Sunday, May 29, 2016

HAL Wants To Pick Your Teacher

Are you responsible for hiring teachers in your district? Do you hate all the mess and bother of actually interviewing other carbon based life forms?

Well, meet TeacherMatch.

If  you are a teacher, TeacherMatch will help you find work, and if you are a hiring department, TeacherMatch can perform flat out magic through the power of Predictive Analytics. From the TeacherMatch blog...

Previously, school districts sorted through a stack of resumes and asked scripted questions during the teacher recruitment process. Today, human resource (HR) departments use predictive analytics to improve the quality of hire, because it allows districts to predict which teachers will positively impact student achievement before they enter the classroom.

Sigh. First, any districts that use "scripted" questions during recruitment and hiring deserve to get the worst hires available. If you are responsible for hiring and you're using scripted questions to do it, you are in the wrong job. Go do something else.

Second, any district that is looking for teachers who "will positively impact student achievement" aka "raise test scores" has lost track of their actual function.

Third, anybody who believes that a consulting service can predict how good a teacher will be before that teacher enters a classroom should go buy a bridge that leads directly to a swamp in Florida.

But wait! There's more! TeacherMatch points out that the power of predictive analytics can be used on students so that teachers and school leaders can "make informed decisions."

One thing TeacherMatch is not is shy. Their "about us" page notes that they were founded by four super-duper people who "worked in the K-12 education system" and who developed a system so awesome that "the U.S. Department of Education used it to shape their multi-billion-dollar school improvement program." It's a bold claim, considering the USED School Improvement Grants program turned out to be a seven billion dollar bust.

So who are the great educational minds behind TeacherMatch?

Well, there's Ron Huberman, co-founder and executive chair. Huberman was Chicago Mayor Richard M. Daley's Chief of Staff. Later, when Arne Duncan went to DC, Daley moved Huberman into the top spot of Chicago schools. His most recent job before that one? Running Chicago's transit authority. His previous experience of any sort in education? None. Never.

Huberman was hired because of his management experience, and he talked like a manager. He brought in a group of upper managers, all completely devoid of education experience, and he tried to kick his teachers into high gear with a power point presentation described as "Orwellian."

"Be sure to get granular about your metrics before you deep dive into those outcomes and be sure you don't avoid those brutal facts' qualitiatives... Got that?"

Huberman's big accomplishment at the Chicago Transit Authority was to "fix" the budget and the pension problem, and it was that bean-counting skill that was called for at CPS. His clever trick for dealing with the huge debt the system owed teacher pension funding-- get permission to just not pay it. Then he went on to work in the field that he truly loved-- private equity.

Co-founder Don Fraynd taught at a prep school, became a principal, and eventually ran the CPS turnaround office. Co-founder Sanjeev Arora is an investor-entrepreneur who once worked for McKinsey. The rest of the management team includes data science specialists and PR folks-- but not one person who's ever taught in a public school. TeacherMatch has also research partnered with the University of Chicago and NWEA.

Their product is a battery of on-line tests for candidates and teachers to take, everything from QUEST (for recruiting) to EPI, the predictive analytics tool that will identify the four core factors in teacher success. Those four factors are qualifications, cognitive ability, attitudinal factors, and teaching skills, and TeacherMatch can totally tell if you've Got the Right Stuff by using their super-duper proprietary questionnaire.

How shady is this? This shady.

You know when you're on shaky ground? When the National Council on Teacher Quality has questions about your methodology.

Yes, NCTQ, the group that evaluates teacher preparation programs that don't exist, and that evaluates the rigor of a program based on commencement programs-- those guys. Kate Walsh, head of the least serious "research" group in education, said that it's hard to know if TeacherMatch is bunk or not.

“We don’t know whether their predictive analytics are accurate,” she said. “It might be snake oil or it might be great.”

