And I will post a link to the

recording onto Moodle later on.

OK, so that's step one.

People keep everything seems to everybody

seems to be joining as a guest,

which is really confusing,

but then they end up in the lobby

and I have to manually admit him.

So that's those beeps that you hear.

Uh, and it might slightly distract me.

There we go, another one.

Um, right.

So you can ask questions if you

want to anonymously via slider.

The instructions for that

are on the slide as well.

So just go to that web page and put

in that event code and then you can

ask a question and I will kind of go

through those in in order so to say.

Um, I checked this morning.

There were no questions on

the discussion forum yet,

so otherwise I would.

I might, you know,

I might have come back to some

of those here as well.

Uh, if you are comfortable in just unmuting

yourself and asking a question verbally,

that is of course fine too.

All right,

Now let's um get started.

Ohh yeah,

the session codes on the slide as well.

Alright,

I will also put that into

chat because people often are

looking for that information.

I have a suspicion that the

way this meeting was set up

is not ideal anyway. Um,

ohh. Team has done something new where

they just add everybody to the lobby.

Some of my space to do this

right. We're just going to start.

I will stop sharing. Uh well

stop sharing this slide And then

we will go to the we'll we'll get started.

Basically the chats. I wanted to put

the session code in there which is.

OK, so now I need that

information is also in the chat.

Hopefully that is all clear all right now.

Of course, outside of this Q&A,

you can also ask questions on the Moodle

forum or on during your lab session.

The answers to the questions in the

lab activities etcetera are now all on

the website and kind of at the bottom.

So I will show you in a minute.

And and there isn't a quiz,

an unmarked quiz on on rudder.

So I'll link to that somewhere you can

check that you've understood it all. Um,

right. Goodness me. If this goes on,

it's just really distracting.

Goodness. Ohh,

Nope. Right. So let's see.

We'll start with Slider and see

whether there are any questions.

Are there not at the moment.

So please, if you have a question,

doesn't matter what it is,

you can ask it anonymously on Slider.

You know, just get your questions

in and we'll work through them.

In the meantime,

I will show you the website.

I mean you know the website

probably but and where is it

that slide out, but the website

is that's the link is here.

So in the week 11 materials and the

answers you will now find answers

to each of the of the questions.

So you can just kind of unfold the hair? Um,

I hope that is helpful. No.

Let's see Slide or do we?

OK, first question is how do you

know whether to report a P value

as P is smaller than .001 or P is

smaller than .05 or something else?

OK, that's a really good question. So

if so, before we get back in the day when

you did calculations by hand and you looked

things up in a table of significance,

and I did go through that in the lecture.

And then you would look in the right

column and then see whether your team

value would be closer to the column for

PS smaller than .05 for PS smaller than

pointer one or PS more than point O1, um.

So that would kind of guide

what you would be reporting.

But nowadays where we use software

like R Studio to calculate the P value,

the guidelines,

the APA guidelines that we used for

and reporting things in psychology,

they ask you to report exact

P values where you can.

So the P value comes out of the

analysis in R Studio and that

is the P value that you report.

There's one exception and that is

when MP is smaller than .01, so it's.

So that's highly significant of course

because P if P is smaller than .05,

we consider it significant.

But if P is almost zero, so you know,

sometimes it's even reported as a

number that in scientific notation,

which might have a wall and an E

to the minus something which means

there is a lot of zeros,

then you report it SP point smaller than .01.

In all other cases,

you would say P equals and then you go to.

You report the number that comes

out of the analysis.

Um. So I don't.

Let's see whether I can easily show you um.

So for instance,

no,

I would have to go to the

our markdown script.

Hmm

ohh

so hey. In this example the

P value was smaller than .001

so that is why it says that.

So that's liability 2, lab activity 3.

Um, it is not. So it is.

And then we report the exact value.

So, um, let's see how we're doing on sliders.

OK, We have a little bit of time to just go.

I hope that answers your

question wherever asked it,

but I'll just show you where

to look up that value, right?

So then I will let's do this differently.

Let's stop sharing that and I will

start sharing my whole screen and

then we can navigate it more easily

between different bits and pieces.

So, um, if let's see what I

can just run this code on

and

reading the data

and

O, it's reading the data and so

this is what you usually look for,

right? Let's do a bigger plotting

and then we'll do the analysis

and I'll show you where.

So when you conduct a correlation

in that it says using the core dot

test function, it will spit out.

