Education Panel Finds Consensus on the ‘Incredibly Ridiculous’ Test Scores Argument
by Hugh McQuaid | Jan 6, 2012 6:30am
(10) Comments | Commenting has expired
Posted to: Education, Labor
At the governor’s education workshop Thursday, members of a four-person panel on fostering quality teaching agreed that teacher evaluations should not be based solely on student test scores.
Some argue that test scores are the best indicator of a teacher’s performance but Randi Weingarten, president of the American Federation of Teachers, said the argument over testing “has become incredibly ridiculous.”
“Every teacher I know creates tests to figure out where their kids are. But this is about student improvement. Most tests were never created as a teacher evaluation document,” she told the workshop at Central Connecticut State University.
Weingarten argued that test scores should be used to evaluate teachers, but only as one of a number of other factors. She wasn’t alone. On Tuesday, Connecticut’s largest teachers union released a report which included the same recommendation.
Every member of the panel agreed with the unions’ position. Though he disagreed with Weingarten on how to reform teacher tenure, Joseph Cirasuolo, executive director of the Connecticut Association of Public School Superintendents, said there are other ways to gauge a teacher’s success that need to be included.
“The most significant factor should be growth in student achievement multiply measured. Not a test score. Test scores will be a part of it but there are a lot of other ways to measure growth,” he said.
Professional practice should also be considered as well peer reviews and input from the parents of students, Cirasuolo said.
New Haven Superintendent Reginald Mayo said his city already uses multiple methods of assessing teachers.
In New Haven, student performance makes up 50 percent of a teacher’s evaluation, he said. The other 50 percent is based on other things like a teacher’s instructional strategies and professional values, he said.
Gov. Dannel P. Malloy said that, as a person who never in his life did well on a standardized test, he understood they shouldn’t be the sole measure of success. But he cautioned against disregarding them.
“To deny that there are standards that we must hold ourselves to, and that we must simply take a look in the mirror or on the test form and understand that in many areas, in many communities, poor and urban in nature in our state, we are failing students and we can see it year after year by using, in part, those tests to measure success, would also be equal folly,” he said.
Patrick Riccards, CEO of the private, non-profit, pro-charter school group ConnCAN, who was not on the panel, said test scores should be the most important factor when teachers are evaluated.
Riccards said people who seek to discredit the usefulness of testing want to measure what teacher puts in to their jobs. Test scores are critical to measuring the outcomes of a teacher’s work, he said.
“I think there is still no dispute that test scores have to be a primary driver to that formula,” he said.
On Tuesday Riccards said that states which benefit the most in federal programs like the Race to the Top-Early Learning Challenge funds Connecticut missed out on last month are the states that significantly rely on standardized testing results for evaluations.
“At the end of the day the most effective way to know whether our schools are succeeding or not is to know if our kids are succeeding and that needs to be the primary measure,” he said.
Weingarten said test scores are mostly useful to teachers as a source of data to help teachers hone their craft. The scores allow teachers to gauge where their students are and how well they are learning, she said.
Focusing too much on those numbers turns the evaluation process into “a numbers game and a compliance issue,” she said.
Besides, there is a lot that goes into teaching that can not be measured by a test, she said.
“It reminds me of Albert Einstein, who was an AFT member, who said ‘not everything that counts can be counted,’” she said.
Tags: education, teacher evaluation, american federation of teachers, malloy
(10) Comments
posted by: Noteworthy | January 6, 2012 10:28am
Riccards comments need to be taken with a grain of salt. ConnCan and Achievement First, a related entity stands to profit from their aggressive effort to take over more schools. Aside from that, it is not an all or nothing benchmark. Absolutists are rarely correct.
posted by: THREEFIFTHS | January 6, 2012 11:28am
What role does the parent play.All I here is blame the teachers.
posted by: brutus2011 | January 6, 2012 1:16pm
If, as the Governor says above, we are failing our students year after year, why are those whose job it is to run things are not held accountable?
posted by: schoolmom | January 6, 2012 8:56pm
“Ridiculous argument?”
Perhaps we should look at the research by Prof. Jesse Rothstein at Berkeley, Professors Briggs and Domingue at Colorado, Prof. Sean Corcoran at NYU, and other peer-reviewed research ALL finding error rates of more than 40-60% when using test scores to evaluate teachers.
