The end of Michelle Rhee

by Andrew Gelman on May 20, 2013 · 34 comments

in Education

Many of you are probably aware of this already, but it was news to me. It’s about education reformer Michelle Rhee. I’ll hand the mike over to Mark Palko:

Rhee’s record mainly shows a pattern of intense self promotion, often the expense of students:

She appears to have started her career by greatly overstating test score improvements during her Teach for America days;

As an administrator, she was charged with abusing her authority to political ends:

and covering up a major cheating scandal;

She lent her political capital to anti-labor measures only tangentially related to education (but vital to her allies);

She oversaw the creation of a convoluted metric that assigned the top ranks to schools she and her allies were responsible for (despite those schools’ terrible performance on the very metrics Rhee had previously championed);

And she endorsed a Bobby Jindal  initiative which pretty much guaranteed wide-spread fraud.

Following the links, I found some further discussion here and here from a couple years ago that makes Rhee look pretty bad.

Palko describes Rhee’s success as an “affinity con”:

Affinity cons work in large part because when people see someone with similar background and cultural signifiers, they assume other similarities: common goals, values, approaches.

Movement reformers, particularly those who came in through Teach for America (and that’s something you see a lot) often get sucked in by something similar. They look at someone like Michelle Rhee and the rhetoric and the resume feel familiar. . . . Lots of leaders in education today have that exact same bio and since the vast majority of them genuinely care about kids, they assume Rhee does as well.

I don’t know about that. I’m just speculating here, but my take on it is, if you give someone good public relations, a lot of money, and a message that people want to hear, he or she can go pretty far before getting tripped up by reality. Palko refers to Rhee’s middle-class background, but I think if she had a lower-class background, that would’ve worked fine too. Look at Barack Obama: his background was different from almost all Americans, black or white, but people just ate up his story.

As for Rhee: I suspect she’s not planning on going anywhere, but all this error, corruption, and cover-up is taking a toll on her reputation. To the extent that her movement is about education reform rather than about Michelle Rhee, at some point they’ll have to find a more credible leader, no?

My impression is that there has been a shift. A few years ago, value-added assessment etc was considered the technocratic way to go, with opponents being a bunch of Luddite dead-enders. Now, though, the whole system is falling apart. We can learn a lot from tests, no doubt about that, but there’s a lot less sense that they should be used to directly evaluate teachers. We’ve moved to a more modern, quality-control perspective in which the goal is to learn and improve the system, not to reward or punish individual workers.

This shift may have not happened yet at the political level, but it’s my sense that this is the direction that things are going. The Rhee story is symbolic of the fallacies of measurement.


Irfan Nooruddin May 20, 2013 at 12:04 pm

That Michelle Rhee is a fraud is hardly news to anyone who has listened to her speak. She’s an anti-union troll and little more than that.

The real tragedy of the Rhee story is the willingness of supposed progressives to jump on her bandwagon. To alter your conclusion, to me, the Rhee story is symbolic of the weakness of “liberals” on their fundamentals as they relate to schools (see Obama and Arne Duncan’s repeated efforts to move to the right on education).

Mike May 20, 2013 at 12:09 pm

We can learn a lot from tests, no doubt about that, but there’s a lot less sense that they should be used to directly evaluate teachers. We’ve moved to a more modern, quality-control perspective in which the goal is to learn and improve the system, not to reward or punish individual workers.

There’s no money in learning and improving the system, though. All the corporate backing for reform was based on the desire to reward and punish individual workers. So if this is true, all the grandiose overhauls should grind to a halt.

Matt May 20, 2013 at 12:48 pm

Is it actually true that testing doesn’t provide good information on teacher performance? My understanding is that the year on year variance is pretty high and year to year teacher rankings not super stable, but that running averages over 3-5 years ended up being very informative with teacher rankings pretty stable. Lots of issues and I’ve got mixed feeling about the various parties in the ed reform debate, but just as a statistical issue?

Andrew Gelman May 20, 2013 at 2:29 pm


My impression is that testing does provide good information on many teachers, but that things will start to fall apart if you use test to evaluate teachers. Noise, incentives, governance, etc.

Kate S-L August 20, 2013 at 11:54 pm

Thank you, Andrew, for your excellent explanation.

Matt, there are too many variables involved in testing that are out of the control of the teachers to allow it to be a valid and reliable measurement of teacher performance.

