I will protect your pensions. Nothing about your pension is going to change when I am governor. - Chris Christie, "An Open Letter to the Teachers of NJ" October, 2009

Friday, January 13, 2012

"Pay Teachers More, But Until We Do..."

We need a name for a certain species of education "reformer." The characteristics of this particular animal are:
  • He believes we really need to pay teachers more.
  • He has no plan for how to raise the revenues to pay teachers more.
  • Until he figures out how to pay teachers more, he thinks we need to fire bad teachers immediately. And replace them with good teachers, who will be paid with...
  • Uh...
For an excellent example of this reformy creature's thinking, let's check in with education tourist Nick Kristof:
This latest study should elevate the issue on the national agenda, because it not only underscores the importance of education but also illuminates how we might improve schools.
An essential answer: more good teachers. Or, to put it another way, fewer bad teachers. The obvious policy solution is more pay for good teachers, more dismissals for weak teachers.
He's talking, of course, about the Chetty, Freidman, and Rockoff study that the Times wrote about last week. This is undoubtedly only the first of many such punditations that will call for a radical upending of the American school system on the basis of a single study done in a single city that has not yet been published and peer-reviewed.

Both Bruce Baker and Matt DiCarlo have dealt with the limitations and the media's misuse of the study. But let's put that aside for a minute, however, and ask a simple question: what polices would Kristof actually implement if he could?
Suppose that the bottom 5 percent of teachers could be replaced by teachers of average quality. The three economists found that each student in the classroom would have extra cumulative lifetime earnings of more than $52,000. That’s more than $1.4 million in gains for the classroom.
Oh, it's that simple, huh? OK, let's just do that - just a few little problems to work out:

- How are you going to identify these bottom 5% of teachers? With unreliable standardized tests? Graded by low-paid workers? With high error rates? That aren't applicable to the vast majority of teachers? Doesn't sound like you're going to do a very good job of finding out who's at the bottom to me.

- Where are you going to get these new "average" teachers? Do you think they are lining up to teach in the most difficult schools for pay and job security based on unreliable standardized tests? Does that sound like a great career for the best and the brightest?

- What makes you so sure that getting kids to win better scores on bubble tests is what America needs to retain its economic supremacy? Are the 21st Century skills this country craves to be found on Scantron sheets? Remember: this study tracked earnings only until age 28 and in only one city; is that really good enough to extrapolate to our entire economy?

- The "lifetime earnings"in this study amount to about $250 per student per year. Seriously.

Kristoff doesn't much seem to care about this stuff; he's got bigger, stinkier fish to fry:
One of the paradoxes of the school reform debate is that teachers’ unions have resisted a focus on teacher quality; instead, they emphasize that the home is the foremost influence and that teachers can only do so much. 
That’s all true, and (as I’ve often written) we need an array of other antipoverty measures as well, especially early childhood programs. But the evidence is now overwhelming that even in a grim high-poverty school, some teachers have far more impact on their students than those in the classroom next door. Three consecutive years of data from student tests — the “value added” between student scores at the beginning and end of each year — reveal a great deal about whether a teacher is working out, the researchers found.
[...] 
The blog of the Albert Shanker Institute, endowed by the American Federation of Teachers, praised the study as “one of the most dense, important and interesting analyses on this topic in a very long time” — although it cautioned against policy conclusions (of the kind that I’m reaching).[emphasis mine]
Kristoff is so eager to indulge in a little union bashing that he admits to ignoring what Matt DiCarlo wrote just so he can catch a teachers union engaging in... honestly, I don't know what. But it must be bad, because its a teachers union, and all serious people know they are the root cause of all of our problems, right Nick?

So, yeah, we'll get around to that "array of other antipoverty measures" real soon - no really, I promise. And we'll come up with all that extra money for good teachers just as soon as we figure out how to raise it without taxing the rich, because everyone knows that's just impossible.

Until then, however, we can do exactly the opposite of every other developed country in the world and begin overemphasizing bubble tests. And I'm sure lots of bright, young people will show up to teach if we just keep talking about paying them more. 

