Monday, September 22, 2014

Obvious relationships are often illusions



I am going to use a controversial example so that I can get everyone mad enough to comment.  But as a strawman, the concept behind this message is really important to get, so please don’t let belief-resonance prevent you from seeing it.

We know that one of the strongest predictors of whether an adult might commit acts of domestic violence is whether they experienced them as a child.  This can be either as a victim of child abuse or as a witness of spouse abuse.  Unfortunately, it seems to be a self-reinforcing cycle of violence.

We also know that low income families are more likely to have acts of domestic violence.  I am not going to guess at the reasons, but it is a very unfortunate reality.

We also know that the NFL has a greater representation of people who grew up in low income families.  They are hardly low-income after a few years as overpaid pro athletes, but this doesn’t seem to break the cycle of violence I mentioned earlier.

If you control for this chain of factors, NFL players do not commit domestic violence any more than the general population.  It is not due to the violence of pro football.  It is not due to the culture of acceptance that has emerged in the league. It is not the pathetically weak response of league officials.  The origins happened much earlier. 

It is easier to visualize the more direct and concurrent effects of these other variables.  That makes them jump to mind much easier and stronger.  But . . .

Of course, the reverse might be a stronger argument anyway.  A strong culture against domestic violence could counteract and perhaps break the chain of domestic violence.  Stronger responses from teammates, coaches, teams, and the NFL central office could do a lot towards preventing it.  Better awareness through the kinds of programs that are just now being proposed could nip tendencies right in the bud.  So these things all should be done.

But when looking for the original cause, attributing it to these things is an illusion caused by representativeness, confirmation, availability, salience, and other cognitive processes that frequently lead us wrong when looking for cause-effect relationships.

Now back to my first caveat – I am using this as a visceral, controversial strawman.  I have seen some evidence that this is the case, but not enough to be sure. On the other hand, there is also not enough data to say that the league (culture, behavior, or otherwise) is the cause.  My point is simply that it is easier to jump to the conclusion for proximate causes than it is to look for precursor variables and root causes.