Over the last few post, we have been looking at various types of variables used in quantitative research. In this post, we look at several more variables. The variables examined in this post include the following.

• moderating variable
• mediating variable
• confounding variable

Moderating Variable

A moderator variable is a variable that affects or modifies the relationship between two variables. This variable is common in experimental studies. For example, let’s say you are looking at physical activity (aerobic exercise and weightlifting) influence on academic achievement. In this example, we currently have two variables as listed below

independent variable: physical activity
dependent variable: academic achievement 

For physical activity, there are two types, which are weightlifting and aerobics. Suppose that you believe that gender plays a role in the impact on academic achievement. You may think that girls will perform better academically when exposed to aerobic exercise and boys will perform better academically when exposed to weightlifting.

The results of the study show no overall difference, however, when looking at aerobic exercise, the girls who were exposed to it did better academically than the boys exposed to aerobic exercise. In addition, the boys exposed to weightlifting outperformed academically the girls exposed to weightlifting.

What is happening here is an interaction effect. Girls perform high academically when exposed to aerobic exercise while boys perform poorly. However, boys exposed to weightlifting perform better academically while girls perform poorly when weightlifting. When boys go up in performance girls go down and vice versa. This interaction effect is due to gender. Thus, gender is the moderating variable of the study as it impacts the relationship between physical activity and academic achievement. Below is a list of the three variables in the study.

independent variable: physical activity
dependent variable: academic achievement
moderating variable: gender

The conclusion from this is that a teacher should include exercise in order to boost academic achievement. However, it may be beneficial to have specific exercises available for males and females as it appears that different forms of exercise are more beneficially for one sex over another. If boys lift weights and girls do aerobics you can potentially expect maximum academic achievement.

Mediating Variable

Mediating variable is a variable that is between an independent and dependent variable. Mediating variables transmit the effect of the independent variable to the dependent variable. Let’s look at an example

Returning to our study on physical activity and academic achievement. Let’s say that we believe that physical activity leads to higher overall confidence and that the higher confidence is what directly leads to higher academic achievement. Below is a list of the variables.

Independent variable: physical activity
Mediating variable: confidence
Dependent variable: academic achievement

Mediating variables help to further explain what appears to be a simple relationship. There is more to academic achievement than just physical activity. Confidence plays a part as well. Thus mediating variables help to further explain cause and effect. Models can become endlessly complex when including mediating variables. Therefore it is up to the researcher to determine what to include when deciding how to explain a dependent variable.

Confounding Variable

Confounding variables are variables that are not a part of a study that has an unaccounted for influence on the results of a study. Returning, to our example of physical activity and academic achievement. Let’s say that we did not account for age in the study. In other words, participants in the study were as young as five and as old as 40. If we did not control for age it may impact the validity of our study. Since the academic performance of children and adults is different, age is something that needs to be controlled when doing the study. If not, the results may not be accurate.

Confounding variables are a  type of extraneous variable. Extraneous variables are any variable that is not a part of a study. When a variable that is left out of a study impacts the results it goes from only be extraneous to be a confounding variable because it is confusing the results of a study.

Conclusion

The number of available variables that can be included in a study is intimidating. The goal is not to try and figure out what variables to include and what not. Instead, the focus should always be on the research problem and the research questions that come from the problem. If this is clear, the variables that are needed will emerge and the study will go well.