Tests should not be used to evaluate ANYONE, not even students. Tests should inform instruction. Anything else is a perversion and misuse of the tool.

Tracy Lightcap May 20, 2013 at 4:42 pm

This is education we’re talking about; i.e. the home of Campbell’s Law.

The whole process is hard to define, much less measure, and running experimental trials (real world ones, that is) has run up against the steadfast refusal of teachers to vary treatments randomly in their classes. The result is fads a go-go and, since this is America, a debilitating tendency to substitute new techniques for thinking about the problems. It was inevitable that our politicians would embrace Rhee and the whole use-testing-to-determine-whose-career-gets-smashed-and-which-neighborhood-has-its-school-closed mystique. It was also inevitable that the entire process would get corrupted nation-wide for reasons just enumerated.

This attempt to avoid raising taxes to provide decent education – that’s what’s behind it, of course – was bound to come a cropper. Query = will the public wake up soon enough to insure that the other nations presently eating our lunch in educating their populations solidify their lead in ways that seriously impact our economy? Oh, the suspense.

Lenny May 20, 2013 at 4:54 pm

This attempt to avoid raising taxes to provide decent education.

Early in your comment you astutely diagnose the prevalence of fads in the field of education and then you invoke this faddish belief as being the prime driver of the other fads. That seems a tad outlandish to me.

mpledger May 21, 2013 at 7:58 pm

The rise in education costs from the 1970 is in large part due to special needs education.

Kate S-L August 20, 2013 at 11:57 pm

mpledger, I wonder how much of the rise in special needs education services is due to misidentification and mislabeling of students…

Mark Palko May 20, 2013 at 5:01 pm

In terms of affinity, I’ll admit the class component is pretty small. The Ivy League /Teach for America is the primary component. TFA alumni in particular are disproportionately represented in the reform movement and have a strong sense of group identity.

Rhee has been very good at shaping her persona and framing her argument in ways that appeal to the group.

ceolaf May 20, 2013 at 5:47 pm

Wow. If only…

But none of this is new. And there have not been a lot of signs that things are changing on the other stuff either.

They embraced Rhee because she said what they wanted to hear, not because she was like them. This is not an affiniity thing. This is a confirmation bias thing.

There have been deades of complaints about how tests narrow curriculum. And we’ve been putting more and more weight on them. We not pressured the assessment industry to produce better tests. There’s no real sign of a change there — unless you think a few parents in Park Slope making better use of the internet to publicize their views counts as progress. (Yeah, hyperbole.)

I was just at NCME (National Council on Measurement in Education)’s annual meeting, a couple weeks ago. No signs of shifts there. More emphasis on autmated scoring, even though those products are not even close to ready for primetime and actually force content specialists to dumb down the content of the tests.

Value-added? It’s gaining more and more steam, without any signs of a loss of acceleration, even. Teachers being evaluated with this, schools being evaluated with. Districts. Teacher preparation programs. It’s not slowing down at all, let alone turning around.

Union bashing. Still strong. Getting stronger.

The only change I see is greater emphasis on cheating. Journalists are looks there more. Psychometricians are thinking about it more (perhaps because the outgoing president of NCME cares about this topic?). And districts are starting to pay more attention. But that’s not a policy shift. That’s simply paying more attention to how to properly implement the same policy umbrella we’ve seen.

Anon May 20, 2013 at 11:19 pm

I think this is a biased, careless, and naive assessment of Rhee. I mean one of the linked sources is the WAPO colum “Class Struggle”, and so on.

Do you know what Rhee is up against? A well organized, sophisticated, and rich teachers union. So yes, character assassination comes with the turf.

So here, lets balance the story from the other side. The truth – remember that? – is probably somewhere in the middle.

Andrew Gelman May 21, 2013 at 7:27 am


Being up against a well organized opponent doesn’t mean you can’t be involved in cheating and fraud.

Andrew Gelman May 21, 2013 at 9:11 am

To say it another way: The education reform movement isn’t going away. But I think Rhee’s usefulness for this movement is going away. They should be able to find a leader who is not so covered in scandal.

Anon May 21, 2013 at 10:10 am

Whatever happened to the presumption of innocence? You should say allegations of cheating and fraud, and note that many of those allegations are hardly from dissinterested parties.