I'm prepared to make the great-or-snake-oil call, but let's see what Huberman has to say about the origins of this magical tool:

Huberman said the company enlisted the help of researchers who analyzed reams of data on student performance, looking for teachers whose students consistently made large gains on standardized tests and teachers whose students consistently did not make those gains.

Then the company surveyed both kinds of teachers, looking for patterns in the way they answered questions about how they might respond to classroom misbehavior, for example, or how they might teach a certain academic standard. That work became the backbone of the inventory now used in dozens of districts.

So what TeacherMatch has done is reverse-engineer a Value Added Measure system-- only sloppier, lazier and with even less basis in solid data and research. Huberman and his buddies actually found a way to make VAM worse! Though when it came time to pitch this, they whipped up a video that is both full of jargoneque vaguosity, and yet somehow promises more than Huberman's quick explanation. Get out your business bullshit bingo card, and play along:

Note that "teachers matter most" as we repeat the old misrepresented data about teachers being the single largest in school factor. And look--here are some logos of "renowned universities" (and the Gates Foundation). I like the part where they call themselves "a team of dedicated educators," which may be one of the more loose definitions of "educators" I've encountered in a while. And look-- they flipped through decades of research to "extract actionable conclusions" which sounds so much cooler than "to write some multiple choice questions" (any experts on test design in your group, there?) But at least they used "advanced scientific techniques that provide multilevel analyses of nested groups, test causal relationships of abstract variables and measure comparative outcomes" and how in the name of God does this NOT set off everybody's bullshit detectors? I can't even imagine writing this with a straight face.

The EPI is reportedly just a 100 item multiple choice question test. And Huberman does say that you should have other parts of your hiring process. But that's not the question-- the question is, why would you use this at all ever? Although you know who might use this? Other people with no first-hand knowledge of teaching, like charter school operators and Broad-trained superintendents. So maybe there is a market.

The Bad News

There must be a market, because another company just bought TeacherMatch. Here's the lead from the May 24 press release:

PeopleAdmin, the leader in talent management software for education, announced today that it has acquired TeacherMatch, joining forces to offer the industry’s most comprehensive talent management platform and analytical solution for identifying, hiring, and developing educators most effective at driving student achievement.

If you want to learn more about PeopleAdmin, you could attend their big June convention in Austin, a "can't miss industry event for Higher Ed and Government customers, thought leaders, and PeopleAdmin employees."

In the meantime, we are left to conclude that TeacherMatch actually is actually making money, which is sad. It's the kind of thing that really shakes my faith in capitalism and the free market, because this is clearly a company that deserves to die.


  1. Amazing what people will do in spite of beloved data. Cash is king. Reposted on

  2. At some point, will some districts figure out that approaching teacher hiring this way is guaranteed to mean the best teachers will go teach somewhere else ? Interviews are a two way street -- if you are being asked canned questions, or being put through a machine, what does that tell you about the district making these choices ? Everything you need to know.

  3. Philly uses this. They also use an in-house system called Site Selection. There is a committee, headed by the principal, vice-principal, counselor, dean of students, and some teachers. But mostly people who have been out of the classroom for a number of years or have never taught in a classroom situation at all. The principal writes the questions, with no input from actual teachers, makes the decisions on whom to interview and whom to hire. If a new hire doesn't work out, the principal claims it was a group decision, which everyone knows is a lie. The whole process is ridiculous since TeacherMatch is used to hire secretaries, NTAs, etc.

    1. It was used for teachers in Philadelphia last year.

  4. This video did not pass the sniff test with my adult mathematician son. He explanation of its fallacies is beyond my comprehension, so I cannot share.

  5. Just when I think it cannot get more inane, it does.

    Wish I had $575 to burn, I would crash their little party.

    I'm in Austin and I wish there were a way to find out who is attending this. Did you note that they have a form letter you can use to "justify your visit"?

    Stuff like this us why our schools ares are broke.

    Been reading you for years--thanks for your excellent work.