[Auto-generated transcript. Edits may have been applied for clarity.]
So in this video I'll walk you through how to build a scatterplot in R and how to conduct a correlation analysis and interpret the results.

First things first. As always, we need to tell R which libraries to load or which packages to use.

Now we as before we use the library function for this.

And previously we have used the here package and the tidy first package.

So those are the ones that we've been loading so far. Um to be able to do correlation analysis we also need to load the easy stats package.

So easy stats is really a set of packages that handles many different types of statistical analyses.

Um, and it includes the function that we need to run correlation analysis.

So as always what we do is we um type library and then the name of the package in brackets.

After that, we need to read in the data.

I won't dwell on that too long either, because we've done that before, and you can watch other videos to remind you of how to do that.

But very briefly we use the read the read underscore CSV function and then um, yeah,

we need to tell our where to look here and what the name of the data file is, including the extension here.

And this time we are reading in a data file that is called data underscore BMW three underscore subscales.

Now as we've been talking about it's really important to visualise your data.

Um first or and to do visualise an association between two variables.

A scatter plot is a really good way to do it. Um, so far we've been using ggplot in combination with geom histogram to create a histogram.

Um, to build a scatter plot we again use ggplot, and we tell it what the data is that we want to use.

And then we tell it what the mapping is. Now for a histogram you only have to specify what's needed to go on on the x axis.

So that is the axis that goes from left to right.

Um for a scatter plot you need to tell R what to put on the x axis, but also what to put on the y axis.

So this is in this scatter plot we are visualising the association between two variables um present in this data set.

Um one is BMW three underscore un for unintentional mind wandering.

So kind of combined score for unintentional mind wandering is on the x axis and on the y axis.

We use the combined score of meta awareness of mind wandering.

Um, as I said previously, we've used geom point to create a histogram.

This time we want to use this to create a scatter plot. So we need to use a, a different a geometrical object.

And we use geom point for that. So make sure to add a plus sign at the end of this first line and then geom point.

And this is what that looks like. So one variable BMW three underscore un or unintentional mind wandering on the x axis.

And BMW three underscore m e or meta awareness of mind wandering on the y axis from the bottom to the top.

Now, each data point that you see here is, um, kind of indicates the responses of one participant on both of these measures.

So for instance, this one here, um, scored two on, uh, kind of an average of two on the questions that measured unintentional mind wandering.

And it's they scored well, about 0.25 on average on the questions of on meta awareness of mind wandering.

So that's why they end up in this position. Um, if we take another kind of extreme case,

then we have somebody who scored four on average on the questions on unintentional

mind wandering and also four on questions on meta awareness of mind wandering.

Um, so that's that particular participant. So each of these data points represents the responses of one participant.

Um, now you can kind of see a bit of a pattern in this in these data.

So I have a think about what you think this what kind of of association or what kind of correlation this might turn out to be.

Um, there's kind of a clear lack of data points in this bottom left corner.

So it is kind of. Yeah, a concentration over over here.

Right. This is where most of them are now to visualise kind of a trend in the data.

Um, you can add a line of best fit.

And we do that by adding a layer. So we add a plus sign here on the, on the line of geom point.

So we basically say plot the data points using the geom point.

And then and as in plus add another layer where you add a line of best fit.

And we use GM smooth um to add a line of best fit.

And you can use different types of line of best fit. And we use the method that we use is called LM, which stands for linear model,

which basically means that you get a straight line rather than a curvy one or something else.

So we specify that and then we say standard error.

S stands for standard error around a line and we say true.

Um, and that basically means that it um, plots the um, standard error around the line of best fit on the plot as well.

So this is what that looks like. So the blue line here is the line of best fit.

And this grey band is the standard error.

So if you would say in the code standard errors equals false or SSE equals false instead of true, we wouldn't get this grey band.

So this line of best fit gives you an idea of the trend in the data.

And we can see that it kind of starts higher at the top left and then ends lower at kind of towards the bottom right.

So kind of indicative of a negative association between these two variables.

So if people score higher on unintentional mind wandering they score lower on, um, meta awareness of mind wandering.

Now let's tidy that up a bit. Let's tidy up that scatterplot a bit.

Um, to make it more informative and easier to read.

Um, I have changed the background by using a different theme, and I have added, um, informative labels to the x axis and the y axis.

Um, revisit one of the other How-To videos if you want to kind of change more of those features.

In for instance, the colour or something else. So this is what that looks like.

Uh, I don't like that grey background. So I've created the white background and I've added unintentional mind wandering at the bottom here,

and metta awareness of mind wandering for the y axis title so that it is easier to understand for the reader who might not know what you know.

The the previous variable names that we had here, um, refer to.

Okay. Then the correlation analysis. So we use the function correlation to tell to to do the correlation analysis.

And there are four bits of information that you need to kind of specify within that function.

Um, to to tell exactly how you want the correlation analysis to be conducted.

We need to tell it what data to use. So this refers to the data frame.

Then we need to tell it which two variables to correlate.

So again we've specified here um unintentional mind wandering and better awareness of mind wandering.

So those are the two variables between which we want to calculate a pairwise correlation.

And then there are different methods to use to to calculate the correlation.

The default is Pearson. I've specified it here so that I can show you where to do that.

Because sometimes you want to use a different method. And we'll talk about that in later weeks.

So the default is Pearson. And.

Then there is a additional um argument that you need that, that, that you can specify or that you sometimes need to specify.

And we'll talk about that in a couple of weeks time. So don't worry about that for now.

Um, at the moment we basically don't want to adjust the p value okay.

So remember this is a function from the correlation from the easy stats um package.

So make sure that that that you have loaded that package. Um and then yeah this is the code to run a correlation analysis.

And when we run that this is what we get. This is the output.

So you will see that in your console or also in your quarter document.

And it tells you um a number of things.

It tells you what methods, what method you used. It tells you which two variables you correlate it.

And then importantly here we have the R value.

So this is Pearson's r. And here it says -0.32.

So that is the value for Pearson's are. And there is two key bits of information.

It is it specifies whether that is positive or negative.

Um and it is negative in this case and the strength of the correlation.

So how far this is removed from zero. So in this case that is 0.32.

It gives you some other information as well. Um it gives you the 90% 95% confidence interval and the t value and a p value.

We'll talk about those in a few weeks time. Now then, it reminds you that you have not used an adjustment method, which is fine in this case.

And it tells you how many observations there were.

And we will that is relevant information as well, because you do need to, um, kind of report that or a number related to it.

Um, this is the degrees of freedom. Um, sorry.

You derive the degrees of freedom from that. The degrees of freedom is here.

Um, but for now, I want you to focus on these kind of key bits of information,

the two variables that you have calculated a correlation between and this R value.

So in this case that was -0.32. And if we go back to the scatter plot then then that makes sense right.

We did see that the line is kind of is indicative of a negative correlation, because it is higher on the left than it is on the right.

Um, and 0.32 also suggests that it is not a particularly strong association or correlation,

because quite a few data points are relatively far removed from the line of best fit.

Um, but yeah, it does suggest that if people who score higher an unintentional mind wandering score score lower on matter awareness of mind wandering,

and in a way that that makes sense, right?

Because if you would score high on meta awareness of mind wandering, you would notice quite quickly when your mind wanders.

Um, an unintentional mind wandering kind of happens spontaneously where you don't notice that very much.

So that kind of makes conceptual sense as well.

Okay, that's it for now.