The New York Daily News reports on an observational study by the Harvard School of Public Health and concludes:
“Men should back away from the bacon if they want to make babies.”
Here’s the nut graf – no pun intended:
“Researchers analyzed more than 350 semen samples from 156 men who visited a local fertility center and answered questions about their diet. They found that eating processed red meat has a negative effect on “sperm morphology” — the size and shape of sperm’s cell structures. Having abnormal sperm can contribute to infertility.”
No, they did not find that. You can’t “find that” by having men fill out a questionnaire and then comparing it with their sperm size and shape. You may be able to point to a statistical association – “we found more men who ate more bacon had more screwed-up sperm” – but you can’t prove a negative effect by this kind of study. No cause and effect. Just a statistical association.
And, in the way online news is often assembled these days, note how the Daily News story links to past “related” news:
Not necessarily related: ASSOCIATION and CAUSATION.
The Daily News was not alone on this. A web search turns up about 100 stories on the study, with headlines such as:
But why couldn’t every story include a simple caveat from a researcher who said, in a Huffington Post story:
“”We’ll continue looking into this question including the possibility that it is not processed meats that is driving the association but what they are replacing.”
Or, as the Boston.com Daily Dose reported:
“…the researchers could only make statistical associations and couldn’t prove that exercise or dietary habits directly affected sperm production.”
Between bacon and oreos, it has been an awful week for journalists purportedly reporting on studies.
For what feels like at least the 100th time, we remind people that we offer a primer, “Does the Language Fit the Evidence? – Association Versus Causation.”
Follow us on Twitter: