Breath of Clarity

My Questions: Experimental Design

1. Is the presence of confounding variables a prevalent issue in quasi-experimental studies because there is no control group?

2. Why is casual comparative study using ex post facto variables only a speculative causal relationship while experiments is described as revealing a true casual relationship? Since the type of study using ex post facto variables leverages the natural variation we see in our environment, my hypothesis is it would have greater accuracy than an experiment involving intervention because there is less room for human error in the stage of implementing the intervention. In the case of a study using ex post facto variables, there is no question whether the variable the researcher is measuring was received by the participants. Further, the participants of the study would not know they are going to be analyzed until afterwards. So, there is no concern the participants would change behavior due to knowing they are part of a study.

3. Would researchers always elect to conduct a causation instead of correlational study if resources allow? A causal relationship seems to be stronger compared to a correlational relationship. What are cases where a correlational study is preferred?

Response by Edward Segura:

In regards to your first question, confounding variables can potentially pose an issue or influence the independent and dependent variable as well. Studies which incorporate quasi-experimental designs do control some confounding variables, but in no means do they have the same level of control compared to true experimental designs. (Tharenou, Donohue, and Cooper 2007, 36).

Your third question revolves around the implementation of correlational studies. Correlational studies are utilized when you seek to test a theory which incorporate dependent, independent, mediator, and moderator variables. Alternative explanations are also evaluated instead of simply only considering the direct relationship between the dependent and independent variables. This type of study focuses on real-life settings and the realistic use of people in daily situations. The hypothesis is tested on a large sampling size. The size of the sample generalizes the results and findings of the study to the general population tested (Tharenou, Donohue, and Cooper 2007, 47). In the end, researchers could possible lean towards using causal relationships if their resources available to them are appropriate, and depending exactly what he or she seeks to find in their research study (Tharenou, Donohue, and Cooper 2007, 15).


Tharenou, Phyllis, Brian Cooper, and Ross Donohue. 2007. Management Research Methods. Cambridge: Cambridge Univ. Press.

Response by Professor Kirwan:

You have some good questions and Edward answered your third one well. So here are my thoughts on the other two:

1. There are always potential confounding variables in any research involving hypothesis testing.

2. Because a true experiment controls the variables and the data is collected within that controlled environment. In ex post facto research the data is analyzed after the occurrence so the researchers have no control over the research environment. That means that the researchers can make a good guess but not actually show cause-and-effect in the right conditions.

Does that answer your questions?

My Reply:

Yes, that makes sense, thank you!