Or we can read the recent Congressional Briefing submitted by the American Education Research Association and National Academy of Education, which after citing voluminous research, concludes that value added models are too unstable to use as a method of teacher evaluation.
Principals all over New York State have signed a letter protesting the use of tests in teacher evaluations, based on their experience and the research. And the New York Times has reported several times on the unworkable nature of Tennessee’s teacher evaluation system, which also uses standardized test scores.
So while true believers may view this argument as “ridiculous,” it is clear from the mountain of evidence that what is ridiculous is using a BAD measure to rate teachers. And saying that test scores will be part of some unmentioned “multiple measures” is a farce. We all know that those other “measures,” whatever they will be, will fall by the wayside and all people will focus on will be test scores. This of course means that our children will be subjected to an endless series of mind-numbing tests in all subjects.
posted by: Tom Burns | January 6, 2012 11:46pm
Wow—first off Joe Cirasuolo mustn’t have ever been a superintendent himself because he doesn’t understand the tenure law in CT—this is embarrassing—to him and our state for listening to him—-Tenured teachers can be fired at will—with no defense—at present the arbitration proceeding in ANY tenured teacher termination case gives the BOE carte blanche in firing ANY tenured or non-tenured teacher any time they want—arbitrators only suggest, even when the teacher wins—the teacher loses anyway—Is this guy really the head of our school Superintendents—-thats scary—-Peer review and Parent review of teachers is a flawed concept at best—these supposed educators need to get back in the classroom to see what is really needed—more no-it-alls, far away from the classroom making bureaucratic decisions to make themselves look good while hurting children are just what we need—right???
This new Conn-Can guy doesn’t have a clue—in his eyes its more important to get Federal $$ than to serve the kids by getting away from the testing mantra of the corporations—“No pig ever got fatter from sitting on a scale” but I guess our kids will get smarter—Infantile thoughts at best from these two—
posted by: Tim_Holahan | January 7, 2012 1:43am
Using raw test scores for teacher evaluation is an obviously bad idea, and few people seriously propose it. The idea that the value-added methods provide a significantly better insight into teacher quality is appealing, but inaccurate.
I’d strongly urge Patrick Riccards, and anyone else who believes that value-added analysis should be “the most important factor” in teacher evaluation to watch this long but important video all the way to the end:
http://www.youtube.com/watch?v=43BoGn5oSm4
The first half is heavy sledding for a non-statistician, but it’s important to understand this stuff if you’re going to offer policy opinions as Mr. Riccards does. Two applied statisticians, both academics, who have studied the details of the value-added method, present a polite but withering critique of the method. They do this in the presence of the original researcher who proposed it, Dr. Eric Hanushek.
In the face of this dispassionate and evidence-based critique, Dr. Hanushek admits that “the Holy Grail of all researchers has been, through observational means, through data steps, to collect various survey information… to find correlates of effective teaching. And it’s just failed. I know that it won’t be found, the one thing… won’t be found in my lifetime, but I’m old.” (1:00:10)
This should have been a humbling moment for a researcher who built his career on the idea that finding those correlates was possible, and that they implied causation. Unfortunately, Dr. Hanushek has continued, since this event, to make wildly tenuous assertions about the impact of “bad teachers” upon the economic future of their students (http://bit.ly/gNtk0p).
Identifying “bad” teachers is not something that can be done by algorithm. It’s irresponsible to suggest otherwise. The effects of implementing an evaluation system that gives primacy to value-added measures, as Mr. Riccards suggests, are likely to be harmful, in that they will lead to the mis-labeling of teachers and depress the morale of the profession as a whole.
No one argues that we shouldn’t push for rigor and excellence in our schools. Advocating the use of statistically invalid techniques, however, is bad policy. I respect many of ConnCAN’s efforts, but I strongly disagree with them on this point. I continue to hope they will affirm their commitment to evidence-based policies, and abandon this one.
posted by: THREEFIFTHS | January 7, 2012 12:55pm
Check this out.RSA Animate - Changing Education Paradigms
http://www.youtube.com/watch?v=zDZFcDGpL4U&feature=youtu.be