The Basic Understanding of
Correlation
DR. RAJEEV KUMAR, M.S.W. (TISS, MUMBAI), M.PHIL. (CIP,
RANCHI), UGC-JRF, PH.D. (IIT KHARAGPUR)
What is correlation?
The components of correlation
What is the meaning of ‘r’
‘r’ is the coefficient of correlation, it tells the strength or weakness of
correlation
Direction of correlation
The direction is denoted by sign of positive or negative.
Few things to remember in Pearson’s
Correlation
 It shows the association between two continuous variables.
 Correlation is always bidirectional
A B . It means, if A is increasing B is also increasing. In case of
positive correlation.
In case of negative correlation, A is increasing, B is decreasing.
But correlation does not show causative relation.
Means B is not increasing due to A.
A is not influencing B
A is not predicting B
It is a linear relation
What happens, when (x) value is
increasing, but (y) is constant.
Here r=0 (zero correlation)
X= Government expenditure on infrastructure, Y=
condition of infrastructure. Inference: No
development and corruption is suspected
When (y) is increasing but (x) is
constant. Again (r=0) means zero
correlation.
X= Police action Y= crime rate
police is taking some action and crime is increasing. It
means, there is failure of police. And many reasons may be
behind it. Corruption is one of them
Another example of zero correlation.
If both (x) and (y) are constant. (x) is neither increasing or decreasing
and (y) is also neither increasing or decreasing.
Means, neither government is spending money, nor there is
development
x= no study further y= no improvement in result.
When both (x) and (y) are increasing
together
X= study hours, y= grades or marks
x= government expenditure, y= improvement in
infrastructure and health. It shows good
governance and development
When (x) is increasing but (y) is decreasing. This is
called negative correlation. Here the direction is
negative
Negative correlation may denote both positive and
negative meanings. X= government expenditure
on health, y= reduction in diseases. Good
governance
The mixed pattern of (x) and (y)
Fluctuation in trends over a time.
The best example is corona cases in
country
The summary of correlation plot
the line join those dots is called best fit
line
How to calculate the correlation
How to interpret the strength of correlation?
If df is up to 30. Here df= n-2
Table of critical values of Pearson
correlation
Exercise -1: just check whether the correlation is significant
??
A study among 35 college students to see the correlation
between social support and GPA
Exercise-2: A study among 12 students to see the
correlation between motivation and success.
Whether correlation is signification?
Exercise-3: A study was conducted among 10 individual
to assess correlation between age and cholesterol. See if
the correlation is significant ?
Understanding p value
How to interpret the ‘p’ value
 If value of ‘p’ is less than or equal to
.05, then p is significant at 0.05. we
can say (p≤.05) and CI=95%
 If value of ‘p’ is less than or equal to
.01, then p is significant at 0.01. we
can say (p≤.01) and CI=99%
 If value of ‘p’ is less than or equal to
.001, then p is significant at 0.001. we
can say (p≤.001) and CI=99.99%
It means it tested at CI=99.99%
( p= 0.001)
Exercise- 4
find out the significant ‘p’ values and at what alpha
level
 .86 .76 .0007 .0008 .000001 .93 .0003
 .98 .06 .006 .63 .054 .049 .051
 .56
 .007
 .09
 .54
 .18
 .01
Exercise-5: See if ‘p’ value is significant
Exercise-6: Tell me, which ‘p’ values are significant?
Thanks!
Keep learning!
In the next session, we will learn how
to calculate the Pearson Correlation,
manually.

Lecture 8 basic concepts of correlation

  • 1.
    The Basic Understandingof Correlation DR. RAJEEV KUMAR, M.S.W. (TISS, MUMBAI), M.PHIL. (CIP, RANCHI), UGC-JRF, PH.D. (IIT KHARAGPUR)
  • 2.
  • 3.
    The components ofcorrelation What is the meaning of ‘r’ ‘r’ is the coefficient of correlation, it tells the strength or weakness of correlation Direction of correlation The direction is denoted by sign of positive or negative.
  • 4.
    Few things toremember in Pearson’s Correlation  It shows the association between two continuous variables.  Correlation is always bidirectional A B . It means, if A is increasing B is also increasing. In case of positive correlation. In case of negative correlation, A is increasing, B is decreasing. But correlation does not show causative relation. Means B is not increasing due to A. A is not influencing B A is not predicting B It is a linear relation
  • 5.
    What happens, when(x) value is increasing, but (y) is constant. Here r=0 (zero correlation)
  • 6.
    X= Government expenditureon infrastructure, Y= condition of infrastructure. Inference: No development and corruption is suspected
  • 7.
    When (y) isincreasing but (x) is constant. Again (r=0) means zero correlation.
  • 8.
    X= Police actionY= crime rate police is taking some action and crime is increasing. It means, there is failure of police. And many reasons may be behind it. Corruption is one of them
  • 9.
    Another example ofzero correlation. If both (x) and (y) are constant. (x) is neither increasing or decreasing and (y) is also neither increasing or decreasing. Means, neither government is spending money, nor there is development x= no study further y= no improvement in result.
  • 10.
    When both (x)and (y) are increasing together
  • 11.
    X= study hours,y= grades or marks x= government expenditure, y= improvement in infrastructure and health. It shows good governance and development
  • 12.
    When (x) isincreasing but (y) is decreasing. This is called negative correlation. Here the direction is negative
  • 13.
    Negative correlation maydenote both positive and negative meanings. X= government expenditure on health, y= reduction in diseases. Good governance
  • 14.
    The mixed patternof (x) and (y)
  • 15.
    Fluctuation in trendsover a time. The best example is corona cases in country
  • 16.
    The summary ofcorrelation plot the line join those dots is called best fit line
  • 17.
    How to calculatethe correlation
  • 18.
    How to interpretthe strength of correlation? If df is up to 30. Here df= n-2
  • 20.
    Table of criticalvalues of Pearson correlation
  • 22.
    Exercise -1: justcheck whether the correlation is significant ?? A study among 35 college students to see the correlation between social support and GPA
  • 23.
    Exercise-2: A studyamong 12 students to see the correlation between motivation and success. Whether correlation is signification?
  • 24.
    Exercise-3: A studywas conducted among 10 individual to assess correlation between age and cholesterol. See if the correlation is significant ?
  • 25.
    Understanding p value Howto interpret the ‘p’ value  If value of ‘p’ is less than or equal to .05, then p is significant at 0.05. we can say (p≤.05) and CI=95%  If value of ‘p’ is less than or equal to .01, then p is significant at 0.01. we can say (p≤.01) and CI=99%  If value of ‘p’ is less than or equal to .001, then p is significant at 0.001. we can say (p≤.001) and CI=99.99% It means it tested at CI=99.99% ( p= 0.001)
  • 26.
    Exercise- 4 find outthe significant ‘p’ values and at what alpha level  .86 .76 .0007 .0008 .000001 .93 .0003  .98 .06 .006 .63 .054 .049 .051  .56  .007  .09  .54  .18  .01
  • 27.
    Exercise-5: See if‘p’ value is significant
  • 28.
    Exercise-6: Tell me,which ‘p’ values are significant?
  • 29.
    Thanks! Keep learning! In thenext session, we will learn how to calculate the Pearson Correlation, manually.