Well it will split and you then pipe

the output of that function to some

different function will tidy and this

is what the output will look like.

So there you have here you have the P

value and you can see that this has a

lot of zeros before any numbers appear.

So this is smaller than .1,

which is why you say P smaller than .01 in.

In the other in this third line activity,

P was .014, so this is where

to to look for that.

So here it reports that P value.

Can you actually see my cursor?

And I hope that clarifies things.

Let's go to practise slider.

Will we just using percents or

also experiments rank correlation?

We will be looking at experiments rank

correlation as well in that kind of

the upcoming weeks materials, yes.

So we'll be using both and you

will learn be learning about both.

And there will be questions on the class

test about what it could be questions.

But on the class tests about either

when will the content for

next week be available,

that is a very legitimate question.

I was not very well on Tuesday and Wednesday,

which is why it's been a little bit delayed.

I will make it available today.

So there's just a few things

that I need to to amend and

then I will make that available.

So apologies for that.

Yeah, because I wasn't very well.

I didn't have the time to make

those amendments and get it online

earlier this week in following weeks.

That should it should come

and be available earlier.

OK, so that is those questions. Um,

let's see whether anybody has

asked a question on the forum.

So while I'm on the forum,

you can see that anyway. Um,

here is a guide with, um,

some information on what the

APA expects us to report when it

comes to numbers and statistics.

So if you're not sure how to write

write up the results of an analysis,

this is a helpful guide.

This is a summary of the various

statistical formulas that you encounter.

And here is the Week 11 quiz.

So that is what you can use as well.

No, no questions on the forum.

No questions on slide out.

Let's see whether there's

anything additional in the chat.

So the session code is 318117.

If you want to ask a

question on Slido Um, it is

449-8100. So yeah, ah, someone else

hopefully pointed out there as well.

Thank you. Um, right. So if you know,

feel feel free to to to answer

your questions if you've got any.

If not, we can.

I'll talk maybe through lab activity 3 and

because I don't know whether everybody

had the time during the lab session to,

um, yeah, work your way through that.

Before I do that, I'm actually talk,

I'll talk about these issues with

uploading things. So

I've added some guidance to the

website because in particular in the

Monday session there were quite a

few issues with uploading your files

to the server and then the server.

You know you thought you'd

done it and it wasn't feasible.

So if that happens,

here are a few things you can do.

You can close on the R Studio server,

close your browser and start afresh,

and you can open the Our Studio

server in a different browser.

So if you've been using Chrome,

try Edge if you're on a Windows machine,

or if you're on a Mac,

and try Safari or something.

Um, it it it's basically a browser

issue and it's very unpredictable

so it's it's difficult to have

a have a clear fix for it.

But try a different browser.

If that doesn't work,

then there is a kind of a workaround

where you use a bit of code to download

the file from the Internet directly

to your folder on the server and

that code I've kind of added that to

where you needed in the lab activity.

So if you go to lab activity 2

No2 here. So if you have

issues with uploading files,

you can use this code which you can copy

just by clicking on this little clipboard.

Here you don't have to type down the line.

You can't even see it all because

it's too long and so you copy that,

put that in your script,

run that line of code,

and then you will see the data

file in your working directory.

So hopefully that will, um, you know,

prevent people losing too much time

with trying to upload various files.

So that was the first thing I wanted

to kind of bring to your attention.

And then I thought,

maybe let's walk through

lab Activity 3 together.

Just check again when there's

any other questions.

No. Um, And we can,

you know,

I can talk you through the different

bits and pieces of the code there,

so OK. Um, umm.

This is about hazardous

alcohol use and impulsivity.

It's a real study published.

Well, some published somewhere,

but these are kind of the

measures that are being used.

A questionnaire about alcohol use and

A questionnaire about impulsivity.

20 participants, and yeah, that's the

date that we're going to work with.

So what this assumes is that you have,

uh, a markdown script open.

Um, you know, you might.

If you use one markdown script

per lab session, for instance,

then you should have already

loaded the the right libraries.

So if if it's all new,

then obviously of course you would

need to add code to to do that,

and and if it's a new session,

you would need to rerun that code

chunk before you do anything else,

because otherwise our doesn't

know which functions to use.

So just to maybe tell you a little bit

about what on earth these libraries are,

and they're basically a collection

of code snippets, you could say,

of functions.

So a function in when you when

you talk about programming is

a set of instructions in code

to do a particular thing,

like reading a file or download a file.

So if or make a histogram or whatever.

