OPINION: Stop holding us back



I recently received an email from a group called, “Don’t Hold Us Back.” They’re a coalition who have taken out full-page ads in major newspapers proposing that United Teachers Los Angeles and The Los Angeles Unified School District complete negotiations on a contract within 30 days, and they’re encouraging readers to call and email leaders of UTLA and LAUSD to encourage adoption of their agenda. Among their demands is a proposal to incorporate student test scores in teacher evaluations, despite the 25% error rate, as has widely been reported.

I called and emailed the following organizations, as listed on the website, to find out why they would propose such a flawed evaluation system: Alliance for a Better Community, Asian Pacific American Legal Center (APALC), Community Coalition, Families in Schools, Families That Can, InnerCity Struggle, Communities for Teaching Excellence, Los Angeles Urban League, Union de Vecinos, United Way of Greater Los Angeles, Watts/Century Latino Organization, and the Southern Christian Leadership Conference.

I asked each the same question I recently asked of Obama, Duncan, etc. “According to the US Department of Education report, Error Rates in Measuring Teacher and School Performance Based on Student Test Score Gains, an effective teacher could be rated as ineffective 25% of the time, and an ineffective teacher could be rated as effective 25% of the time, so, my question is, what is an acceptable rate of error when your job is on the line?”

One of the organizations indicated that they would not be issuing an official response. Another stated they would not be commenting. Another indicated that they would respond by my deadline — they didn’t follow-up.

Executive Director Angelica M. Solis of Alliance for a Better Community issued a 380-word response reiterating their support for the proposal, and referenced a “2011 study published in the journal Labour Economics.” When I inquired as to the specific study referenced in the email, I did not receive a response.

Taulene Kagan, Marketing Communications Director for the United Way of Greater Los Angeles, issued a 122-word response indicating, in part, that, “the research report you refer to clearly states that potential statistical misclassification would be mitigated if multiple measures over time are used.” When I responded, “please kindly indicate where the research report specifically states that potential statistical misclassification would be mitigated by using multiple measures over time,” I did not receive a reply.

By far the most curious response was from Elizabeth Blaney & Leonardo Vilchis of Union de Vecinos. Both separately voiced disapproval for efforts to tie teacher pay to student performance on standardized tests. Leonardo wrote: “we do not support connecting teacher’s pay to student standardized test performance, for the reasons you describe and others that include the problems with standardized testing. However, we believe that there has to be more evaluation of teachers that includes community and student input. We also believe that principals and supervisors need to be more thorough in their evaluation process.”

I inquired further, in part, writing: “You do realize that you’re listed as a supporter of ‘Don’t Hold Us Back,’ don’t you? Secondly, one of their proposals is to incorporate ‘academic growth over time’. Are you now withdrawing your support, or, can you explain, please, why you’re listed as a supporter, yet, based on your own statement, you don’t support the proposal itself?” I did not receive a response.

Simply incorporating a meaningless, random number that in no way reflects the complexity of teaching will not address bridging the achievement gap. How would it affect your on the job productivity if we flipped a quarter four times, and every one out of those four times you were given a below satisfactory evaluation no matter your actual real life job performance?

Given Union de Vecinos’ position on using standardized tests as part of teacher evaluation, how many of the other groups that signed on to this platform also oppose using test score information in teacher evaluations? What’s an acceptable level of error when your job is on the line? Care to flip a coin to determine your response to that question?

OpEd: Linking teacher evaluation to student test scores: wrong 25% of the time



Originally published on Huffington Post

Nowhere is the disconnect between billionaires and public school teachers more stark than when it comes to merit pay proposals. So why are self-anointed “reformers” pushing this agenda, and why do public school teachers so overwhelmingly oppose these efforts?

The Los Angeles Times launched a series on “Value-Added” assessments last year, and they continue to stand by it, despite the reportedly high error rate. This year, LAUSD Superintendent John Deasy (who previously worked for The Bill and Melinda Gates Foundation) has launched a new proposal, now called Academic Growth Over Time, and unilaterally implemented it, even though teacher evaluation is a negotiated issue. The District has even offered money to school sites that participate in this “voluntary” process. United Teachers Los Angeles (UTLA) has filed an unfair labor practice charge with the Public Employee Relations Board (PERB) over its implementation.

