Thursday, June 07, 2007

Accurate test scores

Today's must read is from the Gannet newspaper syndicate via the Tennessean--and not just because I'm quoted. It's not that "test scores don't help" it's inaccurate test scores that don't help. And when the rules, tests, formats, content change regularly--the excuse that we're comparing apples to oranges allows enough room for a Mack truck to drive through. And sure enough...just as we're reviewing all of this the state is, again, revising the curriculum and tests. It's a moving target.

Tennessee has one of the largest gaps in the nation between how well students score on federal and state standardized tests, a new analysis of testing data has found.

'Cheap tricks' vary

States use a number of "cheap tricks" to create the illusion that students are doing better than they really are, said Dan Koretz, a Harvard University testing expert.

Those include designing tests easy enough for almost all students to pass or lowering passing scores to make sure most students make the grade.

Yet, somehow that isn't working well either.

I think the authors cut my quote awkwardly before printing it.
"Because taxpayer money is being taken, there must be accountability."
I actually started by saying until we have actual choice, where the free market can decide (which schools to utilize) and because we're using taxpayer money and educating other people's children some sort of accountability is required. This is a standard phrase of mine--as some of you will know. To me, it's always been more about the children than about the money. Both are important but children only get one childhood--somehow the legislature always manages to find more money for the things that are first on it's agenda.

Here's the link to the Gannet interactive map. Click on the map and then the mouse rollover feature will let you slide around from state to state to view the differences between their state scores and the national NAEP test.

To save you some time here are the differences for surrounding states in 8th grade Math Scores from 2005.
13 Kentucky
10 Missouri
11 Arkansas
39 Mississippi
48 Alabama
46 Georgia
52 North Carolina
48 Virginia
60 Tennessee
No other state in the nation has as much difference between state and national scores as Tennessee in math. The next three other states, NC, WV and NE all score in the low 50's.

And here 4th grade Reading.

37 Kentucky
2 Missouri
22 Arkansas
71 Mississippi
61 Alabama
61 Georgia
53 North Carolina
55 Virginia
66 Tennessee

The Tennessean provides a graphic of all the states in order for 8th grade math but not for reading. Wonder why they only show the one where we're not at the bottom.


Nashteach said...

Thanks for showing this. It is sad that the state has been trusted to devise it's own measurement and it has failed so badly in giving honest feedback to students, parents, and teachers.

The clearinghouse mentioned in the previous post is good. More authority ought to be given to the College Board, which makes and administers excellent tests, and the Southern Association for Colleges and Schools, which offers an excellent self-improvement program for schools via the reaccreditation process. It is the best tools schools have, though with the various levels of bureaucracy, SACS reports and recommendations unfortunately get put on the shelf too quickly and often.

What we really need to do is remember who our customers are: families here in Nashville. While accountability is necessary, there is no way bureaucrats in Washington, D.C. can truly measure a school and its progress, especially when it lets the state devise the measuring stick. We're raising children, not merely test takers. Too many kids need an assessment that's a challenge. Not all, but many students will rise to the occasion- they will jump as high as we ask them to. Unfortunately for many in Tennessee, the stick is way too close to the ground so everybody can make it over.

Eric Holcombe said...

Why else would TN escape the apples to apples comparison of the Stanford 9? Who then can question your "good" grades (and they are worried the homeschoolers aren't accountable).

I agree on the challenge of the assesment. If a child is constantly 99th percentile, the test isn't a test - or they are in the wrong grade.

What do you think the low bar for the lottery dollars is going to do to the SACS schools? We don't want to offend certain people groups (those that can't score the state average of 21 on the ACT) and are doing everything we can to keep them enrolled at the college level.

Bert D. said...

When NCES mapped the state cut scores for proficiency onto the NAEP scales one thing became clear. There is no clear, consistant relationship between the rigor of a state's cut score and the state's overall achievement. For evidence see:

Bert D. said...

the whole URL