Chapter 3 – Part 2

Numerical Measures for

© 2022 Wayne Horn (excluding images)

2
The Dependent Variable and Independent Variable
▪ We are usually interested in investigating the relationship between
two quantitative variables because we wish to predict the value of
one variable based on the value of the other variable.
▪ Examples:
▪ A real estate agent wants to predict the selling price of a house
based on the number of bedrooms.
▪ A personal trainer wants to predict the number of calories
somebody burns based on the total time spent exercising.
▪ A student wants to predict their final exam grade for a course
based on their midterm exam grade.

3
The Dependent Variable and Independent Variable
▪ The variable whose values we want to predict is referred to as the
dependent variable and is denoted by Y.
▪ The variable on which we base our predictions is referred to as the
independent variable and is denoted by X.
▪ As we saw earlier, a scatterplot of the observed (x, y) values will
show us the nature and strength of the relationship between X and Y.
▪ In this chapter, we will focus our attention on variables that have an
approximately linear relationship.

4
Population Covariance
▪ The population covariance of two variables X and Y, denoted by  XY ,
is a measure of the linear dependence between X and Y.
▪ Covariance can assume any real value.
▪ If the covariance is positive, then as one variable changes the other
variable tends to move in the same direction.
▪ If the covariance is negative, then as one variable changes the other
variable tends to move in the opposite direction.

5
Sample Covariance
▪ In order to estimate the population covariance, we take a random
sample of n pairs of values

(x1, y1), (x2, y2), …, (xn, yn)

and compute the sample covariance sxy as follows:

 n  n 
n x i y i 
 i =1   i =1 
n

 ( x i − x )( y i − y )  x y
i i −
n
s xy = i =1
= i =1
n −1 n −1
6
Exercise 1
In the undergraduate statistics project Beers BAC Beers BAC
at Ohio State University, the relationship 5 0.100 3 0.020
between blood alcohol content (BAC)
2 0.030 5 0.050
and the number of beers consumed
appear to have an approximately 9 0.190 4 0.070
linear relationship. 8 0.120 6 0.100
Compute and interpret the sample 3 0.040 5 0.085
covariance between BAC and the 7 0.095 7 0.090
number of beers consumed. 3 0.070 1 0.010
5 0.060 4 0.050

7
Exercise 1

8
 0.20

0.15
xi − x
(x i , y i )

yi −y
BAC

0.10

y = 0.07375

0.05

x = 4.8125
0.00
0 1 2 3 4 5 6 7 8 9
Beers

9
Population Correlation Coefficient
▪ The population correlation coefficient of two random variables X and
Y, denoted by  , is computed as
 xy
= .
 x y
▪ The correlation coefficient is a unitless measure of the direction and
strength of the linear dependence between X and Y.

10
Population Correlation Coefficient
▪ The population correlation coefficient always equals a value between
–1 and 1, inclusive.

–1 0 +1

perfect negative no linear perfect positive

11
Sample Correlation Coefficient
▪ In order to estimate the population correlation coefficient, we take a
random sample of n pairs of values
(x1, y1), (x2, y2), …, (xn, yn)
and compute the sample correlation coefficient r as follows:

s xy
r= .
s xs y
▪ The sample correlation coefficient always equals a value between –1
and 1, inclusive.

12
Examples of the Sample Correlation Coefficient

16
30.0

14
27.5
12

25.0
10

8 22.5
y

y
6 20.0

4
17.5

2
15.0
0

0 1 2 3 4 5 0 1 2 3 4
x x

r=1 r = –1
perfect positive perfect negative
linear relationship linear relationship
13
Examples of the Sample Correlation Coefficient

35
90

30 80

70
25
y

y
60
20

50

15
40

10 30
0 1 2 3 4 0 1 2 3 4
x x

r = –0.874 r = 0.408
strong negative moderate positive
linear relationship linear relationship
14
Examples of the Sample Correlation Coefficient

60
30
50

40
25
30

20 20

y
y

10

0 15

-10
10
-20

-30
0 1 2 3 4 5 10 15 20 25 30 35
x x

r = 0.165 r = –0.009
weak positive no discernible
linear relationship linear relationship
15
Examples of the Sample Correlation Coefficient

59

58

57

56

55

y
54

53

52

51

50

7 8 9 10 11 12 13
x

r=0
perfect quadratic relationship

16
Exercise 2
In the undergraduate statistics project Beers BAC Beers BAC
at Ohio State University, the relationship 5 0.100 3 0.020
between blood alcohol content (BAC)
2 0.030 5 0.050
and the number of beers consumed
appear to have an approximately 9 0.190 4 0.070
linear relationship. 8 0.120 6 0.100
Compute and interpret the sample 3 0.040 5 0.085
correlation coefficient between BAC 7 0.095 7 0.090
and the number of beers consumed. 3 0.070 1 0.010
5 0.060 4 0.050

17
Exercise 1

18
The Least-Squares Regression Line
▪ If X and Y appear to have an approximately linear relationship, then
we can approximate the relationship using the equation of a line
Y = a + bX .
▪ The value a is the y-intercept. It is the value of Y when X= 0.
▪ The value b is the slope. It is the amount by which Y will change if we
increase X by 1 unit.
▪ The best-fitting line relating Y to X is called the least-squares
regression line.

19
The Least-Squares Regression Line
▪ The least-squares regression line is found by minimizing the sum of
the squared vertical differences between the data points and the line.
y
. .
. . Y = a + bX

. . .
. . .
x
20
The Least-Squares Regression Line
▪ The least-square regression line is Y = a + bX , where

sy
b=r and a = y − bx .
sx

▪ Note: Since sy and sx are both positive, b and r have the same sign.

21
Exercise 3
In the undergraduate statistics project Beers BAC Beers BAC
at Ohio State University, the relationship 5 0.100 3 0.020
between blood alcohol content (BAC)
2 0.030 5 0.050
and the number of beers consumed
appear to have an approximately 9 0.190 4 0.070
linear relationship. 8 0.120 6 0.100
(a) Find the least-squares regression line 3 0.040 5 0.085
for predicting BAC based on the number 7 0.095 7 0.090
of beers consumed. 3 0.070 1 0.010
(b) Predict the BAC for a student who 5 0.060 4 0.050
consumed 5 beers.
22
Exercise 1

23
Exercise 3
Fitted Line Plot
BAC = - 0.01270 + 0.01796 Beers
0.20

0.15
BAC

0.10

0.05

0.00
0 1 2 3 4 5 6 7 8 9
Beers

24