All things excellent are as difficult as they are rare.

SED OMNIA PRAECLARA TAM DIFFICILIA QUAM RARA SUNT

06 June 2011

Alter v. Ravitch: A Call For Facts

Courtesy of Joanne Jacobs' excellent edublogging, I've been following what seem to be two sides of an argument that's shaping up. It started with Diane Ravitch in the NYT, which Joanne blogged about here. Then recently, Jonathan Alter (who I'm always confusing with Eric Alterman for the obvious morphological reasons) wrote a response to Ravitch, which Joanne blogged about here.

Except it's not really two sides of an argument at all. It's Diane Ravitch presenting an argument (that's probably wrong, I should note), and Jonathan Alter launching a rhetorical diatribe about her recent positions based not on their falsity, but apparently on their being, well, depressing:

While healthy skepticism is a virtue, Ravitch seems bent on extinguishing any hope that our teachers and schools can do better. In an op-ed in the New York Times on June 1, she derided the impressive progress made at three public schools as “a triumph of public relations” based on “statistical legerdemain.”

Yes, but is she wrong?

Also from Alter:
“This was a very cynical statement that she doesn’t believe teachers and schools can make a difference in high-poverty areas,” says Colorado State Senator Mike Johnston, a former teacher and principal whose sweeping tenure-reform law is a national model. “We can debate facts at particular schools but you just can’t deny that some places are getting phenomenal results -- results that should be celebrated, not called out as fraudulent.”

Except that Ravitch does deny it. And just because she's cynical doesn't mean she's wrong. That's not an argument: it's an insult coupled with a flat-out contradiction.

Let's take the case of Bruce Randolph School. Here's what they have to say. I apologize for the lengthy quotes, but I want to get their arguments right next to each other on the page so you can see what's going on.

Ravitch:

True, Randolph (originally a middle school, to which a high school was added) had a high graduation rate, but its ACT scores were far below the state average, indicating that students are not well prepared for college. In its middle school, only 21 percent were proficient or advanced in math, placing Randolph in the fifth percentile in the state (meaning that 95 percent of schools performed better). Only 10 percent met the state science standards. In writing and reading, the school was in the first percentile.

Alter:
Her so-called evidence that the school is cooking its books is that Randolph’s ACT scores are far below the state average, as if such comparisons to wealthy districts somehow disqualify Randolph’s impressive year-over-year improvement in most areas. (And since when does Ravitch credit test scores?)

Ravitch also goes after the performance of Randolph’s middle school without mentioning that the results from sixth- graders -- one-third of the school -- merely reflect how poorly the students were prepared by the schools they previously attended, a significant though hardly atypical example of her misuse of statistics.


I'd like to point out first that the ONLY before-and-after statistic that we have for this school from either of these two is the graduation rate, which is going up. Ravitch cites low test scores, but how do we know that the low test scores aren't actually a vast improvement? Ravitch is a frackin' professor writing in the NYT. She should know better.

But Alter doesn't really claim improvement in test scores -- he makes vague claims about year over year improvements in "most areas", whatever that means. For all I know, it could be finger painting and plays well with others that have seen year over year improvements. He sure makes it sound like the test scores are improving, but no competent attorney would let a wishy-washy, noncommittal statement like that pass in deposition.

Maybe the test scores are going up. But the real question is this: are the test scores going up enough to justify using a term like "phenomenal results"? If they are, then Ravitch really does seem to be making the untenable argument that the "reform" she's critiquing isn't justified unless the worst schools turn into the best schools, which is ridiculous. On the other hand, if they aren't, then Alter's being somewhat misleading, because I do not read Ravitch as asserting that test scores haven't gone up, merely that the degree to which they have isn't really a cause for joy, celebration, or wholehearted endorsement of certain reforms.

But instead of all this debate, what say we instead have some dialectic? How about some real, actual facts about the schools in question? Here's what I could find:


Bruce Randolph Writing Scores Over Time (2004-2010):
There's some signs of a general trend over 5 years upward. Alter would probably proclaim "Proficiency levels are up almost 100%!" Ravitch would probably claim "Proficiency has inconstantly moved from below 10% to below 20%."

I have to give points to Alter on this one, though. There does appear to be improvement. In fact, if things are as Ravitch posits them -- if the problems of poverty are severe and systemic -- then the results are even more impressive. Nothing like arguing against yourself...

(Still, it looks like the school is suffering major attrition -- just watch the "number tested" fall from grade to grade. That makes its graduation rate a lot less impressive, and could affect how we're looking at test scores, too.)

Math Scores Over Time: No question about this -- there's some pretty remarkable improvement. Point Alter.

Reading? Again, definite improvement.

How about Miami Central?

Reading.
Math.

Imrovement in both cases, it seems. Especially if you're looking in the "meets standards" column over on the right.

Are these results "phenomenal"? I don't think so, but I encourage you to go judge for yourself. Don't trust any of the sides of the debate to give you actual facts, even when they're right. In the case of Bruce Randolph, it looks like hard, steady, grinding improvement, with all the sorts of inconstant ups and downs you'd expect of a work in progress. In the case of Miami Central, it looks a little more impressive.

My final conclusion? Ravitch is mostly wrong: the improvements seem nontrivial and wanting of an explanation that could very well have to do with the reforms in question, even if she (and I) wouldn't characterize those improvements as glowingly as some people might. The alleged spike in graduation rates, though, seem disproportionate to the actual learning being accomplished (at least the tested learning). So that's an area that needs to be looked at more closely.

Alter, on the other hand, is apparently a sophistical polemicist who can't even make a point well when the evidence is on his side. Maybe Ravitch is shilling for the unions. I don't care about her motivations: I care about the facts.

So, based on my incredibly unscientific internet-digging, it seems like maybe there's some considerable improvement. And there's definitely reform. That's a correlation, maybe. We should want of control groups if we're going to be serious about this. There could be some new television program out there that's turning kids into geniuses. But let's say Ravitch is wrong, tentatively, on the facts.

NOW we can start arguing about causation. Anyone got some popcorn?

No comments: