Lesson video

Hello everyone, it's Mr. Millar here.

In this lesson, we're going to be talking about correlation and causation.

So, first of all, I hope that you are all doing well, and over the last few lessons, we've been looking at scatter graphs and this idea of correlation and lines of best fit.

So for the Try This task, we have got a scatter graph with some points plotted on it, and we've got a line of best fit as well.

And as Zaki's statement is saying, "This graph shows that book sales cause car accidents." So what do you think, do you think that this statement is true or false? Why or why not? What I want you to do is pause the video for one minute to write down a couple of sentences about the statement.

Pause the video now.

Okay, great, so I hope that you all said that Zaki's statement is incorrect.

And that clearly is incorrect, because the idea of higher book sales causing car accidents is clearly ridiculous.

There is no way that if more books are sold, then that causes car accidents.

That's just not true.

But what is interesting here is that clearly the scatter graph shows that there is a positive correlation.

So, well done if you mentioned positive correlation in your answer.

So the graph does show that as the number of car accidents goes up, so does the number of book sales, but just because there is a positive correlation here does not mean that one causes the other.

And it's this point that we are going to be talking about in more detail for the rest of this lesson.

So when you're ready, let's move ahead to the Connect slide.

Okay, so let's have a look at the Connect slide now, which says correlation doesn't necessarily mean causation.

So as we saw on the last slide, just because two things happen to be correlated, it doesn't mean that there is any causation going on here.

Okay, so underneath here there are three statements, and in each of these statements there is a correlation, either a positive correlation or a negative correlation.

So I want you to pause the video, to write down a sentence or two on each statement.

And what I want you to think about is is there any causation going on here? What is going on behind the following correlations? Pause the video for two or three minutes, one sentence on each of these statements, pause the video now.

Okay, great, let's go through these statements, then.

So the first one, "There is a positive correlation between lung cancer and smoking." And if you've studied this before, you would definitely know that this is true.

And what is going on here? Well, clearly the fact is that higher rates of smoking, if you smoke, you are more likely to develop lung cancer.

So there is a causal relationship here.

There is some causation happening because smoking causes your chances of getting lung cancer to be higher.

So there is a time correlation going on here.

The next one, "There is a negative correlation between temperature and heating bills." Well, again, there is definitely some causation going on here because if there are higher temperatures, so if the temperature is higher, that means that because it's warmer out, you have the heating on less.

And if you have the heating on less, obviously your heating bill is going to be lower.

So the first two statements, there is a correlation going on and there is also causation.

What about the third one? Well, there is a correlation again here, because when you have a look at the number of ice creams sold and number of barbecues, you would see that when there are more ice creams sold, there are more barbecues, but is there causation here? Does more ice creams sold lead to more barbecues or do more barbecues lead to more ice cream sold? Well, I don't think so.

I think that there is not necessarily any causation going on here.

And what is going on here, of course, is that when it is hot out, then you would typically see more ice creams sold and more barbecues, so it's something else that is causing both of these things to increase at the same time or decrease at the same time.

So just because there is correlation here, it does not mean that one thing is causing the other.

It's also worth pointing out that for the first two statements, the causation only goes one way.

So we say that smoking causes lung cancer, but we wouldn't say that lung cancer causes smoking, and the same thing for temperature and heating bills.

Sometimes there is causation both way, we do see some examples of one thing causing the other, and then the other thing causing the other thing, it's kind of this circular loop, but not necessarily.

So hope that is nice and clear.

Let's now move on to the independent task.

Okay, so for the independent task question one, you have got five different boxes here.

In each box, there are two variables, and I want you to do two things here.

First of all, decide if you think these variables are correlated.

So is there positive correlation, negative correlation or no correlation? And second of all, explain whether there is any causation going on here.

So maybe there is some causation, maybe there isn't, it's up to you.

The second question shows a scatter graph between cheese consumption and number of the PhDs awarded, and there's a statement about cheese makes you clever, so I want you to see if you agree with that statement.

