2. • The Pearson Product Moment Correlation is the most
widely used statistic when determining the
relationship between two variables that are
continuous.
3. • The Pearson Product Moment Correlation is the most
widely used statistic when determining the
relationship between two variables that are
continuous.
Variable A Variable B
4. • By continuous we mean a variable that can take any
valuable between two points.
5. • By continuous we mean a variable that can take any
valuable between two points.
• Here is an example:
6. • By continuous we mean a variable that can take any
valuable between two points.
• Here is an example:
Suppose the fire department mandates that all fire fighters must weigh
between 150 and 250 pounds. The weight of a fire fighter would be an
example of a continuous variable; since a fire fighter's weight could
take on any value between 150 and 250 pounds.
7. • By continuous we mean a variable that can take any
valuable between two points.
• Here is an example:
Suppose the fire department mandates that all fire fighters must weigh
between 150 and 250 pounds. The weight of a fire fighter would be an
example of a continuous variable; since a fire fighter's weight could
take on any value between 150 and 250 pounds.
8. • The Pearson Product Moment Correlation will either
indicate a strong relationship
9. • The Pearson Product Moment Correlation will either
indicate a strong relationship
Variable A Variable B
16. • The Pearson Product Moment Correlation or simply
Pearson Correlation values range from -1.0 to 1.0
17. • The Pearson Product Moment Correlation or simply
Pearson Correlation values range from -1.0 to 1.0
-1 +10
18. • A Pearson Correlation of 1.0 has a perfect postive
relationship. Note two qualities here:
19. • A Pearson Correlation of 1.0 has a perfect postive
relationship. Note two qualities here:
(1) direction
20. • A Pearson Correlation of 1.0 has a perfect postive
relationship. Note two qualities here:
(1) direction
(2) strength
21. • A Pearson Correlation of 1.0 has a perfect postive
relationship. Note two qualities here:
(1) direction
(2) strength
• A +1.0 Pearson Correlation’s direction is positive and it’s
strength is very or perfectly strong.
22. • A Pearson Correlation of 1.0 has a perfect postive
relationship. Note two qualities here:
(1) direction
(2) strength
• A +1.0 Pearson Correlation’s direction is positive and it’s
strength is very or perfectly strong.
• A -1.0 Pearson Correlation’s direction is negative and it’s
strength is very or perfectly strong.
23. • A Pearson Correlation of 1.0 has a perfect postive
relationship. Note two qualities here:
(1) direction
(2) strength
• A +1.0 Pearson Correlation’s direction is positive and it’s
strength is very or perfectly strong.
• A -1.0 Pearson Correlation’s direction is negative and it’s
strength is very or perfectly strong.
• A 0.0 Pearson Correlation has no direction and has no
strength.
24. • A Pearson Correlation of 1.0 has a perfect postive
relationship. Note two qualities here:
(1) direction
(2) strength
• A +1.0 Pearson Correlation’s direction is positive and it’s
strength is very or perfectly strong.
• A -1.0 Pearson Correlation’s direction is negative and it’s
strength is very or perfectly strong.
• A 0.0 Pearson Correlation has no direction and has no
strength.
• A +0.3 Pearson Correlation’s direction is positive and it’s
strength is moderately weak.
25. • A Pearson Correlation of 1.0 has a perfect postive
relationship. Note two qualities here:
(1) direction
(2) strength
• A +1.0 Pearson Correlation’s direction is positive and it’s
strength is very or perfectly strong.
• A -1.0 Pearson Correlation’s direction is negative and it’s
strength is very or perfectly strong.
• A 0.0 Pearson Correlation has no direction and has no
strength.
• A +0.3 Pearson Correlation’s direction is positive and it’s
strength is moderately weak.
• A -0.1 Pearson Correlation’s direction is negative and it’s
strength is very weak.
26. • There is another quality as well. With a Pearson
correlation you are considering the relationship
between only two variables.