7 comments:

Tom Hoffman said...

I'd like to write an agent-based simulation of various theories and systems of teacher retention, just to demonstrate that they're all expensive and most of the reforms have rather obvious flaws.

Duke said...

Tom, that would require actual thought and research and stuff.

Heaven forbid...

Anonymous said...

Why doesn't merit pay solve the dilemma of compensating better teachers more and weeding out the overpaid lousy teachers? When you wash out the 100k deadwood (Duke's term) you'll have more pay for the underpaid all stars. There is a finite pie....is that not obvious?

Oh, right. An industry 100 percent based on testing and grading is impervious to testing and grading of teachers.

Sorry, this blog can really bring out the sarcasm font.

schoolfinance101 said...

Tom is right on target with the agent based simulation approach. An alternative, more structured approach (but still involving system effects, feedback, etc.) would be a system dynamics model. I used to dabble much in both and even wrote sections of a book explaining how to use system dynamics models to understand education policy & school management problems (flop). I had the pleasure of attending a really freakin cool presentation (around 1996) by Chris Langton, one of the founders of agent based simulations (SWARM project).

The problem is that even though these technologies have become so accessible and accepted in their own circles it remains difficult to get academic articles in policy and economics published using these approaches.

So, we continue on our way of providing meaningless linear (or other preset functional form) extrapolations w/o feedback and holding all else equal. This is a classic case of a massive data crunch paper with much to add but still very far to go to provide meaningful insights into how any of it would play out in a complex, dynamic social system.

It's just baffling how little we've moved since I began my academic career around 1997.

Duke said...

I have never used the term "100k deadwood." The notion that there are so many bad teachers making six-figure salaries that it is a serious impediment to student achievement is based on anecdote and has never been confirmed by research that I have seen.

Yes, there is "deadwood" in teaching - like every other profession. While I am certainly for policies that recruit excellent candidates, help current teachers improve, and dismiss poor teachers, I think it's foolish to believe you can do any of that without increasing teacher salaries.

Reformers want to live in a fantasy world where the rules of the labor market don't exist. Like you, they imagine there are boatloads of terrific teachers just waiting in the wings, ready to earn the SAME amount of money as current teachers, with fewer job protections and less security.

It's silly. If you want better teachers, you have to pay more. Firing using unreliable measures without increasing compensation will not work.

"An industry 100 percent based on testing and grading..."

You have stumbled upon the biggest problem in education today. Any good teacher will tell you that testing and grading are only part of assessing.

Duke said...

Tom and Bruce: to be honest, I've never heard these terms. I'l try to look into this soon.

The larger point, however, is well-taken. We have a movement to collect a mound of data led by people who don't seem to have a clue of what to do with it.

Professor, I think there are several factors involved in why we've moved so little, and it extends to many other fields:

1) Academic research is being emulated by the think-tanky (love that term!) world, and journalists, pundits, and public intellectuals have foolishly given that stuff equal weight to peer-reviewed research.

2) Those same gatekeepers and public officials have not been well-trained in statistics and research methods, so they have what can probably be best described as poor taste. This speaks directly to your contention that education leaders need to acquire tools to help them weed through the sea of junk out there to make well-informed policy choices.

3) Too many of your colleagues do not believe they should engage in the public sphere; too often, they are not rewarded when they do.

4) When they do engage the public, as in the case of Chetty et al, they fail to account for the highly-charged political atmosphere in which we now live.

5) There is a concerted effort in this country by those who hold power to demonize rational thought wherever it may be found. 30 years of conservative propaganda has convinced large numbers of Americans that pointy-heads do not have their best interests at heart, while corporations do.

The last, best defense against all of this is a free, high-quality, fully accessible system of public education that trains children to become citizens capable of critical thought.

Thus the battle we see before us today.

Anonymous said...

Would someone please translate "...agent-based simulation of various theories and systems of teacher retention..." into human English. Another sentence that needs translation into simpler English for my dense brain is: "An alternative, more structured approach (but still involving system effects, feedback, etc.) would be a system dynamics model." Huh?