Correlations

Jason Ryan, MD, MPH

Correlation
Coefficient
A correlation analysis uses the correlation coefficient to
describe the linear relationship between 2 quantitative variables,
Pearson Coefficient as in a study evaluating the linear relationship between hours of
sleep and irritability score.

Lifespan
Pack-years of smoking

So Correlation coefficient attemps to describe the relationship between smoking and

lifespan and to describe how much of lifespan is associated with pack years of smoking .
And the way you calculate a correletion coefficient is by fitting a line through the data.This lines
is mathematically calculated to minimize the distance between all the points on the graph and
the line
Pearson Coefficient

• Measure of linear correlation between two variables

• Represents strength of association between two variables
• Number from -1 to +1
• Closer to 1, stronger the relationship. 0 means no relationship

• (-) number means inverse relationship

• More smoking, less lifespan
• (+) number means positive relationship
• More smoking, more lifespan
Correlation
Coefficient
Pearson Coefficient

Lifespan

Pack-years of smoking
The more scattered the data, the further away from 1 the coefficient will be. In the extreme
case, when there is no relationship between two variables, the data look like this:

Note that there is no upward or downward trend; as the value of the

x-axis rises, there is no consistent effect on y-axis values. This
indicates no relationship between variables. The correlation
coefficient, in this case, would be zero.
Correlation
Pearson Coefficient
Coefficient
• Measure of linear correlation between two
variables
• Represents strength of association between two
variables
• Number from -1 to +1
• Closer to 1, stronger the relationship
• (-) number means inverse relationship
• More smoking, less lifespan
• (+) number means positive relationship
• More smoking, more lifespan
• 0 means no relationship
Correlation
Coefficient
Pearson Coefficient
Strength of Relationship

r = +0.5 r = +0.9
(stronger relationship)

The correlation coefficient is denoted by a lowercase r

In the first image there is a lot of scatter in the data
 The correlation coefficient is
denoted by a lowercase r
Strength of Relationship In the first image there is a
lot of scatter in the data

r = +0.5 r = +0.9
(stronger relationship)
d
Direction of Relationship Here ,the
relationskip is
inverse, so the
more you get of
the variable on
d d the x axis,the les
ypu have of the
r = -0.5 r = +0.5 variable on the
Negative Positive r=0
No relationship y-axis
Correlation
Pearson Coefficient
Coefficient
• Studies will report relationships with correlation
coefficient
• Example:
• Study of pneumonia patients
• WBC on admission evaluated for relationship to LOS-lenght of stay
• r = +0.5
• Higher WBC  Higher LOS
• Sometimes p value is also reported
• P < 0.05 indicates significant correlation
• P > 0.05 indicates no significant correlation
Coefficient of
Determination
r2

• Sometimes r2 reported instead of r

• Always positive
• Indicates % of variation in y explained by
x

r2 = 0.6 r2 = 1
(60% variation y is explained by (100% variation y explained by x)
Example this would say thatx)the 60 percent of the variability in length of stay is explained
by the changes in the White blood cell count on the admission, the remaining 40 percent of
the variability in length of stay is explained by other measures