Re: counting up the test failures here in VA

[Art Burke <aburke@VANSD.ORG>]

Mickey ...

Think of a score within the SEM as being "too close to call" on whether the student's true achievement is  passing or not.  The SEM give you a way of quantifiying the unreliability of the test around a decision point.  Note that you can be within the SEM on the high side or the low side of the passing score;  we usually only concern ourselves with scores on the low side.  With a little luck (and some study thrown in) students within the SEM on the low side have a good chance of passing on the re-test.

To start sorting out the number of kids and number of tests issue, first find out how many kids are taking the courses the tests are given in.

Art

>>> PAVURSOL@AOL.COM 03/01 8:40 PM >>>
Hey,
I got some of my division's test data yesterday and using my calculator,
discovered the following numbers:

Our school division has about 10,500 kids.  If you just take the high school
End of Course SOL tests, the ones that are going to count against kids in
2004 for graduation purposes, and you add up the number of tests taken (some
kids would have taken more than one), you get 6,696 total tests during the
2001 administration of the tests.

Of those 6,696 tests given in our high schools (and a few in our middle
schools for the advanced kids), there were 2,012 failures.

Of those 2,012 failures, about 937 of them were within the Standard Error of
Measurement.

At one high school, 301 kids took the  World History from 1000 to Present SOL
test but only 156 took Algebra.  95 took Biology; 95 took Algebra 2.  If
there is an approximate equal number of kids in each grade (and there is),
why would those numbers look like this?  Any thoughts on what is happening?
FYI, there are about 260 kids per grade level.

Mickey

--------------------------------------------------------------------------
To unsubscribe from the ARN-L list, send command SIGNOFF ARN-L
to LISTSERV@LISTS.CUA.EDU.