And so somebody has written

a a function to do something,

and they put them in these little

groups of uh, collections of of,

of, of, of functions.

And they are called libraries.

And ohh has a lot of different libraries,

always open source,

so anybody can contribute to it,

which nobody has to pay for it,

which makes it very powerful.

And you know,

you never have to pay for a

licence or anything like that.

Um,

but it also means that you know

there will be that there's lots of

different libraries that people have

written to do very specific things.

And so at the start of your session,

you always have to tell our

what libraries to use so that it

knows basically where to look

for the functions that you want

to use in that session.

So here we use brew and a

library called tidyverse

and I will just periodically check slider Ah,

what does broom the broom library do

and how is it different to tiny verse?

So basically they have different sets

of functions in them and and the

tiny verse is so so if you want to

know more about any of these things,

um, if we say here, let's try this.

Does it work for? This works with functions.

So if we want to know more about

the broom package,

so I've in the console I've typed

questionmark and then broom package

in single quotation mark and if I um

done press enter it will give me the

help file for that particular library.

So broom is all to do.

So here you can see who's written

it and it is all to do with

how to work well with tables.

So it just does a lot of so to Tidy for

instance comes from the broom library.

Tidy is a function that makes the

output from core tasks look like

this and if we run this code. Um,

without tidy? So if I just

delete that I want it.

This is the output that you get.

So now I just just used core tax

that's the same function to do

the actual correlation analysis.

But this is what the output looks

like and that's kind of how to

read and kind of hard to use.

So somebody thought that was a

bit difficult to use and said OK,

I'm going to write a function to tie

that up and and then if we rerun it,

it makes it looks like this.

So and broom has, you know,

has a kind of a lot of other

functions like that where you know,

there's a lot of contributors

to broom as you can see.

And if you go all the way down here,

you know you can go to the

website that explains more about

this particular package,

and if you discover any bugs in it,

there's also a place to report them.

And Tiny First is a,

so we can do the same thing for Tiny Verse

and

see where it goes. This. Yeah.

Um, so that is actually a bit

unusual as a library because it is

actually a collection of libraries.

This name Hadley Wickham.

He's one of the most important

developers or programmers of our R&R

studio and he's uh really kind of

made some quite innovative changes

in how we use our and our studio

for statistics and tiny vertical

cysts of of many different things.

So if we go to this website,

it has all of this on ggplot 2.

For instance that you know it's a

library that has the GD plot function

in which we use for plotting things.

And it has deep liar in which we use

a lot to make to make new variables,

select variables from a from

a data frame to select rows,

uh to philtre for for for rows, etc.

So there is lot and there is more.

So it is a whole collection of our

packages and they will come back and back

again and it's a really important one.

Basically,

it's one of the main things that we use.

Um yeah,

this is a really good book by the way,

if you want to dive in deeper.

So I hope that answers that question

in the lab activities when it asks us

to calculate how much variance can be

come from and what does that show us?

OK. That is also a very good

question and let's go to that.

Uh, what is the best way to illustrate it?

Maybe we'll go to.

Let's see when we can use this to shut up.

So this was the one of the pre lamp

activities and victory here and it's a

bit all about visualising correlations.

So here we have two measures X&Y and

we can change this number here the

correlation and see what it looks like.

So if we increase this,

let's say we will put it to.

This is a bit too slow.

Can I just do it like that here?

So here we have a correlation of .5.

You can see it's because it is a positive

correlation and if one variable goes up,

the other variable goes up.

So we get a scatter plot like this

and you can see there are some some

sort of relationship between the

two variables even though it's not,

uh perfect right?

If we would say .8 or something,

then the dots fall much closer to the line.

Now what? So what is this shared variance?

It is basically a measure of

effect size and you might have

heard of measures of effect size.

When you talk about

tests to do at compare group

means like a team test in.

You did that in 1-2 one

and so if if our is .5 we can calculate the

the shared variance by squaring that number.

So if you literally square R you get

R-squared and that is .25 in this case.

If you multiply that by 100,

you get the shared variance and it

basically means that the variation in.

Why? So you know the variation in

Y is how much each data point is

removed from this line of best fit.

So you know, here it is a little bit of buff.

This one is a little bit below,

This one is quite a lot of buff.

This is even further above right?

So this variance.

So these these individual differences

between these data points on why?

Well, by calculating R-squared we

can say that 25% of this variation

is explained by variation in X.

So we also have variation in

X and we've got calculated the

correlation between the two.

