Thursday, April 2, 2009

Student data and the $650 million stimulus [Link Update]

To qualify for the $650+ million in federal "fiscal stabilization dollars, what does Kentucky need to do with student data? USED believes that Kentucky meets nine of ten expectations:
  • Statewide student identifier
  • Student-level enrollment data
  • Student-level test data
  • Information on untested students
  • Statewide teacher identifier with a teacher-student match
  • Student-level course completion (transcript) data
  • Student-level SAT, ACT, and Advanced Placement exam data
  • Student-level graduation and dropout data
  • Ability to match student-level P-12 and higher education data
The one expectation we're thought to be missing is:
  • A state data audit system
That information is from a data quality survey reported here, along with added detail for each element. In the new application for stabilization funds, the Governor must either agree that it is accurate or submit corrections.

Are some of those elements are Infinite Campus possibilities we haven't field tested yet, and/or some we haven't yet paid to implement? Things we know are in reach, but not things we are already doing? If we've got all of that truly working, it's impressive progress compared to the last time I looked into how the data system was rolling out.

Whatever the current details, I hope we'll be very frank about anything that still needs work--and then seek stimulus incentive funds to get the whole system up to speed in the next two years.

Update: I'm sorry that the link to the survey won't work above. Try http://www.dataqualitycampaign.org/survey/states/KY/results. Also, at the link for the application, the information on baseline data is Appendix B in the Application PDF.

6 comments:

  1. Susan,

    I would argue, with good evidence, that Kentucky did not meet the “Student-level test data” item, either, because the Kentucky Core Content Tests were not valid and reliable for individual students and we had no test in the program that was for elementary schools – until Senate Bill 1 came along. It’s looking more and more like a good thing that this bill was passed.

    I can’t find where you got your information that US Ed says Kentucky meets 9 of the 10 requirements.

    The link to the “data quality survey” is broken.

    The “application for stabilization funds” link took me to a huge set of options, which at first glance didn’t seem to offer any state-level determinations of the type you discuss, but maybe there is something well-buried in that huge set of links. Can you provide a little more explicit detail?

    ReplyDelete
  2. I'm mysteriously unable to fix the survey link! Above, the update has the whole ugly url and more detail on where to look in the application document.

    Both the KCCT as administered from 2003 to 2006 and as revised and administered in 2007 and 2008 met the NCLB requirement that state assessments that must "be used for purposes for which such assessments are valid and reliable, and be consistent with relevant, nationally recognized professional and technical standards."[20 USC 6311(b)(3)(C)(iii)]

    The augmented-NRT that Kentucky used from 2003 to 2006 never received that approval

    ReplyDelete
  3. Susan,

    No Child Left Behind didn’t require data to be valid and reliable for individual students, just for student subgroups. And, in Kentucky, with our huge NCLB “N Number” loophole, that subgroup was pretty large.

    It’s questionable that the KCCT even met the NCLB requirement, but in any event, NCLB never drilled down to individual students. If it had, Kentucky would have failed, as even the NTAPAA admitted the KCCT wasn’t valid and reliable for high stakes decisions for individual students.

    I have that letter in PDF if anyone would like to see it. It will soon be available through the freedomkentucky.org Wiki site.

    ReplyDelete
  4. Richard,

    Can you give an example of a test that is considered valid and reliable for high stakes decisions about individual students:

    (a) by its own makers?
    (b) by others with the relevant credentials?

    My experience is that test developers are very careful to insist that the data from their tests will not not sufficient basis for high stakes decisions about individuals and to push hard for multiple other data sources to be considered.

    ReplyDelete
  5. Susan,

    The ACT is considered sufficiently accurate by our Council on Postsecondary Education that they use it to determine which students will need remedial courses in college. ACT also plays a factor in KEES scholarship awards. Those are high stakes uses, in my opinion.

    The ACT is certainly far from perfect, but it is a good enough predictor that our legislature and our postsecondary folks are both comfortable enough with the test to use it in those ways.

    Other tests are often used for identification and placement into special programs for students with learning disabilities. That, I would also submit, is a high stakes use.

    In very sharp contrast, the CATS was never designed to generate individual student scores. In fact, the CATS’ matrixed testing design precludes validity at the individual student level. That is another point that NTAPAA raised in their letter, which I mentioned in an earlier comment.

    ReplyDelete
  6. Richard,

    I meant people with psychometric credentials. I can't find any sign that the ACT test developers think that CPE is using the test in an appropriate way.

    ReplyDelete

Updates and data on Kentucky education!