The Dirty Secret of Trends!
Or, how I missed teaching some maths on a sidewalk.
Here’s the story.
It was at the corner of Stevens Boulevard and Santana Row in San Jose, along with a group of petitioners asking why we should refuse salty Cubans from the Sea and accept crispy Cuban from Mexico?
Because they are soggy wasn’t the answer.
At about that time, a college student challenged the petitioners.
Do you know what Communism is; he asked?
Can you even say what’s wrong with Socialism, he piled up?
That was as unlucky as throwing a soccer ball at bored kids and trying to get it back. So, it ended predictably with
You guys never listen anyways!!!
There is a reason why Socialism doesn’t work despite its actual merits, and it is probably the exact reason why that same scenario plays over and again in the corporate world, in our families, sometimes even in our own self-talk.
The reason has to do with Professor Karl Pearson and 19th-century statistics; it is mind-bending, and I think you’ll love it.
Hurried proponents of Socialism today still rely on correlation, as defined by Pearson, to identify marginalized, oppressed identities, despite the last twenty years of research done by Professor Judea Pearl and his students to compute causality from correlations and lack thereof.
Statistical graphs like the one below illustrate correlation. Here, you see the trend (0.9999) between the number of aggravated assaults versus the number of police officers (all per 10,000 habitants) in 108 major US Cities. The complete data is available from the US Census Website.
The bottom axis of a statistical graph is not a scale; it’s a process, the process by which you select US cities.
If you remember one thing from this blog post, let it be this:
Change the process, change the trend.
How bad is that?
If you pick from any of the cities from the US Census data, the number of aggravated assaults trends at 0.9999 of the number of police officers (per 10,000 habitants). Roughly one new crime for each new police officer.
If you limit your search to cities with fewer than 40 police officers per 10,000 inhabitants, the trend that you contend with is now 50% higher.
But there is more, the trend for cities with 40 or more police officers per 10,000 habitants is about 0.4, or half the global trend.
The 2009 US Census data suggest that a police force reduction to below 40 per 10,000 inhabitants may increase the number of aggravated assaults by 20% to 30% in large cities.
Learning to read statistical graphs is no longer optional.
And, what’s more…
it applies to your business life too.
Joel
