Upload
malim-muhammad-siregar
View
223
Download
0
Embed Size (px)
Citation preview
8/12/2019 33851
1/55
8/12/2019 33851
2/55
Correlation &Regression
Dr. Moataza Mahmoud Abdel Wahab
Lecturer of BiostatisticsHigh Institute of Public Health
University of Alexandria
8/12/2019 33851
3/55
CorrelationFinding the relationship between two
quantitative variables without being able to
infer causal relationships
Correlationis a statistical technique used to
determine the degree to which two
variables are related
8/12/2019 33851
4/55
Rectangular coordinate
Two quantitative variables
One variable is called independent (X) and
the second is called dependent (Y)
Points are not joined
No frequency table
Scatter diagram
Y
* *
*
X
8/12/2019 33851
5/55
Wt.
(kg)
67 69 85 83 74 81 97 92 114 85
SBP
mHg)
120 125 140 160 130 180 150 140 200 130
Example
8/12/2019 33851
6/55
8/12/2019 33851
7/55
80
100
120
140
160
180
200
220
60 70 80 90 100 110 120Wt (kg)
SBP(mmHg)
Scatter diagram of weight and systolic blood pressure
8/12/2019 33851
8/55
Scatter plots
The pattern of data is indicative of the type of
relationship between your two variables:
positive relationship
negative relationship
no relationship
8/12/2019 33851
9/55
Positive relationship
8/12/2019 33851
10/55
0
2
4
6
8
10
12
14
16
18
0 10 20 30 40 50 60 70 80 90
Age in Weeks
Heightin
CM
8/12/2019 33851
11/55
Negative relationship
Reliability
Age of Car
8/12/2019 33851
12/55
No relation
8/12/2019 33851
13/55
Correlation Coefficient
Statistic showing the degree of relation
between two variables
8/12/2019 33851
14/55
Simple Correlation coefficient (r)
It is also called Pearson's correlation
or product moment correlation
coefficient.
It measures the natureand strength
between two variables of
the quantitativetype.
8/12/2019 33851
15/55
The signof rdenotes the nature of
association
while the valueof rdenotes the
strength of association.
8/12/2019 33851
16/55
If the sign is +ve this means the relationis direct (an increase in one variable isassociated with an increase in theother variable and a decrease in onevariable is associated with a
decrease in the other variable).
While if the sign is -ve this means an
inverse or indirect relationship (whichmeans an increase in one variable isassociated with a decrease in the other).
8/12/2019 33851
17/55
The value of r ranges between ( -1) and ( +1)
The value of r denotes the strength of theassociation as illustrated
by the following diagram.
-110
-0.25-0.75 0.750.25
strong strongintermediate intermediateweak weak
no relation
perfect
correlation
perfect
correlation
Directindirect
8/12/2019 33851
18/55
If r= Zero this means no association orcorrelation between the two variables.
If 0 < r< 0.25= weak correlation.
If 0.25 r< 0.75= intermediate correlation.
If 0.75 r< 1= strong correlation.
If r = l= perfect correlation.
8/12/2019 33851
19/55
n
y)(y.
n
x)(x
n
yxxy
r2
2
2
2
How to compute the simple correlation
coefficient (r)
8/12/2019 33851
20/55
Example:
A sample of 6 children was selected, data about their
age in years and weight in kilograms was recorded as
shown in the following table . It is required to find the
correlation between age and weight.
Weight
(Kg)
Age
(years)
serial
No
1271
862
1283
1054
1165
1396
8/12/2019 33851
21/55
These 2 variables are of the quantitative type, one
variable (Age) is called the independent and
denoted as (X) variable and the other (weight)is called the dependent and denoted as (Y)
variables to find the relation between age and
weight compute the simple correlation coefficient
using the following formula:
n
y)(y.
n
x)(x
n
yxxy
r2
22
2
8/12/2019 33851
22/55
Y2X2xy
Weight
(Kg)
(y)
Age
(years)
(x)
Serial
n.