27. • There is another quality as well. With a Pearson
correlation you are considering the relationship
between only two variables.
28. • There is another quality as well. With a Pearson
correlation you are considering the relationship
between only two variables.
• Three’s a crowd:
29. • There is another quality as well. With a Pearson
correlation you are considering the relationship
between only two variables.
• Three’s a crowd:
30. • There is another quality as well. With a Pearson
correlation you are considering the relationship
between only two variables.
• Three’s a crowd:
• Bottom line: The Pearson Correlation is used only
when exploring the relationship between two
variables.
31. • Let’s look at a fictitious problem to illustrate how the
Pearson Correlation is calculated.
32. • Imagine you are conducting a study to determine the
relationship between the average daily temperature
and the average daily ice cream sales in a particular
city.
33. • Imagine you are conducting a study to determine the
relationship between the average daily temperature
and the average daily ice cream sales in a particular
city.
35. • Imagine the data set looks like this:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
36. • Notice how as one variable goes up (temperature)
the other variable increases (ice cream sales)
37. • Notice how as one variable goes up (temperature)
the other variable increases (ice cream sales)
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
38. • Notice how as one variable goes up (temperature)
the other variable increases (ice cream sales)
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
39. • One way to look at this relationship is to rank order
both variable values like so:
40. • One way to look at this relationship is to rank order
both variable values like so:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
41. • One way to look at this relationship is to rank order
both variable values like so:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st
42. • One way to look at this relationship is to rank order
both variable values like so:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st 1st
43. • One way to look at this relationship is to rank order
both variable values like so:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st 1st
44. • One way to look at this relationship is to rank order
both variable values like so:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st 1st
2nd 2nd
45. • One way to look at this relationship is to rank order
both variable values like so:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st 1st
2nd
3rd 3rd
2nd
46. • One way to look at this relationship is to rank order
both variable values like so:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st 1st
2nd
3rd 3rd
2nd
4th 4th
47. • One way to look at this relationship is to rank order
both variable values like so:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st 1st
2nd
5th 5th
4th 4th
3rd 3rd
2nd
48. • Notice how their rank orders are identical. And
because their standard deviations are similar as well,
these variables have a +1.0 Pearson Correlations.
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st 1st
2nd
5th 5th
4th 4th
3rd 3rd
2nd
49. • What would a perfectly negative correlation (-1.0)
look like?
55. • What would a zero correlation (0.0) look like?
56. • What would a zero correlation (0.0) look like?
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st
1st
2nd
5th 5th
4th
4th
3rd
3rd
2nd
57. • What would a zero correlation (0.0) look like?
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st
1st
2nd
5th 5th
4th
4th
3rd
3rd
2nd
58. • What would a zero correlation (0.0) look like?
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st
1st
2nd
5th 5th
4th
4th
3rd
3rd
2nd
59. • What would a zero correlation (0.0) look like?
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st
1st
2nd
5th 5th
4th
4th
3rd
3rd
2nd
60. • What would a zero correlation (0.0) look like?
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st
1st
2nd
5th 5th
4th
4th
3rd
3rd
2nd
61. • What would a zero correlation (0.0) look like?
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st
1st
2nd
5th 5th
4th
4th
3rd
3rd
2nd
62. • What would a zero correlation (0.0) look like?
• Note – Pearson Correlation is not just a comparison of rank ordered data
(that is what a Phi coefficient does) but the rank order is one factor that is
considered with a Pearson Correlation. Another factor is the degree to
which the standard deviations are similar.
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
1st
1st
2nd
5th 5th
4th
4th
3rd
3rd
2nd
63. • The Pearson Product Moment Correlation (PPMC) is
calculated as the average cross product of the z-
scores of two variables for a single group of people.
Here is the equation for the PPMC
64. • The Pearson Product Moment Correlation (PPMC) is
calculated as the average cross product of the z-
scores of two variables for a single group of people.
