Reading Across Disciplines

I'm a firm believer in reading across disciplines.  I have a social science background (anthropology / archaeology) and I come from a family where my father was a statistician / sociologist and my mother was a behavioral geneticist / developmental psychologist.  So it's quite the social science melange.  But I read a lot of history, politics and science books.  But my obsession recently has been

I'm currently spending a lot of time reading economics, specifically macroeconomics, because I'm fascinated and repelled by the economic dialog in our country (fascinated that the smart people, like Krugsandra are ignored while simultaneously repelled that twits like Tim Pawlenty or Paul Ryan are taken as Serious Economic Thinkers™).  I'm also obsessed with the soverign crisis in Europe, Greece in particular.  But I digress.

What I learned in class today was Goodheart's Law.  It's so obvious as to be comical, but it applies to every day life.  Here it is in a nutshell:

Any observed statistical regularity will tend to collapse

once pressure is placed upon it for control purposes.

Any metric you use to exert control over a system will, in turn, impact that metric adversely.  Or, if you measure it, you will tend to achieve your targets rather than conform to the underlying principles that were the reason to have the targets in the first place.  An example from my own workplace is in order.  More below the fold.

Last year it was determined that it would be in the company's best interest for everyone in our department to grow their technical skills in some manner.  We're an IT shop so "Putting the T back in IT" is not a bad thing.  A number of us senior technologists had been encouraging this for years.  After all, you only need so many project managers before you'll need people to do actual work.

So did we simply throw it out there to the employees? You know, work with your manager, build a plan and get trained.  No, of course not.  Here's how we did it.  We established a technical training target (20 hours) for everyone in IT.  We then wrapped that metric in a scorecard and used the scorecard to beat people up for not achieving the target.  We tracked and measured and issued reports.  We made sure that everyone was aware of their obligation to get trained.  Rather than treat the technical training as a benefit and an opportunity to encourage employees to improve their skills, we turned it into a compliance exercise.

So guess what happened?  All kinds of things started qualifying as "technical training" that were neither technical nor training.  Managers, measured on the performance of their employees conformance to the targets, approved a wide range of non-technical non-training.  Our criteria for "technical training" deteriorated to the point of utter silliness.  

• Attended a vendor presentation?  Technical training! 
• Watched a video on PBS about computers? Technical training! 
• Sit through a lecture on cloud computing? Technical training!  

Pretty much anything you could construe as vaguely technical and vaguely educational would qualify.  Meanwhile, those of us who had encouraged the business to focus more on the T in IT started looking like this in the meetings:

Instead of measuring people on attending actual technical training, we created a metric that was easily circumventable and everyone got "trained."  Then, being the company that we are, they pulled the target in from December to October and said everyone had to be done by October, because that's how we roll.  See how good we are?

Our measurement system destroyed the purpose of assigning the training.  Special.