Friday, May 20, 2011

Correlation does NOT equal Causation

In the area of research, there is a distinct difference between a correlation between two factors/events and a causation between them. Even when just googling "correlation definition" the word "causal" shows up in many definitions. In research, however, the two terms are not interchangeable.

A correlation is merely two things that within the study both changed either in the same direction (positive correlation) or in the opposite direction (negative relationship.) However, this does not imply that one causes the other. In many cases, there is an unknown third variable that both factors/events are actually causally related to instead. This type of relationship is seen most when a variety of information is gathered and analyzed.

A causal relationship is one in which one variable is shown to cause a change in the other. This can be demonstrated by an ABAB design, for example, where a variable B is measured, then variable A is changed (B is measured again), A is dropped back to original level (B is measured again), and A is changed again. If B changes along with A, there is strong evidence that changes in A cause changes in B. This type of relationship is seen more in experiments.

0 comments:

Post a Comment