education, and nutrition, that may be playing a role in the outcome. When other variables are not taking in the account or are not measured, the researcher has no guarantee that the treatment is the single cause of the outcome.

Moreover, the pretest-posttest can be used as a way to improve the deficiencies caused by two nonequivalent groups by decreasing the threats of internal validity in quasi-experimental designs and through the previous evaluation of the measurements.