- We have learned how confounding variables  obscure the effect of an independent variable and how identifying a confounding variable may  lead us to be skeptical about a hypothesis. In this module, you will learn about  another concept, intervening variables. It is common for students to mix up  confounding and intervening variables, but the two concepts are entirely  different. In the social sciences, to find that one variable causes change to  another is not enough. For good social science, we also need to know how and why one variable  causes change to another. To investigate how and why one variable causes changed to another, we  must master the concept of intervening variables. The objective of this video is to teach  you to define intervening variables, propose an intervening variable  when presented with a hypothesis, and identify the difference between  confounding and intervening variables. Suppose you're a policymaker in your  school district and you read a paper that shows that reducing class sizes is good for  learning. Schools where classrooms have fewer students do better at standardized tests than  schools where classrooms have more students. Now suppose that you, as a policymaker who cares  about your district, decide to reduce class sizes so that students can learn better.  Should you believe that reducing the class size without any additional step would  be sufficient to help students learn better? You shouldn't, because to put  scientific knowledge to good use, we must not only know whether a policy  intervention works. We must also know why it works. And to know why something works, We  must master the concept of intervening variables. Let's assume that reducing the class size works because smaller classes allow teachers to spend  more time giving students individual instruction. In this case, we have a dependent variable, an  independent variable, and an intervening variable. The dependent variable is student performance,  the independent variable is the class size, and the intervening variable is the  time spent on individual instruction. Notice that the time spent on individual  instruction explains why reducing the class size leads to better student performance. Class sizes  make it possible for the time spent on individual instruction to increase, and when teachers spend  more time - more one-on-one time - with students, students learn more and perform better.  This means that one-on-one time, which is our intervening variable,  helps us understand the causal process. Be sure to note that intervening and confounding  variables are different concepts. In my experience as a TA, I've seen many students confuse them.  Let's see how those two concepts are different. Intervening variables are caused  by the independent variable and cause the dependent variable. An intervening  variable explains the causal link between the independent and the dependent variable.  Confounding variables, on the other hand, obscure the causal link between the  independent and the dependent variable. To see why one-on-one time is  not a confounding variable, let's look at the criteria  of confounding variables. So one-on-one time causes change in the dependent  variable. That is, it causes change in student performance. It checks the first box. One-on-one  time is correlated with the independent variable. That is, where we see smaller classes, we  tend to see teachers spend more one-on-one time with students, so it checks the second box.  However, one-on-one time comes after, not before, the independent variable. The smaller  class size is what allows teachers to spend more one-on-one time with  students. It does not check the third box. Therefore, one-on-one time is not a confounding  variable. In our example, one-on-one time is an intervening variable. An intervening variable  is the middle link in the causal chain. In conclusion, the goal of this video is to  teach you to define intervening variables, propose an intervening variable when presented  with a hypothesis, and identify the difference between confounding and intervening variables.  Whether or not you become a policymaker, be sure to know the difference between  confounding and intervening variables.