Here is the equation for the PPMC
𝑟 = ∑ (𝑍 𝑋 ∙ 𝑍 𝑌)
𝑛
65. • Let’s calculate the Pearson Correlation, for the
following data set:
66. • Let’s calculate the Pearson Correlation, for the
following data set:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
67. • Let’s calculate the Pearson Correlation, for the
following data set:
• It is very important to note that the Pearson Correlation
can be computed in a matter of seconds using statistical
software. The next set of slides is designed to help you
see what is happening conceptually as well as
computationally with the Pearson Correlation.
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
68. • When computing a Pearson Correlation you will
normally have two variables that DO NOT USE THE
SAME METRIC:
69. • When computing a Pearson Correlation you will
normally have two variables that DO NOT USE THE
SAME METRIC:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
70. • When computing a Pearson Correlation you will
normally have two variables that DO NOT USE THE
SAME METRIC:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
The metric here is degrees
71. • When computing a Pearson Correlation you will
normally have two variables that DO NOT USE THE
SAME METRIC:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
The metric here is degrees The metric here is number
of ice cream sales
72. • So we have to get these two variables on the same
metric. This is done by calculating the z scores or
standardized scores for the values from each
variable.
73. • So these raw score values in separate metrics are
transformed into standardized values which
converts them into the same metric:
74. • So these raw score values in separate metrics are
transformed into standardized values which
converts them into the same metric:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
75. • So these raw score values in separate metrics are
transformed into standardized values which
converts them into the same metric:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
76. • So these raw score values in separate metrics are
transformed into standardized values which
converts them into the same metric:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
77. • So these raw score values in separate metrics are
transformed into standardized values which
converts them into the same metric:
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
Different Metric
(raw scores)
78. • Note – this is done by subtracting each value from
it’s mean (e.g., 900 minus 700 = 200) and dividing it
by it’s standard deviation (e.g., 200 / 14.1 = 1.4)
Ave Daily Temp
900
800
700
600
500
Ave Daily Ice Cream Sales
560
480
350
320
230
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
Different Metric
(raw scores)
79. • Once the values are standardized we multiply them
80. • Once the values are standardized we multiply them
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
81. • Once the values are standardized we multiply them
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
82. • Once the values are standardized we multiply them
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
83. • Once the values are standardized we multiply them
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
X
X
X
X
X
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
84. • Once the values are standardized we multiply them
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
X
X
X
X
X
1.9
0.4
0.0
0.6
2.1
=
=
=
=
=
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
85. • Once the values are standardized we multiply them
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
X
X
X
X
X
1.9
0.4
0.0
0.6
2.1
=
=
=
=
=
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
These are called cross products
because we are multiplying
across two values
86. • Once the values are standardized we multiply them
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
X
X
X
X
X
1.9
0.4
0.0
0.6
2.1
=
=
=
=
=
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
1.9 + 0.4 + 0.0 + 0.6 + 2.1 = 5.0
88. • Finally, divide that number (5.0) by the number of
observations
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
89. • Finally, divide that number (5.0) by the number of
observations
𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝑛
The number of observations
(in this case 5)
Ave Daily Temp
+1.4
+0.7
0.0
-0.7
-1.4
Ave Daily Ice Cream Sales
+1.5
+0.8
-0.3
-0.6
-1.3
1
2
3
4
5
91. 𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝟓
The number of observations
(in this case 5)
𝑟 =
𝟓
𝟓
92. 𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝟓
The number of observations
(in this case 5)
𝑟 =
𝟓
𝟓
Sum of the cross products
1.9 + 0.4 + 0.0 + 0.6 + 2.1 =
5.0
93. 𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝟓
The number of observations
(in this case 5)
𝑟 =
𝟓
𝟓
Sum of the cross products
1.9 + 0.4 + 0.0 + 0.6 + 2.1 =
5.0
𝑟 = +𝟏. 𝟎
94. 𝑟 =
∑ (𝒁 𝑿 ∙ 𝒁 𝒀)
𝟓
The number of observations
(in this case 5)
𝑟 =
𝟓
𝟓
Sum of the cross products
1.9 + 0.4 + 0.0 + 0.6 + 2.1 =
5.0
𝑟 = +𝟏. 𝟎
This is the Pearson Correlation
which in this case is a perfect
positive relationship
96. • In summary:
• The Pearson Product Moment Correlation can range
from -1 to 0 to +1.