Once we've calculated the correlation,

from that we can derive R squares,

also called to the

coefficient of determination,

and that tells us that 25% of the

kind of the individual differences

between all these data points.

Uh, why?

Is explained by or is accounted

for by changes in X?

Does that make sense? Maybe you can.

Whoever asked that question can put

up a follow up comment with yes or no

and

so that that is kind of what that

what that means and and if we if we

make it higher you know if you have

a correlation of .9 we can see that

81% of of the variance basically

between these two variables is

shared or is what you know fair

variance in Y is accounted for by X.

You know in a correlation it's you

could choose to put if this was weight

and that was height and you know you

could choose to put to put weight

here and height on the on the X axis.

And so it's it's kind of about

shared variance rather than one

predicting the other in in regression

that is a little bit different,

right? I hope that helps. Thank you.

Thank you for confirming UM.

So let's go back to our uh ask right?

Uh, we were trying to walk through.

Um, no. I was going to do that.

We were going to walk through

a lot of Activity 3, right?

So let's do that. Um,

there we are. Lovely activity.

Normal pre enough activity. Lab activity 3,

right, So boom and tiny.

First we talked about that step one,

reading the data,

and you should now see an object

containing the data in the environment.

OK, so let's just check that

we have actually done that.

So I'm just going to clean this

environment by using this little broom

on the on here in the top right.

That just means that it's.

I tend to do that between, you know,

between two different analysis,

because then you know we don't

get confused by things that might

have similar names or something.

Um, so we want to first do reading the data.

We use read underscore CSV to do that.

So read underscore CSV is a function.

We tell that function that we

want to use this data file and

we assign the output of that to

a data frame called data.

Or you could have called it alcohol,

use slash impulsivity or a

Lim or whatever you could have given it.

Or data lab activity three. And the name

of that data frame is is up to you.

It is good actually to to use

sensible names for it, and Data is

actually not that much of A sensible.

But there, you know,

I really should have been a little

bit more specific with that.

So in the environment you now see the

data file which I've called data.

Uh. If you click on this little

blue circle with the white triangle,

you can have a look at what what's in there

just here without having to kind of you know.

You can also click on it of course

and then you will see it here

um in the in this in this pain.

So we have 3 columns,

participants as it is alcohol,

use and impulsivity,

and we have 20 participants.

Always worth checking that what

you expect the data to look

like is actually in the file.

You know,

if it says that there should

be 20 participants,

you should be able to locate 20 participants.

So that's um because I this is

run in a code chunk and it doesn't

say so if I would put echo as

false here and here it wouldn't.

It wouldn't give me these kind of bits here.

So if I rerun this, that goes away.

If you find that quite useful

to actually see those things,

you leave that out.

And then and there you go, there they are.

And so you can connect,

you kind of have a preview of it in

your our marketing document as well.

So lots of different ways to

look at the same thing.

Now did ask us any questions about

that How many variables does it have?

So the hint was ohh yeah you need to do that.

What we did that that's the code.

How many variables does it have?

It has three variables right?

We we concluded that here 1 variable,

2 variable, 3rd variable SO3 variables.

Now then, you're asked to plot

a relationship between hazard,

alcohol use,

and impulsivity using its

catapult and a line of best fit.

No, this is how we do that.

We use G+,

so that's the function that we need.

And Gigi plot requires certain information,

otherwise it can't plot anything.

So it needs to you know what

data we're using,

so that's what we put there.

And then it needs to know what we put on

the Y axis and what we put on the X axis.

So that's what happens here.

Now ggplot is a little bit

different from other functions,

but what it does is that it

builds a figure layer by layer.

So first you tell it what data to use.

You do that here,

then you put a plus sign and then you say OK,

please give me the data points.

So if we only run

this bit of code,

um, would you like to just run that code?

Should ensure run selected lines.

I should run that without that plus,

otherwise it will be waiting for more

inputs. And let's try that again

one selected lines.

Umm, just something that it doesn't like.

Why does it not look?

Well, let's run the whole thing.

I wanted to show you something specific,

but it doesn't quite like to work with that.

So if I only run this bit,

for some reason it gives an error,

but it shouldn't really.

And then I would only get the black line,

the black dots and in this line of code.

So we give, we set,

we add another plus and then say

please add a line of best fit and

for that we need do you have smooth?

There's different methods to do it,

but we also always use Allan linear model.

We don't want a standard error, sorry.

We could actually add a step in that

area and show you what that looks like.

