Scribd
Upload
35 views12 pages

Partial Correlation

Uploaded by

Joysri Roy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
35 views12 pages

Partial Correlation

Uploaded by

Joysri Roy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 12

CORRELAT

ION
Dr. Ishita Chatterjee, Associate Professor, Applied
Psychology, University of Calcutta, Rajabazar Science
College
NEED & IMPORTANCE
• While conducting studies (especially in the field of education and / psychology),
the researcher often find that the relationship between two variables is greatly
influenced by a third variable or an additional variable. In such a situation, it
becomes quite difficult to have a reliable and an independent estimate of
correlation between these two variables unless there is some possibility of
nullifying the effects of third variables (or a number of other variables) on the
variables in question.
• Thus, Partial correlation measures relationship between two variables (X & Y)
while eliminating influence of third variable (Z) from the relationship.
• Hence, the purpose is to find the unique variance between two variables while
eliminating the variance from a third number.
DEFINITION
Partial correlation can thus be described as a special correlation
technique, which is helpful in estimating independent and reliable
relationship between any two variables by eliminating and ruling out
any undesirable influence or interference of a third or an additional
variable on the variables being correlated.
CONTD…
• ASSUMPTION: Control the main variables (two
or three) and they will automatically control so
many other variables since other variables are
related to these controlled variables.
• A partial correlation is a type of Pearson
correlation coefficient that can range in value
from -1 to +1.
• The simplest partial correlation involves only
three variables, a predictor variable, and a
control variable.
WHEN IT SHOULD BE CONDUCTED?
• When the third variable has shown a relationship to one or both variables
under investigation.
• For example, a researcher want to know the relationship between height
(X) and weight (Y). However, she is aware that people’s bone and muscle
structure vary according to age. She can calculate a partial correlation
between height and weight while removing (holding constant,
eliminating, adjusting or controlling) the effect of gender on the
correlation.
• The height (X) and weight (Y) of 10 young males aged 12 to 19 (Z) was
taken.
Here, correlation between height (X) and wight (Y) will be high because
older = taller on average, and older = heavier on average. So it has been
recognized the presence of a third variable, age (Z).
COMPUTATION OF PARTIAL
CORRELATION
1. FORMULAE FOR THE COMPUTATION OF FIRST ORDER
PARTIAL CORRELATION
r12 - r13 – r23 • First order partial correlations are those in which the
relationship between two variables is estimated while
rp = r12.3 = making the third variable constant or partialled out.
√ 1-r132 √ 1-r232

• Here, 1 = Achievement Scores


r13 - r12 – r23
2 = Participation in co-curricular activities
rp = r13.2 = √ 1-r132 √ 1-r232 scores
3 = IQ Scores
r23 - r12 – r13 • Hence, in the first order partial correlation, r12.3 /
r13.2 / r23.1 means the correlation between 1 and 2
rp = r23.1 = √ 1-r122 √ 1-r132 (achievement and participation in co-curricular
activities) while making the third variable
“intelligence” as constant or partialled out.
COMPUTATION OF PARTIAL
CORRELATION
2. FORMULAE FOR THE COMPUTATION OF SECOND
ORDER PARTIAL CORRELATION
r12.3 - r14.3 – r24.3 Hence, in the second order partial
rp = r12.34 = correlation, r12.34 means correlation
√ 1-r14.32 √ 1-r24.32 between 1 and 2 (achievement and
participation in co-curricular activities)
while making the third and fourth
Here, 1 = Achievement Scores (“intelligence” and “socio-economic
status”) as constant or partialled out.
2 = Participation in co-curricular
activities scores
3 = IQ Scores
4 = Socio-Economic Status Scores
COMPUTATION OF PARTIAL
CORRELATION
3. FORMULAE FOR THE COMPUTATION OF THIRD ORDER
PARTIAL CORRELATION
r12.34 - r15.34 – r25.34 Hence, in the third order partial
rp = r12.345 = correlation, r12.345 means correlation
√ 1-r15.342 √ 1-r25.342 between 1 and 2 (achievement and
participation in co-curricular activities)
while making the third, fourth, and fifth
(“intelligence”, “socio-economic status”,
Here, 1 = Achievement Scores and “age”) as constant or partialled out.
2 = Participation in co-curricular
activities scores
In this way, the number to the right of the
3 = IQ Scores decimal point represents variables whose
4 = Socio-Economic Status Scores influence is ruled out and the number to
there left represents those two variables
5 = Age Score whose relationship needs to be estimated.
EXAMPLE WITH ILLUSTRATION
• PROBLEM: From a certain number of schools in Delhi, a sample of 500 students
studying in classes IX and X was taken. These students were evaluated in terms
of their academic achievement and participation in co-curricular activities.
Their IQ’s were also tested. The correlation among these three variables was
obtained and recorded as follows:
r12 = 0.18, r23 = 0.70, r13 = 0.60,

• SOLUTION: The independent and reliable correlation between academic


achievement and participation in co-curricular activities – the main two
variables – can be found by computing partial correlation between these two
variables, i.e., by computing r12.3 (keeping constant the third variable).
CONTD…
0.38
rp = r12.3 = r12 - r13 – r23 = 0.5712
√ 1-r132 √ 1-r232
3800
0.80 – 0.60 × 0.70
= 5712
=
√1 - 0.60 × 0.60 √1 - 0.70 × 0.70 1900
0.80 – 0.42 2856
=
=
√1 - 0.36 √1 – 0.49
0.38
= = 0.67
√0.64
√0.51
0.38
= [Ans. Partial correlation = 0.67.]
0.8 × 0.714
APPLICATIO SIGNIFICAN
N CE
• Partial correlation can be used as a • The significance of the first and second
special statistical technique for order partial correlation ‘r’ – can be
eliminating the effects of one or more tested easily by using the ‘t’ distribution:
variables on the two main variables,
for which researchers want to N–2-K
compute an independent and a
reliable measure of correlation.
t= √ 1 – r2
Where,
• Besides its major advantage lies in K = Order of partial r
the fact that it enable researchers to r = Value of partial correlation
set up a multiple regression equation
of two or more variables, by means of N = Total frequencies in the sample
study
which researchers can predict another
variable or criterion. Therefore,
Degree of freedom (df) = N-2-K

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy