Re: proposals for national test from Fordham

[Kenneth Bernstein <kber@earthlink.net>]

Aaln

I can only respond to you on ARN as I am not a member of the other two groups.   As you may know, I am friends with your outgoing governor Tom Vilsack because of our common interest in education -- he was a member of our education panel at Yearlykos in Las Vegas in June.

In the first direct conversation I ever had with Tom he told me that Iowa had no exit exams and would not have them so long as he was governor.   It helps that his wife was both a teacher and a school librarian, and has had some influence on him on these matters.

You are lucky in not having YET succumbed to the beast of exit tests and all that go with them.  I ahve recnetly testified before a zblue ribbon committee in Prince George's County MD (thanks to my name being suggested by Sue Allison)mon our high schoo exit exams and a lot of problems with them, and apparently I caused the guy from the state testing office to get unhinged ... let's leave it at that.  He made a lot of really bad statements, including dismissing a reference by the other witness to a study by Berliner and Amriend on the gournds that it as "self-published" because it appeared in EPAA.
I was asked to give some followup written response, and that was one of a number of issues on which I further took him apart.  

How I wish we didn't ahve the monstrosity of these exit exams.  I pointed out numerous errors in question that had no correct answer or more than one correct answer, how the 2005 test asked the same question 3 different ways (out of a total of 58 items), how there were too many items on economics and the like.

And the guy tried to defend a 70% interrater reliability rating in grading constructed responses by saying below that level they fire them and that there  real reliability rating is in the 90 % range, which does not answer my charge that it is acceptable to have 71% reliability that is adacent scores, say a 2 and 3, on a sacle of 0-4.   He said that was not a problem because in that case the studetn receives the higher score.  I noted in my written response that my complaint was accuracy of measurement, and merely giving the student the higher score did not address my complaint.  I also noted that there is no correction for guessing (unlike on the AP or SAT) on the selected response items.

Believe it or not, when I pointed out that every professional organization opposes desires the use of multiple measures for high stakes purposes his response was that students can take the test over, as many as 3 times a year.  That is NOT multiple measures, that is multiple applications of one one measure, and does not address my complaint.  And he is the pscyhometrician, with the Ph. D. in the relevant discipline, while I am merely a classroom teacher (yeah, I am ABD in Ed admin and policy, but I have had exactly one course in testing and statistics only through multiple regression).

Oh what a mess we are creating with what we are doing with education.

By the way, I offer several items for your additional reading

1) a diary I did this weekend on dailykos that got a lot of discussion:  http://www.dailykos.com/story/2006/9/3/8489/00713

2) a post I did to the group blog on education The Wall: 
   http://thewallofeducation.blogspot.com/2006/09/i-think-we-have-lost-our-w_115728667793468467.html

Peace all.


Ken Bernstein  (aka teacherken)

Kenneth J. Bernstein