Thank you that and we run it.

We get this grey bar and it kind of adds

a standard error around the best of it

that becomes more important

with more complex data.

Basically so it adds the points,

then it adds a line and then I've

asked it to change the thing.

There's lots of different things in

in duty plots to make the figures

look exactly like you wanted to.

Basically you can explore that

and we will be using, you know,

different ones across the module.

And then I ask it to use to add kind of

different labels for the X axis and Y axis,

and I know what it was.

So I know that impulsivity here

because I just talked with is is there.

Now let's just check whether anymore

questions have come in. No. Um.

So hopefully that is all clear.

What can you tell from Southport about

the direction of the relationship?

So we can tell that there is a positive

relationship because of the line goes

up from startup kind of the bottom

left and goes up to the top right.

And so there's there is a positive

association between hazardous

alcohol use and impulsivity.

And that means that as a participant

score or has this Elka use goes up,

their score impulsivity also goes up.

That is what we can tell

from the scatterplot.

You can actually tell more

even from scatterplot.

You can tell that there is somebody

who schools really high on impulsivity

and and somebody else who scores

really low on impulsivity, right.

Somebody.

Some people score really low on hazardous

alcohol use and others score quite

high normally hazardous alcohol use.

So there's other things that you can tell

that the scatter plot is really useful for.

And so,

So if you know,

let's say that somebody would

have an impulsivity score of 900,

then that is clearly a data entry mistake,

right?

Because that looks like not a possible score.

So that's something else that

you can see in the scatter plot.

And there's other stuff that

we will look at next week.

So then in step three, you're asked

to conduct a correlation analysis.

So for that we use and the the the

function called core dot test and in

core dot test we need to tell it what

is the the data and the and the X

variable and what is the Y variable.

What method won't do you want to use.

So here we use person and whether you

want it to be 1 sided or two sided.

There's very few cases where you

really want one sided test if you.

So the difference between A1 sided

and A2 sided correlation test

is if you are you know,

I guess you based on previous

research you say,

well I know that this correlation is

going to be positive and I only wanted

to test whether it is significantly.

You know whether it is significantly

different from zero,

but I'm never going to expect

a negative correlation.

Then you could do A1 sided test or

the other way around of course.

So you could say OK,

I'm gonna test whether it is and

I I from all previous research

suggests that we were expecting

a negative correlation here.

So I'm going to only test where

it whether it is negative enough

to be significant,

then that's also a possibility.

But you can't.

With one side attached,

you can't test both directions in one go.

Basically so 99% of the cases

you're just wanted two sided tests,

so that's what's what is and specify to him.

So what we do here?

So here we have the the function name,

then all the information that it needs,

and then we assign that to

something called results and that

will appear in our environment.

Once we've run that and as

I showed you before,

we take the output from this function

and put it through another function

called type from the broom package just

to make the results eating more easily,

more more easy to read.

So if we run this I've actually

done a lot of stuff in one go but

so here we've got our results.

So here you've this is just kind of,

you know,

we could click on this and it would open it.

Here,

um,

it's just a table with the results where

you estimate is the correlation coefficient,

the statistic is the T statistic for

that correlation coefficient that

you that we don't typically report,

but that's what it is.

Here we have the P value parameter is

the degrees of freedom and he would

remind you what kind of test you've

done and whether it was 1 sided or two sided.

And this is the correlation,

sorry,

the confidence interval

around the correlation.

So this isn't a 90 percent,

95% confidence interval around this number.

So you know,

we it is highly likely

that based on these data,

the correlation coefficient in full

somewhere between points 13 and .79.

So sometimes confidence in force around

these estimates are very informative,

but you don't need to

worry about it right now.

So just to kind of recap,

if you look at output like this estimate,

it's the correlation coefficient.

So R P value is the P value

and you report an exact P

value up to three decimals.

So you would say

P = 0.014 in this case.

That is what the APA wants us to do,

and for the degrees of freedom

you look at the parameter saying.

And so that is that now.

So you can just look at that in the table.

You can also pull it out

and using the pool function.

So here all I've done is said OK,

take the results,

use the pull function and give me

others in the estimate cell and

round down to two decimals and

assign that to something called R

So we've got that here as a value.

So here I can just.

I could, I could.

I could ask for it in the console because

it was our because that's what I've

decided to to .54 and and here it is.

It's. You can see it here,

you can see it here,

and you can see it in the original table.

And same for degrees of freedom.

