Correlation and Causation

Correlation and causation are two important
concepts.   It is important to understand the difference
between correlation and causation.   Two variables can be
correlated without being causally related.  Thus, the use of
the word "correlation" does not mean causation.

What is Correlation?

It is important to have an understanding of the
definition of correlation.  With respect to research findings,
if there is a significant correlation between two things it is
merely stating that we can predict one thing from the
other.  Thus, the correlation definition with respect to
scientific research is simply
a finding that one variable
predicts another variable.

Correlation Interpretation

If A and B are correlated, it could indicate that A
caused B, B caused A, or a third variable caused both A
and B.
Two things can be correlated without there being a
casual relationship
.   Thus, in some situations, correlation
is not causation.

Correlation Examples

There are many possible correlation examples.  One
example of a correlation is the association between a
supervisor's perceived humor and job satisfaction.  Decker
(1987) found that a supervisor's perceived sense of  humor
was positively correlated with people's job satisfaction.

What is Causation?

It is important to have an understanding of the
definition of causation.
Contributory causation would
reflect the following:  when A is manipulated it results in an
increased likelihood of a change in B.  Thus, for example,
we may observe a difference in the averages between an
experimental group and a control group when A is
manipulated in an experiment.

Types of Causation

There may be different types of causes.   Randomized
experiments allow one to make conclusions about
contributory causation, which reflects that one variable
increases the likelihood of a change in another variable.
With respect to human behavior, sufficient causation
reflects what is sufficient for a behavior to occur.
Also, I think we can make a distinction between single
causation and multiple causation.  Single causation implies
that a variable is sufficient:  it can influence another
variable by itself.   In contrast, multiple causation can be
conceptualized as a combination of variables that are
sufficient as a group, but none of the variables are
sufficient individually.

References

Decker, W. H.  (1987).  Managerial humor and subordinate
satisfaction.
Social Behavior and Personality, 15, 225-
232.