14449841271
643648862
14464961283
10025501054
12136661165
169811171396
y2=
742
x2=
291
xy=
461
y=
66
x=
41
Total
8/12/2019 33851
23/55
r = 0.759
strong direct correlation
6
(66)742.
6
(41)291
6
6641461
r22
8/12/2019 33851
24/55
EXAMPLE: Relationship between Anxiety and
Test Scores
Anxiety
(X)
Test
score (Y)X2 Y2 XY
10 2 100 4 20
8 3 64 9 242 9 4 81 18
1 7 1 49 7
5 6 25 36 306 5 36 25 30
X = 32 Y = 32 X2= 230 Y2= 204 XY=129
8/12/2019 33851
25/55
Calculating Correlation Coefficient
94.
)200)(356(
1024774
32)204(632)230(6
)32)(32()129)(6(22
r
r = - 0.94
Indirect strong correlation
8/12/2019 33851
26/55
Spearman Rank Correlat ion Coeff ic ient
(rs)
It is a non-parametric measure of correlation.
This procedure makes use of the two sets ofranks that may be assigned to the sample
values of x and Y.Spearman Rank correlation coefficient could becomputed in the following cases:
Both variables are quantitative.
Both variables are qualitative ordinal.
One variable is quantitative and the other isqualitative ordinal.
8/12/2019 33851
27/55
Procedure:
1. Rank the values of X from 1 to n where nis the numbers of pairs of values of X and
Y in the sample.
2.
Rank the values of Y from 1 to n.3. Compute the value of di for each pair of
observation by subtracting the rank of Yi
from the rank of Xi
4. Square each di and compute di2 which
is the sum of the squared values.
8/12/2019 33851
28/55
5. Apply the following formula
1)n(n
(di)61r2
2
s
The value of rsdenotes the magnitude
and nature of association giving the same
interpretation as simple r.
8/12/2019 33851
29/55
Example
In a study of the relationship between leveleducation and income the following data was
obtained. Find the relationship between themand comment.
Income
(Y)
level education
(X)
sample
numbers
25Preparatory.A
10Primary.B
8University.C
10secondaryD15secondaryE
50illiterateF
60University.G
8/12/2019 33851
30/55
Answer:
di2diRank
Y
Rank
X(Y)(X)423525PreparatoryA
0.250.55.5610Primary.B
30.25-5.571.58University.C
4-25.53.510secondaryD
0.25-0.543.515secondaryE
2552750illiterateF
0.250.511.560university.G
di2=64
8/12/2019 33851
31/55
Comment:
There is an indirect weak correlation
between level of education and income.
1.0)48(7
646
1
sr
i
8/12/2019 33851
32/55
exercise
8/12/2019 33851
33/55
Regression Analyses
Regression: technique concerned with predictingsome variables by knowing others
The process of predicting variable Y usingvariable X
8/12/2019 33851
34/55
Regression
Uses a variable (x) to predict some outcome
variable (y)
Tells you how values in y change as a function
of changes in values of x
8/12/2019 33851
35/55
Correlation and Regression
Correlation describes the strength of a linear
relationship between two variables
Linearmeans straight line
Regressiontells us how to draw the straight line
described by the correlation
Regression
8/12/2019 33851
36/55
Regression
Calculates the best-fit line for a certain set of data
The regression line makes the sum of the squares of
the residuals smaller than for any other lineRegression minimizes residuals
80
100
120
140
160
180
200
220
60 70 80 90 100 110 120Wt (kg)
SBP(mmHg)
8/12/2019 33851
37/55
By using the least squares method (a procedure
that minimizes the vertical deviations of plotted
points surrounding a straight line) we are
able to construct a best fitting straight line to the
