Newspapers provide their advertisers with lots of reader data.
My first apprentice-level research job out of grad school was with a Scripps-Howard newspaper in Memphis, Tennessee. One of my assignments was to seek correlations without regard to causality. The goal of this research was not to learn, but to find promotable items of interest to advertisers. This strategy is an old one, and very successful, too.
Correlation and causality are among the most misunderstood of all research terms, and the ones most likely to be misused, especially by marketers and the media.
Correlation is no more than some sort of undefined connection between things. Causality is the more complex cause-and-effect.
Correlation is very popular with Big Data, where automated tools search for relationships. Blindly groping in the dark is another way to say it.
Cause-and-effect is applied incorrectly more often than not. Another term is jumping to conclusions.
Red cars are twice as likely to be in accidents than blue cars.
The obvious conclusion for some is to buy a blue car. The question researchers would ask is this: “Is it cars that are in accidents, or people driving cars?” Is it possible that people who buy red cars are more aggressive drivers and less careful? Sure, and that’s only one possible explanation of many.
A friend bought a motorcycle.
HIs mother had read most motorcycle accidents occurred during the first six months of ownership. Unaware that those accidents were not related to the calendar (correlation), but to the new rider’s inexperience (causation), her advice to him was to put it in the garage and not ride it for six months.
And don’t forget, if you can’t get divorced without being married, marriage must cause divorce.