Hello there bloggies! Hope you all had a great Halloween and your extra hour of sleep for Daylight Saving Time.
Welcome to Science Sunday!
By now, many of you would have seen the headlines that bacon and other processed meats cause cancer. Of course, with the Western world’s love of bacon, these headlines made the rounds through popular and social media. There are so many examples of similar headlines, many of which include chocolate and red wine and their effects on human health.
A common problem with these headlines is that these over-simplifications are based on observed associations between certain factors and human health and disease. So, today I wanted to talk about what an association actually means and what it doesn’t mean.
An association is simply a statistical relationship between a factor like bacon consumption, amount of exercise, or even a genetic variant with an outcome like cancer diagnosis, performance on cognitive tests, or longevity. That is, variation in the factor has a relationship with variation in the outcome. That’s all.
This statistical relationship could be strong or weak, and still be a statistically significant association. For example, many genetic variants associated with human traits from GWAS (check out this previous post on GWAS!) are strongly associated with really really small effects. These associations are true, but they don’t have major effects on the traits being studied.
A statistically significant association does not mean that a factor causes an outcome. It only means that there is a statistical relationship between the factor and outcome. Much more evidence is needed to show that a factor directly causes an outcome. Direct, controlled experimentation and randomized clinical trials are way to show that there is a causal relationship between factors. To be overly cliché, correlation does NOT mean causation!
Something else that association does not set is temporality. An association may occur from the factor happening before the outcome or as a result of the outcome. Again, an association does not necessarily mean that factors cause an outcome!
Thank you for visiting Science Sunday! Next time you see a headline claiming that something causes cancer, heart attack, or stroke, ask yourself whether the evidence is just based on associations or if they have causative evidence. Let me know what you think in the comments below or on twitter @DrFsThoughts!
See you on Tuesday!