Re: Fw: [LiteracyForAll] change in strategy: just "fix" value-added
- How about this analysis of teacher evaluation?
--- In email@example.com, Eduardo Martinez <ezedmartin@...> wrote:
> More explanation for the uselessness of value-added measures as an evaluation
> As the parent of armies, war encourages debts and taxes, the known instruments
> for bringing the many under the domination of the few. - James Madison
> ----- Forwarded Message ----
> From: Jill Kerper Mora <jmora@...>
> To: LiteracyForAll@yahoogroups.com
> Sent: Thu, November 4, 2010 7:48:44 AM
> Subject: RE: [LiteracyForAll] change in strategy: just "fix" value-added
> Value-added measures (VAM) can't be "fixed" because its underlying theoretical
> assumptions are false. VAM is based on the assumption that such a thing as a
> "teacher effect" can be isolated, statistically at least. This is wrong for two
> reasons: First, students' learning (or the lack thereof) is attributable to the
> totality of their experiences, including all of the teachers they have for any
> portion of their instruction during a school day & year. This number of teachers
> is rarely N=1. This is especially true in grades 6-12 where teaching is
> departmentalized & by courses. The "teacher effect" theory also fails because
> the assumption is that all teachers are fully equipped with everything they need
> (sound & effective curriculum, instructional materials, time & other resources)
> to do a perfect job of teaching so that students' learning is purely
> "value-added" based on teachers' effectiveness & efforts.
> We can't let the likes of Eric Hanushek get away with leading policymakers & the
> public to believe that by just "fixing" the statistical model that we can "fix"
> value-added model research.
> See the pages on my website about merit pay and value-added model evaluations.
> -----Original Message-----
> From: LiteracyForAll@yahoogroups.com [mailto:LiteracyForAll@yahoogroups.com] On
> Behalf Of Stephen Krashen
> Sent: Wednesday, November 03, 2010 4:30 PM
> To: LiteracyForAll@yahoogroups.com
> Subject: [LiteracyForAll] change in strategy: just "fix" value-added
> There has been a change in strategy: supporters of value-added and making
> teacher evaluations public now admit that there are problems but think that they
> can be easily fixed.
> The relevant section of the Star-Ledgrer's editoral:
> "In New York, the teachers union filed a lawsuit to block such a release saying
> errors are rampant: Some teachers were scored based on students or classes they
> never taught, and thereâs a high margin of error. Unpredictable swings rank a
> teacher in the top tier one year and near the bottom in the next. If true, that
> must be fixed."
> I suspect that this confession comes from Hanushek's article in the NY Daily
> News, available at http://susanohanian.org/show_nclb_outrages.php?id=4065.
> My response:
> Fixing value-added evaluations: Not our first priority Sent to the NJ
> Star-Ledger, November 3, 2010 The Star-Ledger feels that value-added scores, the
> gains a teacher's students make in a year on standardized tests, should be
> released to the public (Nov. 2). The Star-Ledger recognizes that there are
> problems with using value-added scores, and states that they must be "fixed."
> It's not that simple.
> Studies show that value-added ratings are unstable. Value-added ratings based on
> one year are weak predictors of value-added ratings the next year. A teacher who
> succeeds in boosting scores with one group will not necessarily succeed with
> others. Studies also show that different reading tests result in different
> value-added scores for the same teacher.
> Value-added ratings may not represent real learning. There are ways of pumping
> up test scores without student learning, including teaching test-taking
> strategies and making sure weak students don't take the test.
> It would take years of hard work and major financial support for research to fix
> these problems. Our schools are facing tremendous financial problems: In high
> poverty areas, science classes lack equipment, libraries lack books, and even
> bathrooms lack toilet paper. Funding complex and subtle studies to attempt to
> create what might or might not be a better teacher evaluation measure is not our
> first priority.
> Stephen Krashen
> Yahoo! Groups Links
> [Non-text portions of this message have been removed]