Correlation vs Causation Errors

It's tempting to believe that researchers know how to effectively draw conclusions from survey data – especially researchers that work for Universities or research institutions. Most take pride in their work, and have years of experience with the craft, so it's believe that to work in the research world, one must have a basic understanding of what data really means. You could think this, but you would be wrong. It seems that no matter how much experience someone has in the research world, there is still a temptation to draw causational conclusions from survey data in a way that creates false facts that supports incorrect arguments.

Perhaps the best examples of these types of conclusions come from studies about relationships. While it's possible to quantify behaviors and develop research hypotheses in a lab, ethics prevents people from studying things like personal relationships in a cause/effect setting, so surveys become the best way to collect and analyze data.

But researchers often forget that their results are often correlational, and as they teach in Statistics 101, co-relation does not equal causation. This was a rule apparently forgotten by a study that came out in 2011. These researchers found that men rated cuddling an important factor of their relationship if they were in a happy relationship than if they were in an unhappy relationship.

These researchers – followed by numerous media outlets, including ABC News, CBS News, and even WebMD, then reported that men need cuddling to be in a happy relationship.

Do men enjoy cuddling? Probably more than they admit.

But do men need cuddling for their relationship to be more satisfying? Probably not. The researchers found a co-relation between cuddling and happy marriages because, quite frankly, there isn't an unhappy couple in the world that really wants to cuddle their partner.

Consider this Relationship 101 – If you don't like your husband or wife, you probably won't want to cuddle with them or find cuddling enjoyable. It's not that complicated to understand. Yet hundreds if not thousands of news media outlets incorrectly reported that the results of the study proved that men need cuddling to be happy in their relationships.

Trained researchers drawing incorrect conclusions like this one are sadly not that uncommon, and no matter how much you believe that you understand survey research best practices, you need to always remember that your data may not prove what you think it proves, so keep an open mind and remember that correlation will never guarantee causation.