Lake Wobegon Effect
Date sent: Mon, 9 Mar 1998 22:44:48 -0500
From: "Richard G. Innes" <70224.434@compuserve.com>
Subject: Re: Terra Nova from Charlene
To: ed-consumers main listing
Charlene (and the LOOP/ECC),
<>
I am not a testing expert, either, but I have looked at some of the
numbers. More importantly, what you are saying is quite interesting and
"unusual."
Terra Nova is marketed as a nationally-normed test. The norming was done
about two years ago, and does not have to be repeated to generate national
comparison scores. If what you say above is correct, then it could be your
school system really isn't going to get a national comparison score. Your
results will only compare your kids to other kids in your state alone. If
so, then if your state is overall weak compared to the rest of the country,
you will have no way to know that.
That could be a nice dodge for your state department of education, but it
would be very misleading to parents. However, I really need to know more
about what is going on in Tennessee to really understand what is being
proposed.
((Dr. Stone, have you heard this?))
<>
This refers to a mythical place from Garrison Kellor's PBS comedy program
that often talked about Lake Woebegon where every student was above
average. In the late 1980's a non-educator discovered that every state in
the country was reporting that its testing scores were above the national
average. Obviously, that wasn't possible. It turned out that when
nationally-normed tests were given more than once, teachers got better at
remembering the questions (which don't change) and making sure their
students were well prepared for just those specific questions. Scores thus
became inflated, and every school and system tested out "above average."
This is still a problem today with just about any test that does not change
questions, but there are some work-arounds. First, better tests are
available in multiple forms. While each form has different questions, they
are all normed at the same time, so scores, at least in theory, are
comparable with high accuracy from one form to the other. The different
forms, however, make it hard for teachers to look good by teaching to just
a narrow set of questions. A school system that really wants to take a
hard look at itself will demand different test forms each year (and a
system that wants to hide something will make darn sure it does not change
testing forms each year).
Kentucky tried another approach. They changed a large number of questions
each year, and then tried to develop a very exotic formula to "equate" the
different tests so scoring would be even. To date, that has not worked (in
fact, it has been somewhat of a disaster, in my opinion, not necessarily
shared by others).
I hope that helped more than confused.
Richard Innes
EDUCATION CONSUMERS CLEARINGHOUSE