# Causality VS Correlation [duplicate]

When I have read this article 'http://www.analyticsvidhya.com/blog/2015/06/establish-causality-events/?utm_source=FBPage&utm_medium=Social&utm_campaign=150725' i still don't understand the difference between difference between Causality and Correlation?

I understand about causality but not correlation. Maybe you have any concrete example?

Thanks!

Correlation means two variables have either a positive or negative relationship with each other (e.g., when one goes up the other goes up) but does not tell us anything about if one variable caused the other to go up- which is what causality aims to answer.

For example, there's a correlation between having a cigarette lighter in your pocket and lung cancer. But the cigarette lighter did not cause the cancer.

Instead, smoking (which is correlated with having a cigarette lighter) has the causal link with lung cancer. Causality ultimately answers the question, if we kept all other variables constant in someone's life (even keeping that cigarette lighter in their pocket), but changed their smoking habits would their outcome (lung cancer) have changed.

• And you can also have causation without correlation.
– mkt
Jul 6, 2016 at 7:48

I always explain correlation vs causation to laymans by showing them the following website:

Spurious correlations

This website shows data that is correlated, but to imply that any causality is part of these graphs is simply absurd.

An example: