StatLit-Blog

Fighting Statistical Illiteracy

Archive for the ‘Association-Causation’ Category

US Healthcare: Third Leading Cause of Death?

with 2 comments

Consider this extract from a recent article: 200909HealthCareCausesDeath3bSource: www.health-care-reform.net/causedeath.htm

Comment: This study linked hospital patient deaths to events they experienced in hospitals.  But this linkage does not mean that the presence of these factors caused these deaths.  These deaths may be associated with, linked to, or attributed to these factors.  But no coronor or physician said these death were caused by these factors.   Confusing association (due to) with causation (“third leading cause of death”) is an all too common form of statical illiteracy.

These scary statistics pull focus away from the real argument: trying to pass off an association as causation. Journalists use words to get our attention.  Instead of using passive verbs like associated, linked or due to, they use action verbs like ups, cuts, boosts, raises and causes.  Whoever wrote this headline went too far. Saying as a result of goes over the line from association to causation. 

This headline writer needs to go back to school or journalism schools need to do a better job of teaching journalists to avoid statistical illiteracy.

For more on how journalists prevaricate on this distinction, see Distinguishing Association from Causation in News Media Headlines by Schield and Raymond, ASA (2009).

Written by schield

September 22nd, 2009 at 12:00 pm