This tutorial talks about variables. Now, variables are the characteristics of people or things. The word variable here comes from the fact that those characteristics can vary. They can change from person to person or thing to thing.
Now, variables of interest are the characteristics of the people or things that are being studied. So there's lots of different characteristics that we could look at, and we narrow it down to just a few variables of interest. A key part of experimental studies or observational studies-- sorry, experiments or observational studies is defining what the variables are going to be. One factor that we look at is explanatory and response.
Explanatory variables are those variables that might cause an effect. And response variables are those variables that might reflect an effect. Now, you'll notice that both explanatory and response use the word might. There's some uncertainty over the cause-and-effect relationship between these two.
Now, until we've proven causality, we can't know for sure that explanatory causes the response to happen. And while these are very similar to independent and dependent variables that you might have heard about in other classes, until we've proven that this relationship exists, that the explanatory does in fact cause the response, we can't call them independent and dependent. So right now, when there's the uncertainty, we just call it explanatory and response.
So let's look at an example. In our examples, we're going to identify what the explanatory and the response variables are. So we'll highlight the explanatory variables in yellow, and we'll highlight the response variables in blue.
So in this first example, it says we have an experiment that wants to examine the effect of time exercised on weight. So the explanatory variable is one that we think might cause the effect. So we think that the time exercised might be the one causing the effect on the response variable, on the weight.
In the second example, it says, do older drivers have more accidents? So here, we think it's the age of the drivers. So the word "older" here is the key to what we think are explanatory variable is. But we're thinking that it's actually the age that's the explanatory variable. And the response variable that we're looking at is the number of accidents here. And we think that more accidents happen for older people, and that's what we're trying to prove with our experiment or observational study. And until we've proved cause and effect, then we still call it explanatory and response.
In our last example, it says, students who are smarter were exposed to classical music as a baby. So here, we think that the classical music exposure is causing a change in intelligence here with the smarter. Now, with this, we're saying that the classical music might be causing the change in intelligence. But until we have proof and certain of causality, it's only a might. So this tutorial has talked about variables, and we're talking about explanatory and response variables.
This tutorial talks about confounding variables. Before we go through what a confounding variable is and some examples, let's start by defining some key terms. Now, a variable, variables are characteristics of people or things. Explanatory variables are those variables that might cause an effect, and response variables are those variables that might reflect an effect. And in both explanatory and response, we've used the word "might." Until we've proven that there's a cause-and-effect relationship and that the explanatory does in fact cause the response, it's only a might situation.
Now, part of what puts the question mark on whether or not this relationship exists would be confounding variables. A confounding variable is an unseen variable that has a significant effect on the response variable and is related to the explanatory. It's unseen in that we're not studying it from the get-go. It's not one of the ones that we think is causing the response variable, but it does, in fact, have a significant effect on it.
For example, if we were talking about the question and trying to study, does wearing uniforms lead to higher grades? If we picked out some explanatory variables from this question, it would be uniforms. And the response variables, there's only one that we're looking at here, and it would be higher grades.
Now, with confounding variables, we're looking for something that has a relationship with the explanatory, it's connected with uniforms, and could also have an effect on higher grades. So, in this case, some confounding variables that I could think of would be type of school, parental income, and student compliance. So, for example, type of school-- private schools are more likely to wear uniforms. They have a connection between confounding and explanatory. And being at a private school could also affect the higher grades and the prevalence of higher grades, in that students who are more intelligent go to higher private schools and earn higher grades or things of that nature.
Parental income could also have a connection with the explanatory variable uniforms. If parents make more money, then they're more likely to send their students to a private school, which is also they're more likely to have students wearing uniforms. And similarly, if parents have higher incomes, then they have more time. They can help their students get higher grades.
Now let's look at a second example. In the second example, does smoking cause cancer? This is one that's been looked at many times. So the explanatory variables, we're only looking at one in this case. It would be smoking. We're thinking that smoking might cause the effect. And the response variable in this case is going to be cancer, causes cancer.
Now, one thing that is associated with smoking and with cancer and kind of puts a question mark on this cause-and-effect relationship is genetics. Perhaps there's a genetic tie that causes you to smoke more, and it also causes you to be more likely to get cancer. So in this situation, genetics would be a confounding variable, because we're not sure whether or not smoking causes cancer because of the association between genetics and smoking and the effect of genetics on cancer.
So this tutorial has explained what a confounding variable is and gone through a couple of examples of what they look like.
