On Thu, Aug 25, 2005 at 12:11:04PM -0400, Chris Browne wrote:
> Consider the political issues surrounding standardized testing in
> schools; both in the US and Canada, there is a lot of controversy
> surrounding "standardized tests." Parents and politicians want there
> to be some form of "accountability", which has led to students
> spending more and more of their year writing such tests, with a
> concomitant reduction in the amount of time teachers can spend
> actually teaching them. They are increasingly 'well tested,' but less
> educated, because the time was spent on tests :-(.
This is only indirectly related to the difficulty of writing good
tests. Good tests _are_ difficult to write, and are extremely
difficult to validate: ask the people who prepare the LSAT and GRE if
you doubt it. But that's not what the problem is with standardised
testing of schoolchildren. The problem there is that everyone _also_
expects that the standardized testing won't result in more children
flunking out. What people want is a standardized test that produces
happy results, rather than tests that are good predictors of academic
ability.
Which is why many certification tests are crap: they're there to seed
the market with enthusiastic product adopters; and giving someone who
just paid $1500 for the privilege of taking the test a failing grade
is not likely to gain new adherents.
What I'd be really interested in seeing, therefore, is an institute
that handled such test scores the way ETS handles GRE scores: you
agree to release such scores to interested parties when you take the
test.
A
--
Andrew Sullivan | ajs@crankycanuck.ca
The plural of anecdote is not data.
--Roger Brinner