During an Education Summit panel discussion August 31 hosted by Patt Morrison that included Deasy, LAUSD Board President Monica Garcia and UTLA President Warren Fletcher, at one point Fletcher said, “And the Superintendent has proposed a system of evaluation called AGT, Academic Growth Over Time, which in most of its aspects is identical to Value-Added models used by the [LA} Times. But the US Department of Education itself says that it’s inaccurate 25% of the time.” Click the third audio file to listen to the discussion about Teacher Evaluation at the Education Summit.

To be absolutely clear, according to the US Department of Education report, “Error Rates in Measuring Teacher and School Performance Based on Student Test Score Gains,” 25% of the time programs like AGT will wrongly label an effective teacher as ineffective, and 25% of the time programs like AGT will also label an ineffective teacher as effective.

Did Garcia or Deasy correct Fletcher’s assertion that AGT is mostly identical to Value-Added models? No. Did Garcia or Deasy dispute, in any way, Fletcher’s assertion that the US Department of Education stated that models like AGT are inaccurate 25% of the time? No.

Instead, Deasy focused on alleging that teacher input was included in the development of AGT. Fletcher responded, “I would prefer that those people who were selected to develop an evaluation system not exclusively be made up of people who were selected by the Superintendent, the school board, and senior management.” Did Garcia or Deasy correct Fletcher on this assertion? No.

In fact, there were four UTLA members and professional staff who did participate in this process that were not hand-picked by the Superintendent, school board, and senior management—and together they wrote and signed a four-page letter sent to LAUSD administrators stating, in part, that while they were provided an opportunity to voice their concerns, “we believe that these concerns were not heard and therefore we must put our concerns and comments on the record as this process seems to be unfolding rapidly.”

Among the numerous unanswered questions and concerns they raised about AGT: “What is an acceptable level of error if your job is on the line?”

While they were offered a meeting in response to their letter, there was no assurance that the consultants designing the program would be available to address their concerns, and, to date, there has been no written response. The Obama administration is now letting individual states opt out of No Child Left Behind, which, in part, incorporates the use of standardized test scores in teacher evaluation. Unfortunately, however, states can only opt out if they agree to certain provisions of Race To The Top, which also require the use of standardized test scores in teacher evaluation, despite the 25% error. One of the main proponents of this effort, Bill Gates, has poured millions of dollars into this proposal, and for unknown, unstated reasons, he’s determined to attempt to apply unproven mathematical models to teacher evaluation, even though, in this case, 2+2 = 5.

I called and emailed the following individuals, organizations and their press representatives, and gave all more than five days to respond to this article: President Obama, Secretary of Education Arne Duncan, Bill Gates, The Bill and Melinda Gates Foundation, and Batelle For Kids, the Ohio-based organization that operates AGT for LAUSD.

I asked each to address the same exact question: “According to the US Department of Education report, ‘Error Rates in Measuring Teacher and School Performance Based on Student Test Score Gains,’ an effective teacher could be rated as ineffective 25% of the time, and an ineffective teacher could be rated as effective 25% of the time, so, my question is, what is an acceptable rate of error when your job is on the line?”

Neither President Obama, Secretary Arne Duncan, Bill Gates, or Batelle for Kids responded directly.

I did receive an email from the LAUSD Media and Communications Department, though I had not directly contacted them. The email did not reveal which of the above parties had contacted them, and did not attempt to answer the central question I had posed regarding the 25% error rate. I also received a 237-word email response from The Gates Foundation Media Team that stated, in part: “The foundation does not support a system of teacher evaluation that is solely based on student test scores.” The response referred to a survey of teachers that they indicated was, “commissioned by Scholastic, Primary Sources.” They did not reveal what I discovered in the small print on page two of the survey after I downloaded it: “This report is a collaboration of Scholastic and the Bill & Melinda Gates Foundation.” The response also did not address the one and only question I had posed: “what is an acceptable rate of error when your job is on the line?”

The question remains.