When experts collide!
I’ve been reading a book called Humble Pi written by mathematician, Matt Parker, about real-life failures caused by bad maths. Did you know that Enron’s downfall was probably caused by a dodgy Excel spreadsheet where someone had =sum 2 numbers rather than =average, causing investment risk factors to be underestimated?
Anyway, towards the end of the book he makes a quite profound statement along the lines of:
“…statistics can only signpost you to the changes in something – the what. You then have to go and find out why those changes have taken place. Statistics won’t give you the why.”
This is quite a weighty idea if you think about it long enough (as a dedicated night time over-thinker, this is exactly what I did). I then got to thinking about another of my fave books – Freakonomics. OK, so I know it was written a while ago now, but the core themes are still relevant and the way Levitt and Dubner apply their formal academic thinking to real world problems shows how maths, stats and science-y stuff is all around us.
In the intro of the book they talk about New York and the Broken Window policies that were enacted at the behest of (now discredited) Mayor Rudy Giuliani to reduce crime. At the time the Big Apple had seen a steady and alarming increase in crime that – if predictions were to be believed (which they were!) was going to end up in a major world city becoming a lawless dystopian no-go area.
Wow! The predictions were sobering. And then the policies were introduced. Every Tom, Dick and 14 year old street corner drug pusher was arrested. No potential murderer was left unbooked. There was nowhere safe for those with criminal intent to hide.
And then the crime rate started to fall, most notably violent crimes.
Imagine the volume of back slapping, emotional man-hugs and willy waggling that ensued. They’d done it. All it needed was people to be tough and show these people (what…these people who were living in dire situations, surrounded by deprivation and hardship with few life choices open to them…is it those people we’re talking about?). The more self-focused and less-smart smart people concurred. All you had to do was take a tough stance on all crime, no matter how petty, and then it would all go away.
And then some more-smart smart people came along and were all like “yoo hoo! Mr Mayor, we don’t think this is right.”
Remember kids – correlation does not mean causation. Just because violent crime was reducing, does not mean it was the zero tolerance programme that was causing it. So what did the more-smart smart people find when they did some proper background research?
Well back in 1973 there was a landmark legal case that was given the designation Roe v. Wade. You may have heard of it over recent years because it was the ruling on this case that gave women more opportunities for managing unwanted pregnancies, which was subsequently overturned. What the more-smart smart people did was realise that the outcome of this ruling opened up the way for women to terminate pregnancies when they felt that they could not support the child properly.
It’s easy to make sweeping and unfair judgments about this, but the fact is that at the time women would find themselves pregnant and would have no options available, other than have the baby and either try and support it (often in poverty) or put it into the social care system. Apart from these not really being choices, they have the effect of creating generations of dysfunction.
When these young people, with few life-choices and living in dire situations were old enough, they would naturally segue into crime, because they were surrounded by it. It was how you ran your life and put food on your table. Then Roe v. Wade came along and the number of babies born dropped which had a direct relationship to the number of young people available to commit the crimes that were plaguing New York.
The broken window policies may have had some effect on the crime rate; they certainly had an effect on the horrifically overcrowded (and corrupt) penal system. However, they were not responsible for the dramatic downturn in crime rates…
…a legal battle which picked up the right of women to manage their own bodies, made 20 years earlier in a completely different part of the US, was responsible.
This is one heck of a story. I didn’t understand the Roe v. Wade debate until I re-read Freakonimics and realised it was the intro. When I originally read the book, the ruling hadn’t been overturned and it was a case that had fallen into obscurity. What will be interesting is to see whether crime rates start to increase across the US sometime around 2040, which is when the next batch of disenfranchised young people, with few life-choices, reach maturity.
What it does tell us is that it’s worth digging into what your data is telling you to see if there are any hidden truths. Things that might not be obvious but could give you ideas of things you can do to increase engagement with supporters or raise more income.
What we do want to be sure to avoid when we’re looking at our data are the two downfalls of the less-smart smart people: 1) correlation rather than causation and 2) confirmation bias. We are the experts and we can make sure you don’t fall into those traps!