So I know that that is in the

cell called parameter sizing.

Pull out at #4 me and there you go.

It's 18.

I'm saying from P value and then

you are asked to calculate R ^2

coefficient of determination

which shows you allows you to to

calculate this shared variance

or the variance accounted for by

1 variable in another variable.

So R-squared is literally squaring

of Rs and multiplying R by itself.

So we can just do that like this

and and then ask where percent or

this shared variance is r ^2 * 100.

And then I asked it to round it to

00 decimals so I get that I get kind

of a number like that and not 29.2.

So we typically around that to 0 decimals.

So 29 percent is the answer

here of the variance in and

OHH impulsivity is accounted

for by hazardous alcohol use.

Now let me check sliders

whether there's any slide out.

No, I don't see any additional questions.

And there we go.

So that is has answered them

of these question right?

What is the correlation coefficient?

This number? What is the P value?

That number? Is it significant?

Yes, it is significant.

So you might wonder what this says.

But if I knit this,

it will look sensible

and it's my markdown file.

This symbol here between the team and

the IT should knitted it, whatever.

If it's still busy knitting,

maybe it is a little bit too much

to ask to be sharing my screen,

have a teams meeting,

recording and also asking. It is.

It is doing some work here.

It's going there.

It's knitting the whole thing,

which is why it's taking you off.

So I've known. I've knitted my

ARM markdown from and you can see

all the output here right? And so

this is not what we're doing here.

This is the activity that we

were concerned with. So umm

is, it's. Is the correlation

significant at the level people?

At the smaller point of higher

level, Yes, because it is

.014. And what other degrees of freedom

you need to report, so that is 18.

You can also calculate that. Fine,

just just doing the maths yourself

because that is always and -2.

So and what's 20?

Because we had 20 participants in

this particular data set and you

subtract 2 because you have two

variables in correlation, so it is 18.

How much variance in impulsivity can be

accounted for by hazard alcohol use?

That is this R-squared thing, right?

So that is was 29%.

Now then, we're asking for 3 logical

possible directions of causality,

indicating for each direction whether

it is possible explanation in light

of the variables involved and why.

So you know. Correlation, of course,

does not infer causation, but it is.

It is helpful to start thinking about OK,

well, what does this actually mean?

Why are these two things related?

So here are just three examples.

Being more impulsive may make

people consume more alcohol and

or consuming more alcohol may be

may make people more impulsive.

With the current data,

we have no way of distinguishing

between these options.

A third one could be that an outgoing

personality might influence both your

level of impulsivity and you are more like,

so you're more likely to be socialising

in a pub and consuming alcohol.

So in a way,

personality having an outgoing personality

might be a third factor that influence

both of our variables of interest.

So again, with just the correlation analysis,

we can't tell these difference.

That kind of, um, explanations for

the relationship that we found apart.

We would have to do further experiments

to figure out which one is correct.

Now finally,

how do you report these results?

Following the APA guidelines.

So here is this kind of written up

saying a Pearson correlation coefficient

was used to assess the relationship

between alcohol use and impulsivity.

There was a significant positive correlation

and then we report these numbers here.

Actually this is wrong.

This should be equal.

So let's that's correct, that's.

This should be equal.

There we go ohh,

there we go.

There we go.

So P equals

.014 and APA requires, UM,

statistical symbols like R&P to

be Italian in italics in italics,

SOCOM like that.

And you can do that by placing

them within these asterisks.

So if I need this again,

we should get,

um, we should get that

there we are, P = 0.014.

So you can see R as in

italics and P as in italics.

And the way to do that in your market is

by placing them within these two asterisks.

That's kind of our marketing code,

so let's have to report it.

And then you give it to make it easy

on the reader of your research report,

you would say something like this.

People who reported to consume more alcohol

scored higher on the impressive skill.

Explaining what this correlation means,

right? So that was lab activity 3.

Um, let's see whether

there's any other questions.

So no other questions on slide out,

you don't see any other

questions on in the chat either.

So I think we will leave it at that.

I hope you found this useful,

and materials for next week

will be available later today,

and we'll be doing more work on

correlation and thinking about the kind of

assumptions that you need to be aware of.

Um, looking at how to check those.

We'll be looking at Pearson Rank,

Spearman Rank and Coronation, Sorry.

And when to use that, Yeah.

And we'll be doing also a

little bit of data practising,

data wrangling to to get

the data in the right shape.

So thank you for attending today

and I will see you all next week.

Thank you.