Scribd
Upload
12 views

partial_correlation_intro_1 (1)

The document explains partial correlation, which measures the relationship between two variables while controlling for the influence of one or more additional variables. It differentiates between first-order and second-order partial correlations and provides examples of calculating these correlations along with their statistical significance using t-values. The document also includes practice problems and resources for further learning.

Uploaded by

ashishsharma120513
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
12 views

partial_correlation_intro_1 (1)

The document explains partial correlation, which measures the relationship between two variables while controlling for the influence of one or more additional variables. It differentiates between first-order and second-order partial correlations and provides examples of calculating these correlations along with their statistical significance using t-values. The document also includes practice problems and resources for further learning.

Uploaded by

ashishsharma120513
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Partial Correlation

Review of Correlation

You previously learned about the Pearson product-moment correlation, also known as Pearson’s r:

(∑ ) (∑ )(∑ )
(√ (∑ ) (∑ ) ) (√ (∑ ) (∑ ) )

The statistical significance of Pearson’s r can be tested using a t-value:

Pearson’s r is a measure of the direction and magnitude of the relationship between two variables.

Pearson’s r is also known as a zero order correlation because the influence of zero other variables is
accounted for.

Introduction to Partial Correlationsi

Partial correlations involve more than two variables. Thus, we will discuss partial correlation using the
variables X1, X2, X3, etc.

Unlike the zero order correlation, partial correlations control for the influence of one or more other
variables.

A first-order partial correlation is a measure of the relationship between X1 and X2 while controlling for
one other variable: X3. This is noted as and computed using the following equation:

√( )( )

A second-order partial correlation is a measure of the relationship between X1 and X2 while controlling
for two other variables: X3 and X4. This is noted as and computed using the following equation:

√( )( )
Examples

Given , , and , compute the correlation between X1 and X2 while


controlling for X3. N=266 for all

We are looking for the relationship between two variables while controlling for one other, so this is a
first-order partial correlation.

√( )( )

( )
√( )( ) √

The correlation between X1 and X2 while controlling for X3 is .8000

This is a strong positive correlation. However, you can see that the correlation between X1 and X2
dropped from .87 to .80 after the third variable was taken into account.

Given , , and , compute the correlation between X1 and X2 while


controlling for X3. N=40 for all

We are looking for the relationship between two variables while controlling for one other, so this is a
first-order partial correlation.

√( )( )

( )( )
√( )( ( ) ) √

This is a moderate negative correlation. Again, the magnitude of the relationship between X1 and X2
decreased from -.70 to -.5678 after the third variable was controlled for.
Given , , , , , and , compute the
correlation between X1 and X2 when controlling for X3 and X4. N=506 for all

We are looking for the relationship between two variables while controlling for two others, so this is a
second-order partial correlation.

√( )( )

( )
√( )( ) √( )( ) √

( )
√( )( ) √( )( ) √

( )
√( )( ) √( )( ) √

( )
√( )( ) √( )( ) √
Statistical Significance

The statistical significance of a partial correlation is computed using a t-value:

√ where θ is the order of the correlation

Thus, for a first-order partial correlation, √

For a second-order partial correlation, √

Examples

Let’s compute the p-values for the previous examples. For each, test the null hypothesis that the partial
correlation equals zero at the .05 level.

, N = 266

ii

There is evidence that the correlation between X1 and X2 while controlling for X3 is different from zero.

, N = 40

( )√
( )

There is evidence that the correlation between X1 and X2 while controlling for X3 is different from zero.
, N=506

There is not sufficient evidence to state that the correlation between X1 and X2 while controlling for X3
and X4 is different from zero.

To Do

 Complete the “Partial Correlation” Practice Problems (optional)


 Watch the “Partial Correlation in SPSS” Video

i
All formulas assume listwise deletion.
ii
An online p-value calculator was used to compute an exact p-value: http://easycalculation.com/statistics/p-value-
t-test.php

You might also like