Sowing Confusion and Insecurity

The more I learn about No Child Left Behind, the more discouraged I become.  There is no way that anyone who passed this legislation could believe that the goals are actually achievable.  I came to the conclusion that the law was written so that public schools can be branded as "failures" and lend support to the voucher and privatization movement.

Firstly, how do we ensure that ALL students perform at the proficient level without pushing out special needs kids? 100% proficiency is just not possible, unless we lower proficiency standards to insultingly low levels (as some states have done) or pushing out low-scoring kids?

I am not surprised that only 26% of California public schools are making adequate yearly progress (AYP) and that the school my daughter attended last year is among the 74%.  Failing to meet an arbitrary and stupid rule is not a failure.

What surprised me is how the law binds the hands of teachers and administrators to create maximum panic, confusion and insecurity.

We received a letter in the mail from our school district saying that our neighborhood middle school failed to meet their NCLB targets for all groups of students; they could offer as a transfer to a farther school within the district that had met all the targets.

The letter did not explain which group of students did not make AYP or whether our child is among them.  That's pretty basic information, which a parent might want to know, right?  Well, the law forbids them from telling you.

That's right.  Schools have a long laundry list of targets that they must meet, but they can't tell you which ones they failed to meet.  All you can see is a big fat failure label on the school.  If they wanted to design a law to sow panic among the widest set of people, so that they try to pile into life rafts (aka a handful of school districts with low to no special needs kids), then they've succeeded.  Why else would you withhold this information from parents?

There is one defensible reason to keep mum, that of not blaming or stigmatizing the group(s) that failed to meet the NCLB targets.  I sympathize with that argument, and I dithered about going into specifics on our local situation.  But, I was at a PTSA meeting in which parents were encouraged to inform their neighbors and allay their fears.

I'm going to tell you that, if you are reading this blog, chances are slim that your child falls in the group that "failed" to jump high enough on command.  The chances are slim even among non readers, just 17 out of 915.  Actually, it's lower than that because not all of the 17 kids in the group missed their targets.

I'd also like to make it clear that I am not blaming the kids.  I am blaming the law and lawmakers who must be either too stupid to understand statistics or too craven to care.

Background on the Statistics of Small Numbers:
When you draw a sample from a population with a known standard deviation, the variance of the sample mean goes down with increasing sample sizes.  Intuitively, that makes sense because, as you increase the sample size to the entire population, your variance is for the entire population. [Equation excerpted from this excellent psu.edu tutorial.]

Alternatively, variance increases with decreasing sample size. That's why groups of 10 or fewer are exempt from AYP targets.  Does something magical happen when n jumps between 10 and 11?  Look at the equation above and draw your own conclusion.

Now take a look at  the group that "failed" in 2012 and "failed" again in 2013.  Do small sample sizes of 17 and 18 jump out at you?  If not, they should.

I pored over the statistics from 2011, 2012, and 2013.  The train wreck started when a small group  of English learners in one grade did spectacularly well.  They nearly closed the gap with English proficient students!  But, under NCLB, they needed to "build" on their success and score that high PLUS an additional 10 points.  Yup, they need to make uniform progress of 10+ points every year, regardless of starting position.

Anyway, the average score of the small sample regressed to the mean.  They still did above the state-wide mean for English learners.  In addition, the average score for all of the kids at the school was well above the average for all types of California students.  This school is no where near failing in their mission to educate all kids.

The teachers gave practice tests to the kids and saw that they were not going to make it.  The school decided that burdening the kids with more test prep was counterproductive and would detract from their real job of teaching.

Call us failures, but that's a decision I fully support.

* I've coded up a simulation to demonstrate the statistics of small numbers.