Statistic with Computer Education

Module no 6 Stat.
MEASURES OF CORRELATION
Objectives:
1. Definition of Measures of Correlation
2. The Pearson Product-Moment Correlation coefficient
3. Other interpretation of Pearson’s r
4. Coefficient of Determination
5. Testing the Significance r
6. Exercise
Source:
 Statistic with computer Applications, Second Edition by Priscilla S. Altares, Antonio roland Copo,
Yonardo A. Gabuyo, Angeline T. Laddaran, Leila D.P. Mehia, Immaculata A. Policarpiop, Helen D.
Tizon, Ana Maria S.D. Yao
 https://www.statisticssolutions.com/correlation-pearson-kendall-spearman/

Definition
Correlation is a statistical tool to measure the association of two or more quantitative variables. It
is concerned with the relationship in the changes and movements of two variables. Correlation is also define
as the measure of the linear relationship between two random variable x and y, and is denoted by r. That is,
r measures the extent to which the points cluster about a straight line.
There are three degrees of relationship or correlation between two variables:
1. Perfect correlation (positive and negative)
2. Some degree of correlation (positive and negative)
3. No correlation
The quantitative interpretation of the degree of linear relationship existing is shown in the following
range of values.
±1.00 perfect positive (negative) correlation
±0.91 - ±0.99 very high positive (negative) correlation
±0.71 - ±0.90 high positive (negative) correlation
±0.51 -±0.70 moderately positive (negative) correlation
±0.31 - ±0.50 low positive (negative) correlation
±0.01 - ±0.30 negligible positive (negative) correlation
0.00 no correlation
The Pearson Product-Moment Correlation Coefficient
Of the various correlation coefficient in current use, the most widely used computational formula
for correlation is the Pearson Product-Moment correlation Coefficient, the symbol of which is r. the
size of the correlation varies from ± through 0 to -1. The formula for the Pearson r is
𝑁 ∑ 𝑋𝑌− ∑ 𝑋 ∑ 𝑌
r=
√[ 𝑁 ∑ 𝑋 2 − ( ∑ 𝑋)2 ] [𝑁 ∑ 𝑌 2 − ( ∑ 𝑌)2 ]

where
x = the observed data for the independent variable
y = the observed data for the dependent variable
N = sample size
r = degree of relationship between x and y

Example:
Suppose we want to find out if a relationship exist between the height and weight of 20
UE college students.
Student X Y
Height(cm) Weight
(kg)
1 170 72
2 172 70
3 158 60
4 165 73
5 180 85
6 195 98
7 183 78
8 175 76
9 182 82
10 190 90
11 165 75
12 184 80
13 194 90
14 176 78
15 182 84
16 178 80
17 180 82
18 168 76
19 157 64
20 185 85
 For the height and weights of 20 college students, the computation in r is tabulated as
follows:

Student X Y XY 𝑥2 𝑦2
Height(cm Weight (kg)
)
1 170 72 12,240 28,900 5,184
2 172 70 12,040 29,584 4,900
3 158 60 9,480 24,964 3,600
4 165 73 12,045 27,225 5,329
5 180 85 15,300 32,400 7,225
6 195 98 19,110 38,025 9,604
7 183 78 14,274 33,489 6,084
8 175 76 13,300 30,625 5,776
9 182 82 14,924 33,124 6,724
10 190 90 17,100 36,100 8,100
11 165 75 12,375 27,225 5,625
12 184 80 14,720 33,856 6,400
13 194 90 17,460 37,636 8,100
14 176 78 13,728 30,976 6,084
15 182 84 15,288 33,124 7,056
16 178 80 14,240 31,684 6,400
17 180 82 14,760 32,400 6,724
18 168 76 12,768 28,224 5,776
19 157 64 10,048 24,649 4,096
20 185 85 15,725 34,225 7,225
Total ∑ 𝑋 =3,539 ∑ 𝑌 = 1,578 ∑ 𝑋 𝑌 = 280,925 ∑ 𝑥 2 =628,435 ∑ 𝑦 2 =126,012
.
𝑁 ∑ 𝑋𝑌− ∑ 𝑋 ∑ 𝑌
r=
√[ 𝑁 ∑ 𝑋 2 − ( ∑ 𝑋)2 ] [𝑁 ∑ 𝑌 2 − ( ∑ 𝑌)2 ]

20(280,925)−(3,539)(1,578)
r=
√[20(628,435)−(3,539)2 )][(20(126,012)−(1,578)2 ]

5,618,500−5,584,542
r=
√(12,568,700−12,524,521)(2,520,240−2,490,084)

33,958
r=
√(44,179)(30,156)

33,958
r=
√1,332,261,924

33,958
r=
36,500.16

r= 0.93
Referring to our table of qualitative interpretation of r, we can conclude that the heights and
weights of the 20 students in the sample are having a very high correlation.
Coefficient of Determination
As noted above, correlation coefficients are interpreted in a relative sort of way by talking about
high, moderate or low coefficients. However, the best way to evaluate an r is by the use of the coefficient
of determining, denoted by 𝑟 2 . In our example, our computed r = 0.93. The coefficient of determination 𝑟 2
= 0.86. This means that for the 20 college students in the sample 86% of the variation in weights could be
attributed to the variation in heights. The rest (14%) is chance variation.

Testing the Significance r

𝑛−2
T = r√
1−𝑟 2

4. Decision Rule:
Reject 𝐻𝑜 if t computed is > 2.10 or <-2.10 otherwise accept 𝐻𝑜
(Note: 2.10 coming from the table of critical value t, .05 df of 18)

5. Computation

20−2
t = 0.93 √
1−(0.93)2

18
= 0.93 √
1−0.8649

18
= 0.93 √
0.1351

= 0.93 √133.23
= 0.93 (11.5)
= 10.73
 6. Conclusion:
Since 10.73 > 2.16, reject 𝐻𝑜 and accept𝐻𝑎 . Therefore we can say that there is a significant
relationship between weight and height of the 20 college students.

Exercise:
Twenty students are given a test of general ability and an English achievement test with the
resulting scores:
Mental Ability English
55 205
54 180
52 203
50 201
50 186
48 203
47 195
46 188
46 192
45 195
43 190
40 188
40 194
38 165
38 170
37 185
36 188
36 194
34 180
32 185

Compute r for these data.