2.02 Pearson's r

Loading...

來自 University of Amsterdam 的課程

统计基础

1707 個評分

University of Amsterdam

统计基础

1707 個評分

課程 3（共 5 門，Specialization 社会科学的方法和统计）

Understanding statistics is essential to understand research in the social and behavioral sciences. In this course you will learn the basics of statistics; not just how to calculate them, but also how to evaluate them. This course will also prepare you for the next course in the specialization - the course Inferential Statistics. In the first part of the course we will discuss methods of descriptive statistics. You will learn what cases and variables are and how you can compute measures of central tendency (mean, median and mode) and dispersion (standard deviation and variance). Next, we discuss how to assess relationships between variables, and we introduce the concepts correlation and regression. The second part of the course is concerned with the basics of probability: calculating probabilities, probability distributions and sampling distributions. You need to know about these things in order to understand how inferential statistics work. The third part of the course consists of an introduction to methods of inferential statistics - methods that help us decide whether the patterns we see in our data are strong enough to draw conclusions about the underlying population we are interested in. We will discuss confidence intervals and significance tests. You will not only learn about all these statistical concepts, you will also be trained to calculate and generate these statistics yourself using freely available statistical software.

從本節課中

Correlation and Regression

In this second module we’ll look at bivariate analyses: studies with two variables. First we’ll introduce the concept of correlation. We’ll investigate contingency tables (when it comes to categorical variables) and scatterplots (regarding quantitative variables). We’ll also learn how to understand and compute one of the most frequently used measures of correlation: Pearson's r. In the next part of the module we’ll introduce the method of OLS regression analysis. We’ll explain how you (or the computer) can find the regression line and how you can describe this line by means of an equation. We’ll show you that you can assess how well the regression line fits your data by means of the so-called r-squared. We conclude the module with a discussion of why you should always be very careful when interpreting the results of a regression analysis.

2.01 Crosstabs and scatterplots7:06

2.02 Pearson's r7:19

與講師見面

Matthijs Rooduijn
Dr.
Department of Political Science
Emiel van Loon
Assistant Professor
Institute for Biodiversity and Ecosystem Dynamics

This scatterplot shows the relationship between chocolate consumption and

body weight.

On the horizontal X axis we see the independent variable chocolate

consumption, measured in consumed grams of chocolate per week.

On the vertical Y axis,

we have the dependent variable body weight, measured in kilograms.

The case to study here are 200 female students,

who are exactly one meter 70 tall.

They are represented by the 200 dots in this graph.

The scatterplot shows at a glance that there is a strong correlation between

the two variables.

The more chocolate someone eats, the larger the body weight.

But how strong is this correlation?

We will now turn to one of the most often used measures of correlation,

the Pearson's r.

One of the most important advantages of the Pearson's r,

is that it expresses the direction and

strength of the linear correlation between two variables with one single number.

The relation between chocolate consumption and

body weight can best be described by this straight line, because all cases

closely around a line we can conclude that this is a rather strong correlation.

Another thing to note is that the line goes up, so

more chocolate consumption is associated with higher body weight.

We can therefore also say that there is a positive correlation.

Conclusion, we have a strong, positive and linear relationship here.

However, variables could also be correlated in different ways.

In this graph, we see a rather strong positive and

linear relationship between the variables x and

y, just like in the example with chocolate consumption and body weight.

But in this graph, we have a rather strong negative linear correlation.

The line goes down that indicates that when variable x goes up,

variable y goes down.

In this graph, we also see a positive linear relationship.

However, it is much less strong than the previous ones.

We know that because the individual cases are much further removed from the line.

This is a perfect negative linear correlation.

It is perfect because all cases lie exactly on the line.

But the correlation between two variables need not be linear.

In this graph, we also see a relationship between the variables x and y.

However, the line that best represents the relationship between the two variables

is not straight.

Instead, it is a U-shaped line.

We call this a curvilinear relationship.

A scatterplot helps us to broadly assess whether a correlation is strong or weak.

But it does not tell us exactly how strong the relationship is.

Pearson's r is a measure that can show us exactly that.

More specifically, the Pearson's r tells us the direction and

exact strength of the linear relationship between two qualitative variables.

A positive Pearson's r indicates that a correlation is positive, and

a negative correlation indicates that it is negative.

The size of R expresses how tightly the observations are clustered around

the imaginary best fitting straight line through the cloud of the data points.

Pearson's r is always a number between -1 and 1.

Minus one refers to a perfect negative correlation, and

plus one to a perfect positive correlation.

And zero means that there is no correlation at all.

So, how do we compute a Pearson's r?

Imagine that the study on chocolate consumption and

body weight was not based on 200, but on only four individuals.

This is a data matrix and this is a scatterplot.

You can see that every combination of values on the two variables becomes

a circle in the graph.

This woman consumes 200 grams of chocolate per week and weighs 70 kilograms.

She's represented by this circle.

The other three circles represent the other individuals you see in

the data matrix.

To compute a Pearson's r we need this formula.

What does it mean?

Well, first we change all original scores to z-scores.

In other words, we standardize the values.

The reason is that we want the Pearson's r to be a number between minus one and one.

If we don't standardize, the measure of correlation will be expressed according to

the original metrics.

So, first, we compute the mean for both variables.

This results in the value 162.5 for variable x, that's chocolate consumption,

and 71.25 for variable y, that's body weight.

Then we compute the standard deviations for both variables.

The results are 110.9 for x and 18.4 for y.

We the get the z-scores by applying this formula to every case.

We subtract the mean from every value, and then divide it by the standard deviation.

So, 50-162.5/110.9=-1.01.

We do that for

every value of the independent variable, chocolate consumption.

And for every value of the dependent variable, body weight.

In the next step, we compute the products of every z-score on x for

every z-score on y.

So for the first case, that's -1.01, times -1.15 equals 1.17.

These are the results for the other three cases.

To get this part of the formula, we add up all these scores, that's 2.78.

To finish, we have to divide it by n minus 1.

N = 4, so in our case, n - 1 means 4 - 1, that's 3.

The Pearson's r is 2.78 divided by 3, that equals 0.93.

What does this mean?

It means that there is a strong positive linear relationship

between chocolate consumption and body weight.

One important note,

though, you can always compute a Pearson's r, even if the relationship is not linear.

Therefore, it is very important that before you compute a Pearson's r, you can

always check the scatterplot first to see if your variables are linearly related.

If not, do not compute a Pearson's r,

because it doesn't tell you much about relationship between your variables.

This scatterplot, for

instance, shows that there's a strong curvilinear relationship between x and y.

If you compute the Pearson's r you will get a very low value, -0.15.

This doesn't tell you that there is a weak correlation,

it only tells you that there is a weak linear correlation.

It is rather easy to compute a Pearson's r with only four cases.

However, you can probably imagine that it becomes an almost

impossible task when you have, say, 200 cases.

Luckily, every statistical computer program

can compute a Pearson r in no time.

Nevertheless, it is important for

you to understand what the Pearson r exactly tells you.

And it's also important to understand what the formula means.

It helps you to better understand how variables are related.

And it might help you to decide for

yourself how much chocolate to eat every week.

探索我們的目錄

免費加入並獲得個性化推薦、更新和優惠。

Coursera 致力於普及全世界最好的教育，它與全球一流大學和機構合作提供在線課程。

© 2018 Coursera Inc. 保留所有權利。

通過 App Store 下載

通過 Google Play 獲取