Not all extraneous variables become confounding variables. To explain when extraneous variables become confounding variables, it is helpful to discuss confounding variables within the context of internal validity. In this respect, the results from a study are internally valid when we can conclude that there is only one explanation for our results. To explain what we mean by this, let's go back to our example:
Study #1
The relationship between background music and task performance amongst employees at a packing facility
Let's imagine that we have collected the data from our experiment. From this data, we know the average (i.e., mean) number of tasks performed each hour (i.e., our dependent variable: task performance) for the control group (i.e., without background music) and treatment group (i.e., with background music). We use statistical analysis to compare the differences in the number of tasks performed for the two groups. We find that (a) there is a difference between background music, (b) the difference is statistically significant, and (c) the difference equates to a 10% increase in task performance. We find that (a) there is a positive relationship between background music and task performance; (b) this relationship is statistically significant; and (c) the gain score is about 10%. In other words, we found that the addition of background music improved the task performance of employees by 10% compared to the control group that had no background music. Since our statistically analysis shows that the relationship between background music and task performance was statistically significant, we conclude with some confidence that background music improves task performance in the packing facility.
The question arises: Can we conclude with confidence that background music improves task performance in the packing facility? If we can, we can say that the results from our study are internally valid.
The simple answer to this question is NO. No experiment can be 100% internally valid. However, the goal of quantitative research designs, and experimental research designs, in particular, such as the one we illustrate in our example, is to reduce the possible threats to internal validity [see the article: Internal validity]. We want to make our results as internally valid as possible. This means trying to ensure that there are as few explanations for our results as possible; remember that our results are only completely internally valid when there is only one explanation for our results.
The question arises: Could any of the other extraneous variables that we identified provide an alternative explanation for our results?
The answer to this question is YES. Any of the extraneous variables that we identified, whether those relating to individual differences (e.g., existing employee task performance, employee age and gender, etc.), the environment in which the study was conducted (e.g., the climate inside the facility, especially if the packing facility is not air conditioned/heated; the weather outside, which could affect employee mood, etc.), as well as factors relating to the independent variable (e.g., type of music, loudness of music, time of day), and the dependent variable (e.g., employee tiredness - number of shifts - employee motivation, job satisfaction, etc.), could have all provided an alternative explanation for our results. In other words, they could have all threatened the internal validity of the results. However, just because an extraneous variable can be a threat to internal validity does not necessarily make it a confounding variable. This fact - that all extraneous variables are not necessarily confounding variables - is really important in understanding those factors that pose a threat to the internal validity of your study. After all, of the many hundreds or thousands of extraneous variables that relate to your study, only a very small number may act as confounding variables. This means we need to think about extraneous variables in two ways:
There are too many extraneous variables in experiments to worry about whether each one is a possible confounding variable. However, when considering the more obvious extraneous variables in your experiment that might become confounding variables, ask yourself two questions: QUESTION #1: What academic evidence is there that the extraneous variables that you can identify affect the dependent variable? and QUESTION #2: Are there any logical or practical reasons to assume that an extraneous variable might become a confounding variable?
For example, imagine our example study examining the relationship between background music and task performance amongst employees at a packing facility. Some of the potential extraneous variables that we have identified that could have affected task performance (i.e., our dependent variable) include the time of day that the study was conducted, which could affect the tiredness of employees; the weather, which could affect employees? mood on the day of the experiment; the loudness of the background music being played, which may affect employee concentration, and so forth. The question arises: Does the academic literature suggest that any of these variables (e.g., time of day, tiredness, the weather, music volume, etc.) affect task performance? In other words, are there any other studies that have been published on task performance that examined these extraneous variables and their relationship with task performance? If such academic research exists, what did it tell us about this relationship? For example, perhaps a previous study of employees in a call centre showed that tiredness reduced task performance. If this was the case, there would be justification to think that the time of day that the experiment was conducted; or more precisely, the beginning or end of an employee's shift might affect employee tiredness, and therefore, task performance. This may lead us to believe that such an extraneous variable could become a confounding variable [see the section: Extraneous variables that could become confounding variables]. However, if there was limited or no academic evidence to suggest that an extraneous variable might affect the dependent variable in a way that could be confounding, we could choose to ignore the extraneous variable.
Whilst it is possible that there is limited or no academic evidence to suggest that an extraneous variable could affect the dependent variable in a way that could be confounding, we may still consider there to be logical or practical reasons to think that this might happen. For example, imagine that there was no or limited academic evidence to suggest that tiredness reduced task performance. Just because this had not been studied and discussed in the literature, it would still be logical to assume that in a physical job, such as in a packing facility, where employees are constantly 'on the move', tiredness would play some role in task performance. Therefore, we may still feel that it could be a confounding variable.