Scribd
Upload
133 views

Correlation & Regression-I

This document discusses various types of correlation analyses that can be used to assess relationships between variables, including Pearson's correlation coefficient (r), Spearman's rank correlation coefficient (rs), Kendall's tau, partial correlation, and point-biserial (rpb) and biserial (rb) correlations. It provides examples of how to calculate and interpret these coefficients using SPSS and discusses assumptions, limitations, and the difference between correlation and causation.

Uploaded by

drnareshchauhan
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
133 views

Correlation & Regression-I

This document discusses various types of correlation analyses that can be used to assess relationships between variables, including Pearson's correlation coefficient (r), Spearman's rank correlation coefficient (rs), Kendall's tau, partial correlation, and point-biserial (rpb) and biserial (rb) correlations. It provides examples of how to calculate and interpret these coefficients using SPSS and discusses assumptions, limitations, and the difference between correlation and causation.

Uploaded by

drnareshchauhan
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 43

Correlation & Regression-I

Dr Naresh T Chauhan
Community Medicine, GMC SURAT
Research methods

Correlational or
Cross-sectional longitudinal research
The tertium quid

a third person or thing of indeterminate character


Experimental research

an effect should be present when the cause is present and that when the
cause is absent the effect should be absent also John Stuart Mill(1865)

confounding variables
Correlation

Positive
No relation
Negative
covariance

The simplest way to look at whether two variables are


associated is to look at whether they covary.

If two variables are related, then changes in one variable


should be met with similar changes in the other variable.
Therefore, when one variable deviates from its mean we
would expect the other variable to deviate from its mean in
a similar way.
imagine we took five people and subjected them to a certain
number of advertisements promoting toffee sweets, and
then measured how many packets of those sweets each
person bought during the next week.
Graphical display of the differences between the observed data and the means of two variables
Calculating the covariance is a good way to assess whether two
variables are related to each other
Standardization and the correlation coefficient

standardized covariance is known as a correlation coefficient


and is defined as follows:
This coefficient, r, is known as the Pearson product-moment correlation coefficient or
Pearsons correlation coefficient and was invented by Karl Pearson.
bivariate correlation, partial correlation

By standardizing the covariance we end up with a value that has to lie


between 1 and +1
if you find a correlation coefficient less than 1 or more than +1 you
can be sure that something has gone hideously wrong
A coefficient of +1 indicates that the two variables are perfectly
positively correlated, so as one variable increases, the other increases
by a proportionate amount.
Conversely, a coefficient of 1 indicates a perfect negative
relationship: if one variable increases, the other decreases by a
proportionate amount.
A coefficient of zero indicates no linear relationship at all and so if one
variable changes, the other stays the same.
The significance of the correlation coefficient

In the case of a correlation coefficient we can test the


hypothesis that the correlation is different from zero (i.e.,
different from no relationship)
There is one problem with Pearsons r, which is that it is
known to have a sampling distribution that is not normally
distributed.
Second method

The resulting zr has a standard error of:


CRAMMING SAMS TIPS

A crude measure of the relationship between variables is the


covariance.
If we standardize this value we get Pearsons correlation coefficient, r.
The correlation coefficient has to lie between 1 and +1.
A coefficient of +1 indicates a perfect positive relationship, a
coefficient of 1 indicates a perfect negative relationship, and a
coefficient of 0 indicates no linear relationship at all.
The correlation coefficient is a commonly used measure of the size of
an effect: values of .1 represent a small effect, .3 is a medium
effect and .5 is a large effect. However, if you can, try to interpret
the size of correlation within the context of the research youve done
rather than blindly following these benchmarks.
A word of warning about interpretation: causality

The third-variable problem or tertium quid:

Direction of causality- correlation coefficients say nothing


about which variable causes the other to change
Data entry for correlation analysis using SPSS

SELF-TEST Enter the advert data and use the chart editor to
produce a scatterplot of the data (number of packets bought
on the y-axis, and adverts watched on the x-axis).
The general process for conducting correlation analysis
Linearity
Normality check !

Sample size
Modelling
Confidence interval
Exam Anxiety.sav

The main dialog box is accessed by selecting


Using R2 for interpretation

The correlation coefficient squared (known as the coefficient


of determination, R2) is a measure of the amount of
variability in one variable that is shared by the other
However, although R2 is an extremely useful measure of the
substantive importance of an effect, it cannot be used to
infer causal relationships
Spearmans correlation coefficient -rs

a non-parametric statistic based on ranked data


so can be useful to minimize the effects of extreme scores or
the effects of violations of the assumptions
Spearmans test works by first ranking the data, and then
applying Pearsons equation to those ranks.
The Biggest Liar.sav
Kendalls tau (non-parametric)

Kendalls tau, , is another non-parametric correlation


should be used rather than Spearmans coefficient when you
have a small data set with a large number of tied ranks.
This means that if you rank all of the scores and many scores
have the same rank, then Kendalls tau should be used.
Kendalls statistic is actually a better estimate of the
correlation in the population than Spearmans
Biserial (rb)and point-biserial(rpb) correlations

These correlation coefficients are used when one of the two


variables is dichotomous (i.e., it is categorical with only two
categories).
The difference between the use of biserial and point-biserial
correlations depends on whether the dichotomous variable is
discrete or continuous
Discrete-there is no continuum between the two categories
Continuous- it is possible to have a dichotomy for which a
continuum does exist.
pbcorr.sav
CRAMMING SAMS TIPS
Partial correlation
The theory behind part and partial correlation

a type of correlation that can be done that allows you to


look at the relationship between two variables when the
effects of a third variable are held constant.
In this instance we have controlled for one variable and this
is known as a first-order partial correlation. It is possible to
control for the effects of two variables (a second-order
partial correlation), three variables (a third-order partial
correlation) and so on.
Partial correlations can be done when variables are
dichotomous (including the third variable).
The difference between a partial and a semi-partial
correlation
How to report correlation coefficients

There was no significant relationship between the number of


adverts watched and the number of packets of sweets
purchased, r = .87, p = .054.
Creativity was significantly related to how well people did in
the Worlds Biggest Liar competition, rs = .37, 95% BCa CI
[.604, .114], p = .002.
Further reading and exercises, real research examples

You might also like