scatter diagram points and then formulate a
regression equation in the form of:
n
x)(x
n
yxxy
b2
2
1)xb(xyy b
bXay
R i E ti
8/12/2019 33851
38/55
Regression Equation
Regression equation
describes the
regression line
mathematically
Intercept Slope
80
100
120
140
160
180
200
220
60 70 80 90 100 110 120Wt (kg)
SBP(mmHg)
8/12/2019 33851
39/55
Linear Equations
Y
Y = bX + a
a = Y-intercept
X
Changein Y
Change in X
b = Slope
bXay
8/12/2019 33851
40/55
Hours studying and grades
8/12/2019 33851
41/55
Regressing grades on hours
Linear Regression
2.00 4.00 6.00 8.00 10.00
Number of hours spent studying
70.00
80.00
90.00
Finalgr
ade
in
course
Final grade in course = 59.95 + 3.17 * study
R-Square = 0.88
Predicted final grade in class =
59.95 + 3.17*(number of hours you study per week)
Predicted final grade in class = 59.95 + 3.17*(hours of study)
8/12/2019 33851
42/55
Predict the final grade of
Someone who studies for 12 hours
Final grade = 59.95 + (3.17*12)
Final grade = 97.99
Someone who studies for 1 hour:
Final grade = 59.95 + (3.17*1)
Final grade = 63.12
Predicted final grade in class 59.95 3.17 (hours of study)
8/12/2019 33851
43/55
Exercise
A sample of 6 persons was selected thevalue of their age ( x variable) and their
weight is demonstrated in the following
table. Find the regression equation and
what is the predicted weight when age is
8.5 years.
8/12/2019 33851
44/55
Weight (y)Age (x)Serial no.
12
8
1210
11
13
7
6
85
6
9
1
2
34
5
6
8/12/2019 33851
45/55
Answer
Y2X2xyWeight (y)Age (x)Serial no.
144
64
144100
121
169
49
36
6425
36
81
84
48
9650
66
117
12
8
1210
11
13
7
6
85
6
9
1
2
34
5
6
7422914616641Total
8/12/2019 33851
46/55
6.836
41x 11
6
66y
92.0
6
)41(
291
6
6641461
2
b
Regression equation
6.83)0.9(x11y (x)
8/12/2019 33851
47/55
0.92x4.675y (x)
12.50Kg8.5*0.924.675y(8.5)
Kg58.117.5*0.924.675y(7.5)
8/12/2019 33851
48/55
11.4
11.6
11.8
12
12.2
12.4
12.6
7 7.5 8 8.5 9
Age (in years)
Weight(i
nKg)
we create a regression line by plotting two
estimated values for y against their X component,
then extending the line right and left.
8/12/2019 33851
49/55
Exercise 2
The following are theage (in years) andsystolic bloodpressure of 20
apparently healthyadults.
B.P
(y)
Age
(x)
B.P
(y)
Age
(x)
128136
146
124
143
130
124
121126
123
4653
60
20
63
43
26
1931
23
120128
141
126
134
128
136
132140
144
2043
63
26
53
31
58
4658
70
8/12/2019 33851
50/55
Find the correlation between ageand blood pressure using simple
and Spearman's correlationcoefficients, and comment.
Find the regression equation?
What is the predicted bloodpressure for a man aging 25 years?
8/12/2019 33851
51/55
x2xyyxSerial
4002400120201
1849550412843239698883141633
6763276126264
28097102134535
9613968128316
33647888136587
21166072132468
336481201405894900100801447010
8/12/2019 33851
52/55
x2xyyxSerial
211658881284611
280972081365312360087601466013
40024801242014
396990091436315184955901304316
67632241242617
3612299121191896139061263119
52928291232320
416781144862630852Total
8/12/2019 33851
53/55
n
x)(x
n
yxxy
b2
21
4547.0
2085241678
20
2630852114486
2
=
=112.13 + 0.4547 x
for age 25
B.P = 112.13 + 0.4547 * 25=123.49 = 123.5 mm hg
y
8/12/2019 33851
54/55
Multiple Regression
Multiple regression analysis is a
straightforward extension of simple
regression analysis which allows more
than one independent variable.
8/12/2019 33851
55/55