So take the first link to WAPO colum on her exaggerating her performance. To me it was no slam dunk, least of all considering the source. As a Bayesian maybe you should start with a prior (presumption of innocence?); observe the data (claims, perhaps discounted for reliability); and state your posterior probability interval that she exaggerated. If binary, exaggerate or not, state the posterior is 0.6 CI .3 .9 etc.

But of course this is not binary, we all exaggerate a little. So the question is if real scores increased by x but she claims x+b, what is substantive magnitude of b, and CI. I already know b is not 0.

I do not think she is getting a fair treatment. I know this is a blog, but presumably one run by social scientists. For us fairness and even handedness light to be second nature.

Andrew Gelman May 21, 2013 at 10:15 am


I’m not talking about throwing Rhee in jail or even fining her. I’m talking about trust. Given that the education reform movement is all about measurement, I and others are troubled by a series of scandals, all of which involve measurement biases.

Anon May 21, 2013 at 11:12 am

Andrew, with all due respect I think your attitude is one of surrender to smear.

If tomorrow you were appointed to reform schools in NYC I bet you within 2 years you’ll be covered in scandal. That is politics. And all of these scandals will involve measurement biases, after all that is what one camp wants to avoid, as well as ad hominem critiques. I’d like to think I’d be there to defend your integrity, not to cast you away as tainted by scandal.

Most democrats don’t yield to IRS smear of Obama but when it comes to an ideological opposite, all they want is an excuse to let go. I try to leave ideology out, and look at Rhee as a human being in a difficult fight. I don’t know much about her, and maybe all claims are true, but my strong held prior is that the probability of scandal , conditional on being innocent, in a position like the one she has been in is pretty close to 1, 95% CI .95 1

Now you might say this is all nice but she is tainted, innocent or not, and hence ineffective. I say if you go by that you create a dangerous precedent, that all it takes is some smear to get an official removed / forgotten. I refuse to live by that, we all deserve a fair hearing, and i will fight for it.

If you haven’t done so already pls read the The Lost Honour of Katharina Blum.

Andrew Gelman May 21, 2013 at 12:07 pm


The cheating scandal looks pretty bad to me, and so does Rhee’s happytalk response to it.

Yes, it’s true that some honorable people get slammed, and other honorable people make unfortunate mistakes. But there are also political figures who are not so honorable. At this point, I don’t think defending Rhee does any good to the cause of education reform. Her priority seems to have been, first to get lots of good press, then to raise lots of money. One way to get good press is to exaggerate the good news and bury the bad news. In this case, though, the bad news eventually came out, reported in USA Today and other outlets. And the responses from Rhee and her organization do not look convincing.

I could see the argument being made that Rhee, as is said in the article you linked to, “prioritizes the needs of schoolchildren” and that she’s playing a larger game. Sure, there were cheating teachers in DC, lots of the claims based on test scores are false, etc etc. But if Rhee truly believes in the larger goals of education reform and defeating teacher unions, she could feel that some misrepresentation and cheating along the way is a small price to pay. Politicians have to make compromises, and I respect that. But then Rhee is playing a different role, not the technocrat that she’s been presented as, but more of a wheeler and dealer.

Anon May 21, 2013 at 7:06 pm

Andrew, I looked at the cheating scandal in USA T, and the article is, in my humble opinion, a collection of unsubstantiated claims with innuendo.

The investigator hired by DCPS to investigate cheating claims to have found some evidence of cheating but qualifies it: “Much of what we think we know is based on what I consider to be incomplete information. So the picture is not perfectly clear yet, but the possible ramifications are serious”. Yet in all that follows this investigation is taken as if it where a RCT, even if separate investigations by DOE and DCPS found no evidence of widespread cheating. Franklly, w/o the raw data and the ability to replicate the analysis I don’t know what to make of this.

In all this the implicit claim is that (a) Rhee was aware of the study. In this John Merrow, who “leaked” the study, has no doubt: “Michelle A. Rhee, America’s most famous school reformer, was fully aware of the extent of the problem when she glossed over what appeared to be widespread cheating during her first year as Schools Chancellor in Washington, DC. ” — based on the testimony of two unnamed sources; (b) she ignored the problem to avoid trouble, pump up here performance.

It all sounds very plausible, but let’s see. First, Henderson, in my quick view, is a defender of the status quo: “the recipe for success includes some mix of strong leadership, committed teachers, an integrated curriculum, the willingness to challenge conventional wisdom and accepted practices, and the moral imperative to care for and about all of our children”. Aka more of the same but trying harder, with better intentions. Second, why would Rhee decide to hide the cheating? After all many teachers hated her, and since they were in the know they could out the scandal at any time via anonymous whistle blower. It’s kind of handing your enemy a dagger.

