NYC introduces "report cards" for schools (incoporating "growth model")

["ElsaHaas" <ElsaHaas@si.rr.com>]

The New York Times today had a major article on a new, citywide system for
rating NYC public schools. The full article ("Fifty New York Schools Fail
Under Rating System") is here: 

http://www.nytimes.com/2007/11/06/education/06reportcards.html  

The crux of the article, describing how the rating is done, is this:

---- 

[...] The grades released yesterday contained many surprises, with some
schools with top-notch reputations receiving B's, C's, D's - and even F's,
to the astonishment of some parents. 

That is because unlike traditional methods of judging schools, this one
involves a complex calculation that assigns the most weight to how
individual students improve in a year's time on standardized state tests. It
also compares schools with similar populations, as judged by demographics
and incoming students' test scores, and assigns final grades based on a
curve. More than 60 percent of the schools earned A's or B's. [...]

The largest portion of a school's grade, 55 percent, is based on the
improvement of individual students on state standardized tests from one year
to the next, a so-called growth model analysis. Thirty percent of the grade
is based on overall student achievement on state tests. Fifteen percent is
based on the school's environment, measured by attendance figures and
parent, teacher and student surveys. [...]

-------

I used the "Search Report Cards" feature over to the left of the article to
find that the school down the hill from us, where our son would be if he
weren't homeschooled (actually, unschooled), got a "B." 

I had figured it would get a good grade - that school has a fine reputation.
But when I visited it, just because I wanted to keep an open mind about it
before we decided to homeschool for "first grade" two years ago, I didn't
like it. 

I would be hard to please, I know. 

Without getting into too many details right now, I was pleased for about
three seconds to see that the fifth-grade class was reading The Lion, the
Witch and the Wardrobe, the first book in the Chronicles of Narnia (not
because I'm a Christian, which I'm not, but because that was one of my
favorite books as a kid). 

The reason I was pleased for only about three seconds is that I learned that
they were reading this book by noticing a hallway display of reports on The
Lion, the Witch and the Wardrobe. Twenty-five or more of them! (However many
kids there were in that class.) 

I thought, "Do I want my kid in a classroom where all the kids have to read
and report on the same book at the same time?" Answer: no.

So the grades mean little to me, but I suppose that if they're going to use
standardized test scores at all, it's better to use a "growth model" for at
least part of the reckoning (a point I've made on this list before because
as homeschoolers in NYS, we have the option of showing, rather than a score
above the thirty-third percentile, a score that demonstrates at least one
year's "growth" since the previous year).

On the other hand, it does seem that schools with very high percentages of
kids doing very well on the tests may find that it's hard to keep improving
every year. 

We homeschoolers can CHOOSE between showing that our kids have met the
thirty-third percentile cutoff, or that they've experienced "growth." What
does a school do if all the kids are near the top from the beginning? 

It reminds me of a homeschool mom who panicked recently because her kid had
previously scored a couple of grade levels ahead, and she thought that they
would have to demonstrate "growth" on the next test, which would mean
leaping over a pretty high bar. I told her it's either the cutoff or the
growth, not both!

I hasten to add that a lot of states don't even require standardized testing
for homeschoolers. Nobody has yet shown this to be a problem.  

Elsa Haas