97. • In summary:
• The Pearson Product Moment Correlation can range
from -1 to 0 to +1.
-1 +10
98. • A correlation of 0 indicates no association between
the variables of interest.
99. • A correlation of 0 indicates no association between
the variables of interest.
• The direction (positive or negative) simply indicates a
positive or negative (inverse) relationship between
the variables.
100. • If POSITIVE, when values increase on one variable,
they tend to increase on another variable.
101. • If POSITIVE, when values increase on one variable,
they tend to increase on another variable.
Variable 1
10
9
8
7
Variable 2
5
4
3
2
-1 +10
102. • If POSITIVE, when values increase on one variable,
they tend to increase on another variable.
Variable 1
10
9
8
7
Variable 2
5
4
3
2
-1 +10
103. • If POSITIVE, when values increase on one variable,
they tend to increase on another variable.
Variable 1
10
9
8
7
Variable 2
5
4
3
2
Pearson
Correlation = +1.0
-1 +10
104. • If NEGATIVE, when values increase on one variable,
they tend to decrease on another variable.
105. • If NEGATIVE, when values increase on one variable,
they tend to decrease on another variable.
Variable 1
10
9
8
7
Variable 2
5
4
3
2
-1 +10
106. • If NEGATIVE, when values increase on one variable,
they tend to decrease on another variable.
Variable 1
10
9
8
7
Variable 2
5
4
3
2
Pearson
Correlation = -1.0
-1 +10
107. • The strength of the relationship depends on the
decimal value.
108. • The strength of the relationship depends on the
decimal value.
-1 +10
109. • The strength of the relationship depends on the
decimal value.
-1 +10
110. • The strength of the relationship depends on the
decimal value.
-1 +10 0.2
weak
111. • The strength of the relationship depends on the
decimal value.
-1 +10
112. • The strength of the relationship depends on the
decimal value.
-1 +10 0.8
strong
113. • The strength of the relationship depends on the
decimal value.
-1 +10
114. • The strength of the relationship depends on the
decimal value.
-1 +10
0.2
weak
115. • The strength of the relationship depends on the
decimal value.
-1 +10
116. • The strength of the relationship depends on the
decimal value.
-1 +10
0.8
strong
117. • The strength of the relationship depends on the
decimal value.
-1 +10
118. • There is a tendency to interpret the Pearson Product
Moment Correlation with causal language as though
changes in one variable causes changes in the other.
119. • There is a tendency to interpret the Pearson Product
Moment Correlation with causal language as though
changes in one variable causes changes in the other.
• Whether to interpret the Pearson Product Moment
Correlation as prediction or causation depends on
the nature of the research design rather than the
nature of the statistic.
120. • There is a tendency to interpret the Pearson Product
Moment Correlation with causal language as though
changes in one variable causes changes in the other.
• Whether to interpret the Pearson Product Moment
Correlation as prediction or causation depends on
the nature of the research design rather than the
nature of the statistic.
• First, analyze the nature of the research design
before interpreting the Pearson Product Moment
Correlation with causal or prediction language.
121. • There is a tendency to interpret the Pearson Product
Moment Correlation with causal language as though
changes in one variable causes changes in the other.
• Whether to interpret the Pearson Product Moment
Correlation as prediction or causation depends on
the nature of the research design rather than the
nature of the statistic.
• First, analyze the nature of the research design
before interpreting the Pearson Product Moment
Correlation with causal or prediction language.
• So, if your research question is focused on the
relationship between two continuous variables the
Pearson Product Moment Correlation would be the
appropriate statistical method to use.