Is Correlation Always Misleading?

How often does correlation get mistaken for causation in data analysis?