So pause the video, five or six minutes, to write down some sentences for these two questions here.

Okay, great, so let's go through these questions.

So the first one we'll go through them very quickly.

Your last math test score and how far you can run.

Well, I doubt that there's any correlation going on here and also no causation as well.

Next one, height and age of secondary students.

Well, hopefully you said that there is a positive correlation because if you're taller, you're more likely to be older.

And what about the causation here? Well, clearly it is age causing height because as you grow older, you grow taller.

So definitely age is causing high here.

Next one, age of the car and price of the car.

Well, we've seen this one in a lesson a few lessons ago.

There is negative correlation typically here.

And you know, because as the age of the car gets older, the price goes down.

So here, the causal relationship is as the age increases, the price of a car typically goes down.

So there is negative correlation and causation here as well.

Next one, number of ice creams sold and air temperature.

That's nice and obvious.

The higher the temperature, the more ice creams sold, so a positive correlation here, and air temperature causing ice creams sold to be higher.

The final one, distance you live from school and the time it takes you to get to school.

Well, again, you should have said a positive correlation here and clearly the longer you live from school, the more time it's going to take you to get to school, so there's causation going on here.

So actually for all of these, excuse me, all of these four statements where we had some correlation, there is also a causation going on here, but this is not always the case.

For example, in question two, we can see a positive relationship between cheese consumption and the number of PhDs awarded, but just because there is positive correlation, does that mean that there is causation here? So is the statement, "Cheese makes you clever," true? Well, hopefully you said that this is very unlikely because clearly there are other things going on here, it's not the case that cheese makes you clever, it's definitely not the case.

So how have a think about the second question here, what could cause this positive correlation here? Why do you think this is happening? Well, one reason might be that if you are wealthier, you are more likely to have a PhD.

And also if you're wealthier, you're more likely to eat more cheese.

That could be one thing that is behind this positive correlation, but there is definitely no causation going on here.

So anyway, let's now have a look at the Explore task.

Okay, in the Explore task, we have got five variables below: cycle ownership, greenhouse gas emissions, et cetera.

So, first of all, what correlation would you expect between each paring and why, and which do you think might have a causal link? So for example, you might expect there to be a positive correlation between cycle ownership and cycle use.

So the causal relationship would be that, you know, the more likely you are to own a bicycle, the more likely you are to use it.

So the causal relationship would run this way like that.

That could also be, you could also have more than two variables in a chain, so you could start off by saying more cycle ownership leads to more cycle use, but that, in turn, could lead to something to do with car use.

So feel free to have a think here about the relationship between these five pairings.

Okay, great, so hope that you wrote some interesting things down here, and we can just talk about this for a second.

There's lots that we could talk about here, because it's a very interesting topic, but typically, if you think about it, people have a choice about how they get around.

So they could use a car, they could use a bicycle.

And so if you own a bicycle, you're more likely to to use it, as we've said.

And then also if you use your bicycle more, this is likely to lead to your personal car use going down.

And if your personal car use goes down, then we know that when you use a car, that is a source of greenhouse gas emissions, so greenhouse gas emissions could go down, but not necessarily, of course, you know, there might be a causal relationship between the amount that you use your car and greenhouse gas emissions, but just because there might be a causal relationship, it doesn't necessarily mean that there's going to be correlation because, let's say, for example, that fewer people use their cars to get around, but at the same time, I don't know, more people are turning on their electricity or their heating, that would lead to more greenhouse gas emissions.

So there's lots of things that affect greenhouse gas emissions, not just car use.

So anyway, that is it for today's lesson.

Hope you've enjoyed it.

It's a really important one to think about the link between correlation and causation.

So the message really is that when you see correlation, you have to think really carefully about what is causing this correlation, because just because there is a correlation, doesn't mean that one variable is causing the other variable to change as well.

So anyway, thanks very much for watching, that is it for today, and I hope you enjoyed it.

Have a great day and see you next time.

Bye-bye.