Again, I do not know enough about this, and you may well be right. But it is perhaps no coincidence that all the attention is on Rhee, and not on the school principals, teachers, and teacher union financed politicians (

The real scandal here is that school performance remains disastrous, especially for poor kids.

Eli Rabett May 22, 2013 at 10:00 pm

Go read Diane Ravitch on Rhee, and yes there was/is a real cheating scandal in DC unless you believe in the tooth fairy eraser.

mm May 23, 2013 at 12:57 pm

Diane Ravitch is hardly an unbiased source on Rhee.

Thorstein Veblen June 21, 2013 at 7:54 pm

OK, then watch Rhee on Frontline. She incriminates herself.

Rhee hired a company to produce a report on the (suspicious) large score increases, and that company told Frontline Rhee told them explicitly not to check for cheating. Rhee didn’t dispute this.

Rhee said she also “saw no evidence of cheating” in the schools in Anacostia with large score increases that got huge pay raises and were subsequently caught cheating.

Rhee’s a complete fraud.

ceolaf May 21, 2013 at 2:22 pm

What presumption of innocence are you talking about?

You just mean the cheating stuff, right? (I’ll get back to that.)

Her firing people on camera, for broadcast? We all know that happened, right? What presumption would be applicable there?

Her refusal to deal with the community relations part of her job, thus costing her benefactor his position as mayor? We all know that happened, right? What presumption would be applicable there?

Her false claims of responsiblity for score increase that occured befure her policies even had a chance to take effect? We all know that happened, right? What presumptions woul dbe applicable there?

etc., etc..

So, you just mean the cheating, right?

Well, presumption of innocence is something for our legal system. It is about criminal penalities. It is about fair trials being necessary BEFORE the state (i.e. government) meets out punishment.

But there is no criminal penalty on the table. And there is no state threatening any. So, your “presumption of innocence” is not really appropriate.

Which shifts us to a personal standard for judgment. We want the state to be sure beyond a reasonable doubt before punishing anyone. But the punishments or condemnation of individuals or non-state actors are not held to that same standard. We don’t need formal trials. The amount of process that is due in the political and publicity realms — what Andrew is talking about — is far far far far FAR less than in the criminal justice system. Furthermore, unlike criminal punishment, our condemnation can be reversed, addtionally suggesting the appropriateness of less due process and a lower burden of proof.

And so, the question becomes, “How much evidence do we need, or whose say-so do we need, to personall think her guilty of cheating, coverup and/or rigging systems to get the answers that suited her?”

Well, this stuff has been fairly well investigated. It’s not done and over. We don’t know everything. But those of us who have followed this stuff know quite a bit.

Which leaves really no rational reason to futher reserve judgment. There is only the political and agenda-driven, not rationality.

Tracy Lightcap May 21, 2013 at 2:50 pm

Right. The standard here is more like it is in a civil trial: a preponderance of evidence that the defendant is liable.

And, as you say, Rhee is toast on that standard, at least as a standard bearer for GERM (Global Education Reform Movement, hat tip to Diane Ravitch). It’s high time that we realized that most of the nostrums pushed by her and her allies are flapdoodle. And, as I said before, the best validation of Campbell’s Law I know of.

Oth, empirical research on teaching methods is, I think, on the verge of an explosion driven by the use of hybrid and, to a lesser extent, online instruction. It’s a lot easier to vary treatments in such courses and the amount of data available on each student will be orders of magnitude greater. But we’ll have to see about that.

mm May 23, 2013 at 1:03 pm

“You just mean the cheating stuff, right? (I’ll get back to that.)”
It is teachers and administrators accused of cheating, not Rhee.

“Her refusal to deal with the community relations part of her job, thus costing her benefactor his position as mayor?”
She met with the community extensively, and communicated more than any prior superintendent, answering just about anyone via email. She mainly met with racism (or accused of racism) in terms of her negative perception. The DC public schools were considered an employment system for friends and family, she broke that quite a bit.

Stop lying, start considering the other side of the situation.

The real problem is that she hasn’t hit back hard enough at the union movement. Her slogan “students first” means “teachers second”. They don’t like that.

ceolaf May 31, 2013 at 9:56 am

No lies. Lots of consideration.

