Precision Teaching In A Fuzzy World

Ever since I wrote the post "Do You Only Get What You Measure", I've been thinking about the impact of precision teaching and measurement as detailed in the links provided by PalisadesK . OT, but I would like to see an entire post sometime on where PalisadesK gets the time to find this stuff. If you're reading this, let me just interject here that you are an amazing resource, sort of like the Wikipedia of education without the half truths. Kudos!

Anyway, I've struggled with why such a common sense, obvious truth (precision measurement and teaching) seems to hit such disproportionate resistance to any such implementation by the education establishment. I'm so old that I have all sorts of ancient oblique anecdotes to sift through, and as I sifted through my old warehouse of stories, it hit me. I've been through this exact scenario before. I just never recognized the connection.

In the late seventies I was a test engineer for a computer company. In those days testing was pretty much an end of the line event. Hundreds of people built stuff, then threw it over the wall to a test system that found defectives and fixed them before shipping to customers. I'm over simplifying of course because there were always interim tests of subunits, but the important point to this discussion is that testing was, culturally at least, a post manufacturing event. Defects were intentionally passed on to others to resolve.

As products grew more sophisticated it became more and more difficult to build a high percentage of good units on your first try and the hypocrisy of passing on defectives became a problem. Eventually, most of what was built in its first pass was in fact bad. Simple probability worked against us as circuits went from tens of thousands of components to tens of millions of components. The cost of test machines skyrocketed to over ten million dollars a copy and the programming costs to bring one of these beasts on line soared along with the purchase price. A point was reached where it was physically impossible to test-in goodness as a post build event.

The forces of stasis in this culture were huge. Entire buildings had been erected to sustain this post test culture but in private enterprise stress causes change. You change or die without a monopoly position. In order to address this increasing complexity, testing evolved from a post build culture to one that was integrated into manufacturing. Instead of testing products after they were assembled you tested people, parts, machines, and processes before and during the process of putting them together. We found that if you perfected the front end you didn't have to test on the back end (as much) and ultimately some products had such high yields that if they were found (at some final system check) to be bad, this was such an infrequent event that it was economically unsound to fix them. Test engineering went from 'find and fix' to 'find and prevent'. This was an enormous paradigm shift that took a decade to complete.

In education we have a very analogous paradigm shift to work through if we want to get to precision teaching.

Today, we have a 'post build' culture. Kids are tested after being 'built' to see if they've gotten it or not. Where this process gets dicey is when it bumps up against the physical infrastructure of the ordinary school. Schools are a collection of rooms. There are 1st grade rooms and 2nd grade rooms and on and on. There are 1st grade teachers and 2nd grade teachers. There are 1st grade students and 2nd grade students. You get the idea. The infrastructure is entirely designed as a sequential set of processes in an assembly line with the kids passing though the system like little computer parts .

Once you understand this locked in, sequential infrastructure it's easy to see why you can't teach with precision. Teaching with precision creates 'rejects'. It creates students with identified deficiencies that need to be fixed. The most effective (for the status quo) testing system in this infrastructure is one that leaves deficiencies in a fuzzier state, otherwise how can you justify passing the 'part' down the line to the next assembly station. Only fuzzy measurement can possibly support this sequential structure. This is exactly like the old computer manufacturing. It knowingly passed defects down the line to the next step in the process. In this culture you don't hang a sign on the defective parts telling the world you just built junk. You just don't mention it.

Precision teaching would blow up this infrastructure!

Precision teaching tests to find out what to do next, not to produce a grade. So if you want to measure and teach with precision, the implication is that you must also build an infrastructure that is responsive to what you learn in your measurement. To navigate a paradigm shift from today's post test culture to a pre test culture, a culture where testing is designed as a prerequisite screen for 'the next step', the infrastructure has to have components equipped with a finer edge than the blunt sword of grade levels, teachers in rooms, and monolithic curricula.

The challenge for precision teaching is to devise an infrastructure to support it. If you do the testing right and uncover lots of little defects that need to be remedied before passing on the 'parts', then you need a structure that has a great variety of paths that can accommodate what you uncover with your precision. Instead of a sequential set of processes you need a web of processes that can provide for all the twists and turns that are the inevitable detritus of human interactions.


I'm convinced that precision teaching can not happen without fundamental change to the infrastructure and it is fear of this fundamental change which creates the illogical resistance to precision measurement and thus precision teaching . It's far easier to pretend that Johnny can add and pass him on, than to prove he can't and then have to do something about it.