kitchen table math, the sequel: The Tyranny of Average

Thursday, February 19, 2009

The Tyranny of Average

Whenever I hear about moves to create national standards I'm reminded of the tyranny of averages. Averages are obtained by a simple algorithm that reduces the great cacophony of human endeavor to a cipher. All things human have a distribution and if you're measuring a performance you can be certain of results from awful to superlative.

If you're setting a standard for performance it will seek to find this middle ground of average capability. Set it too low and you leave money on the table. Set it too high and you've created an unatainable, frustrating goal. Even if you attempt to set it as some sort of forcing function, to set a high bar, you still must be cognizant of the middle.

What, then, happens to performers in a system where the goal is changed to average? Well the people who are awful, remain so. Absent some other change to the stage, a new goal doesn't have the juice to change an awful performance. The people on the other end of the spectrum, the ones whose excellence brought the average up in the first place, will likely lower their performance. They're the only demographic that has the unfettered ability to move to the new goal.

The new average, the one that kicks in after a few years of leveling, will be lower than the original. The high end has moved down. The low end stayed in the same place and the middle probably didn't know anything was happening anyway. This is the tyranny of average.

Without firing a shot or forcing any unpleasentness, the frog gets lowered into the cold water to start its cooking!

I'm reminded of Catherine's comments on the 'convergence of interests' when I think of this effort to create a national standard. Put another way, you could say that the convergence is also a 'synergy of entanglements'. A national standard would be a standard gleaned from national consensus (an average) brought to you be a cabal of vested interests. And, even though each participant in the great leveling may have individual pursuits in mind, they all serve the great convergence. They all strive to keep the beast alive because it is the beast that butters their toast.

I've been entangled in many start up companies and was priviliged to see a few grow to greatness. Most are now gone. They replaced their entrepreneurial, chaotic (successful) beginnings with the stultified average of (failed) convergence. We're in the midst of such a process right now in our politics. We seem unable to embrace failure as the cauldron where future success is born. Instead, we are scrambling to make sure that nobody fails. From the executive taking his gold from the burning plane, to the scofflaw who mortgaged a house they couldn't afford, we are doing everything possible to see them reach the tyrannically average American dream.

Our school systems are a mess and they are jam packed with entanglements. On whole it is a gigantic entangled, convergent, average. Instead of codifying this monster with national standards it should be decentralized down to a few million Darwinian enclaves to see which one is most fit.


RPondiscio said...

At the risk of oversimplifying, I think it's possible to eat our cake and have it too. Here's how: national standards (content standards, please) ought to represent a statement of what we expect our kids to learn and know in school. What if we married those standards to national assessments, with reading comprehension tests tied to those content standards. In other words, the selections on the tests would be culled from the content standards, thus making it a test worth teaching to. Finally, tie federal funding to states adopting the standards and tests with NO SANCTIONS WHATSOEVER based on performance. Thus the federal role is limited to spreading sunshine--a pure apples-to-apples comparison among states, districts and schools. If everyone is taking the same test, it'll be pretty clear who is performing well and not, and up to states and districts to improve their performance. If you're serious about "decentralizing down to a few million Darwinian enclaves" as you put it, then the only way to credibly gauge performance and unlock what works is if everyone is shooting for the same target.

Anonymous said...

Standards don't stand in isolation to the rest of the Gordian knot. The way they're written can drive whether or not your curriculum spirals. They can drive pedagogy, phonics vs. sight words, for example. They can drive calculator usage, or not. Whosoever shall write the standards, shall tie the rest of the knot.

These things drive text book selection, which is driven by the ed schools and professional organizations with skin in the game. Don't forget the publishers and consultant community. Lots and lots of players converge on standards.

I would pose another question, perhaps more thought provoking... Why should the feds give a rats patootie (is patootie a word?)how a school performs? OK, it's rhetorical! They care because they've got their $$$ nose in the tent.

Why is their nose in the tent? It's not a constitutional mandate is it? I think the only people that should care at all are the parents. The objective measure (over time) is; are the graduates successful in their post graduate pursuits? This is far more important than a contrived test.

Until parents own the measurement, and take responsibility for the cost by paying for and having the freedom to not pay if the measurement flops, the camel wins every time.

By advocating for, and accepting federal or state money, schools and parents are making a pact with the devil. They are allowing a convergence of interests (not necessarily aligned with their own) to drive their children's future.

I guess I'm just not convinced that we're collectively smart enough to write a 'standard' that fits millions of kids to a tee.

RPondiscio said...

Point taken. Bad standards are worse than none. But parental interest is a strong reason to support nationalized standards and assessments. As a parent, it's impossible to know whether your child's school is terrific, atrocious or somewhere in the middle unless it's measured by the same yardstick as all the others.