But such accusations really work well end discussion.

KJ June 2, 2013 at 10:53 pm

““You just mean the cheating stuff, right? (I’ll get back to that.)”
It is teachers and administrators accused of cheating, not Rhee.”

Allow me to introduce you to the concept of respondeat superior, meaning that the leader is responsible for what happens on her watch. In tort law this means that the management is liable for the actions of her underlings when the suspect actions are a done within the scope of the underling’s employment. Rhee was in charge, and to paraphrase Stan Lee, with power comes responsibility. In this case, the responsibility was Rhee’s and she did not meet her duty as the leader of DC schools by failing to put in reasonable preventative measures and failing to address the problem once it was brought to her attention.

Steve Sailer May 22, 2013 at 1:29 am

Michelle Rhee fired a lot of black teachers, so white liberals in D.C. like her and blacks dislike her. In modern America, many things are becoming less complicated as everything just turns into a who-whom struggle. For a more detailed analysis, see:

Eli Rabett May 22, 2013 at 10:02 pm

“the steadfast refusal of teachers to vary treatments randomly in their classes”

Classroom teachers are not free to change curricula and methods of instruction. They teach to lesson plans that are approved by administrators.

ceolaf May 22, 2013 at 10:14 pm

1) Ummmm….


Some administrators insist on approving lesson plans, but many do not. I don’t know of any research on the topic, but my guess is that most teachers do NOT have their lessons plans approved. (Many administrators make sure that that lesson plans exist, but that is not the same thing as approving them.)

2) There is room for teachers to do research on their own methods and pedagogy, yes. But few teachers are taught to do research — unlike Prof. Gelman. They also have a far larger teaching load, must spend more time course in class and never have TAs to answer student questions or help with grading. Expecting them to do research on their own teaching — something that, as Prof. Gelman points out, faculty at RESEARCH universities quite rarely do — is simply not reasonable.

3) Actually, many teachers DO do research on their practice. Look up “action research” and you’ll find a ton of stuff. Well, a ton of tons of tons.

Is it high quality research? Not usually. I’m not sure that action research is the best methodology. I think that self-study might be a better approach for practicing teachers. (These are established methodologies, each with their own strengths and weaknesses.)

But without training in research, and support/encouragement of their employers and supervisors, even these methodologies would be difficult to do well.


And so, while there is a ton of tons of tons of action research, most of the 3 million+ k12 teachers in this country do not do it. There are a huge number of reasons why not, many of which are not within their control.

Eli Rabett May 23, 2013 at 12:34 am

Public school teachers are assigned and do not choose their curricula

ceolaf May 23, 2013 at 12:06 pm

Many public school teachers DO write their own curricula.

* Music and art teachers.
* In some schools, senior teachers write curriculum for the entire school (or grade).

States set standards in some content areas, but not all.

Some districts require some grades and some content areas to use particular curricula. And even in districts where this is not done, or not done as much, many schools (i.e. principals, assistant principals, department/grade level leaders) set the curriculum. (In some cases, school leadership actively undermines district-selected curricula.)

But few curricula include include individual lesson plans for each and every day. Fewer still are scripted, actually mandating the words to be said. Even fewer tell teachers how to answer students’ questions or redirect them to their work. None address discipline issues.

There is plenty of room — even with scripted curriculum — for teacher to research their own pedagogy. And the vast majority of teachers have far more leeway than those with scripted curriculum.

nancy raye June 30, 2013 at 11:50 pm

Michelle Rhee is a bitch with a capital B and will receive back everything that she has given to destroy professional reputations and livelihoods of educators. What comes around, goes back around. She is a LIAR along with all the city administrators and judges who stood behind her lies in all her firings.

AN July 28, 2013 at 2:23 am

Detroit,Baltimore and DC school systems are disaster zones. They are beyond fixing. They are a reflection of the corrupt and inept school boards, city councils and mayors. They are a
reflection 0f the corrupt and inept unions. They are a reflection of the most lazy and inept teachers in America. I think Michelle Rhee knew she was going to fail so she decided to make a name for herself before she got fired. Got to give her credit, she sure stirred the shit up good! As it turns out she is just as corrupt and inept as the rest. I forgot one, it stands above all the others the corrupt, inept, and useless Dept. of Education.

Comments on this entry are closed.

Previous post:

Next post: