Sunday, February 20, 2011

Why do we pay teachers for advanced degrees?

First, to all of my liberal, union-friendly (and union-employed) friends and acquaintances-- please don't hate me. I'm an equal-opportunity skeptic; I'm skeptical of everything from homeopathy and astrology to fair-trade coffee and the power of unions to improve society.

So, Max in the comments to my previous post says:
1. Public employees in Massachusetts make LESS (total pay and benefits) than private employees with comparable education. The same is true is Wisconsin. You can look it up.

2. No "public good" for higher than average wages? Huh? Isn't that PRECISELY the way you attract good people to do important teach your children? Back to Finland: one crucial way they took an average educational system (as it was in the 1960s and 1970s) and made it the best (as it is now) was to dramatically raise the pay of teachers. That attracted lots of the best the ones who in the United States go right to law school or business school so they can become investment bankers.

Why the heck wouldn't you want to pay people who perform absolutely essential jobs a top wage?
Is there any research showing that teachers with more education (beyond a basic teaching degree) are better at educating students? All the research I've run across says that advanced degrees don't matter.

As for paying people who perform absolutely essential jobs more: I completely agree, as long as they actually do a good job. We have been paying teachers more and more and yet we're not seeing better results. We have run the experiment of throwing money at underperforming schools, but money doesn't work. Listening to a recent podcast from the Center for American Progress (CAP is a progressive think-tank) drove that point home for me (podcast here).

By the way: I can think of a lot of things that are MORE essential than learning where the private sector seems to do a good job with non-unionized workers. Starting with the food and shelter industries.

One of the things that drove me nuts working at UMass in a unionized position was that my job performance didn't affect how much I was paid, or what benefits I got, one iota. Salary was based on years of service, and the union, while I was there, was actively trying to DEFEAT a pay-for-performance proposal.

Maybe I'm weird expecting that if I do a good job I'll be paid more. Wouldn't we get better teachers if we paid them more when they did a great job?

I think employees should have the right to organize themselves to demand better wages and working conditions. But I don't think it is right that most public employees in Massachusetts must, by law, be unionized.


Anonymous said...

You are talking straight into the teeth of the orthodoxy in this town, Gavin. Good luck with that.

I think we have two problems in this country.

1) We don't pay teachers enough relative to other professions.

2) We can't seem to find a way to pay good teachers what they deserve relative to lousy teachers.

Check out the table accompanying Charles Blow's column in yesterday's New York Times for how far we are slipping relative to other countries on various measures of quality of life, including education.

We've peaked.

Abbie said...

As I have been told, in Massachusetts, within 5 years of receiving your teaching license you have to earn a MA. If this is true, then I don't understand why our district seems to use as a point of pride that many of our teachers have advanced degrees. Any thoughts? Corrections?

Steven Brewer said...

The central problem is that merit systems don't work. They encourage people to focus on maximizing the reward rather than doing what they think is right. This phenomenon has been extensively studied, and found over and over again. Read Alfie Kohn's Punished by Rewards for a good introduction.

At the root, there is another problem, which is that educational tests don't really "measure" -- and we probably shouldn't call them measures.

@Abbie: it might still be meaningful to show that Amherst schools can attract and keep teachers. Some schools are staffed almost entirely with new teachers, who leave as soon as they can.

Bimmerhead said...

I tend to think that raising the pay of teachers will in fact produce worse results. If you really want to have only the best teachers in the system, pay them nothing.

Pay increasing teacher salaries we are now attracting people who may be smarter, better educated and probably better looking. But are we attracting people who are better teachers? I think the best teachers are those who love teaching. They love kids, they love seeing kids learn and discover and be creative. These teachers aren't doing it for the money, they're doing it for the love of the job. Introducing money into the equation attracts the wrong people into the position.

Gavin Andresen said...


So.... "move along, nothing to see here" is the right answer because performance-based compensation is hard?

Nina Koch said...

Hi Gavin,

Can you suggest some measures that you think might work in implementing performance based compensation? For example, at the job where you felt you should have been rewarded for your performance, what were the indicators that you feel should have been used?

I don't think this is at all obvious, in any profession. A law firm might reward people according to billable hours. But what if someone's volunteer work in the community helps to forge goodwill toward the law firm? Does that get counted? Who decides what constitutes performance? This is going to involve value judgments.

Sometimes arguments that are made based on principle can be eclipsed by pragmatic considerations. For example, I could try to argue that the death penalty is morally wrong. People who already felt that way would agree with me; people who don't feel that way would not be swayed by my assertions of morality and no one would budge. Then I could reframe it: what is the probability that we have executed an innocent person? What level of risk are you willing to accept? Most people would say zero. Then I do a probability computation for them. If we are 99% confident that each person is guilty and we have executed 200 people, then there is a 1-.99^200 chance that we have executed at least one innocent person. That's 87% chance of doing something that nearly everyone considers unacceptable. So it boils down to pragmatics: do I think our criminal justice system can be made good enough to boost the 99% confidence rate to 99.99%? No. I don't think it's anywhere near 99% now. We don't have the resources to get it there. So let's just forget about the death penalty and do something different. Morality aside, we can't find a way to implement it. Pragmatics carry the day.

The same reasoning can apply to the arguments about merit pay. Even if you feel that it is justifiable on principle, I don't see how you can possibly implement it. It's not just that it's hard. It's that it would take such an enormous quantity of resources to implement it effectively that you might as well devote those resources more directly to education.

For example, I teach a class of mostly tenth graders in Algebra II. Later this year, they will take the MCAS. The content of the MCAS doesn't include Algebra II, so could I in any way take credit for the students' performance on that test? Probably not. The students are being tested on material that they learned over a long span of years in many different classrooms. Maybe something that Miss Finster said about fractions in third grade finally clicks for one of my tenth grade students when we are doing our unit on rational functions. Does Miss Finster get credit for that or do I? I really don't want to spend a lot of time trying to figure that out.

So even if we agreed that something like NAEP or MCAS captures what we value in a student's education, we are hard pressed to associate student performance with a particular teacher.

And of course, we don't all agree that those tests capture what we value. In fact, we don't even agree about what we value. I value students' ability to think creatively and independently. I value students' willingness to take intellectual risks and to tackle complicated problems that they haven't seen before. I value passion and curiosity about the subject matter. Is there an MCAS for that? Should we have them wear mood rings in class to help us measure it?

I also think that there is a faulty premise about the degree to which merit pay would serve as a motivating force for teachers. More on that later...

Gavin Andresen said...

First: I think the "right" solution is to get the government out of the "provide education" business. Have public funding of private education and all of these issues take care of themselves through competition and innovation driven by parents and students finding the right school that works best for them.

However, I know that ain't gonna happen. So how about asking students and parents and peers to rate teachers? I KNOW who the best and worst teachers are at Wildwood because I talk to other parents and I listen to my kids. I bet teachers and administrators know who the best and worst teachers are, too.

I believe I recall a study that showed that student evaluations of teachers matches student performance pretty well, so even just letting the kids say whether or not their teachers are good would be better than the "it is all about seniority and degrees" system we have now.

Nina Koch said...

I am not sure if it matters if we are talking about private sector or public sector. If you want compensation to be a function of several independent variables, then each of those variables needs to be a quantity that can be stretched out along a number line in a definite order. You should be able to describe a quantity where A>B and B>C implies A>C. I don't think this is easy to do at all but I would be interested to see someone take a stab at it. I guess one arena where we see performance-based compensation is in sales, where your salary is a function of how much you sell. Even with that, however, other factors creep in. Maybe one salesperson has a better territory than another. Does that get taken into consideration?

I think there are very few measurable quantities in the world of organizations. This is why it's difficult to rank sports teams. If team A beats team B and team B beats team C, does that mean that team A will beat team C? Not necessarily. If there were a sure-fire method for ranking teams, we wouldn't need to have March Madness, and arguing about sports would be no fun at all. I think 2008 is the first time that the NCAA Final Four consisted of four #1 seeds. Shouldn't it always be four #1 seeds? If we had a neatly measurable quantity, it would be. But we don't. We have multiple, complicated, interconnected variables and elements of uncertainty. I would say that you have that in most organizations.

I think that systems which attempt to assess performance relative to an average are especially ridiculous. Any time you compute an average, you need to acknowledge that half of the people will be below that average. That's what average means. If you get rid of all the people who are less than average and then compute a new average, you will yet again find that half the people are below average. It would make more sense to have some kind of absolute standard, if such a standard could be found.

So, Gavin, you did make the statement that you felt that you should have been rewarded for performance in your university job. Since you feel strongly on this issue, you must have some thoughts on how that performance would be measured. Similarly, your statement that you KNOW who the best and worst teachers are at Wildwood must be based on some criteria. Since the statement implies a high degree of certainty in your measurement, I would be interested to know what those criteria are.

Gavin Andresen said...

RE: getting rid of people below average:

Interesting you should bring that up-- when I worked at Silicon Graphics they had an explicit policy to fire the bottom-performing X% (I think it was 5 or 10%) of everybody in the company each year. You were measured by what you accomplished, and it was pretty clear which people and groups were getting things done and which weren't.

It was a great place to work-- lots of very smart, hardworking people.

As for criteria: you try to measure outputs, not inputs or effort. You try to evaluate how somebody affect the entire organization (I had colleagues that weren't the world's greatest programmers but who were great at helping make our entire team happier and more productive).

And you realize that change is good-- if somebody isn't being successful, then it better for everybody if they find another job where they can or will be successful.

I disagree that evaluating performance has to take a huge amount of resources. If it does, then there's something wrong with the company or institution.

Nina Koch said...

Here is a link to some interesting work being done by the Wisconsin Center for Educational Research:


It seems like the researchers are taking a thoughtful approach to the concept of compensation. I mostly worry about people oversimplifying the issues. That's why I have an adverse reaction to "everybody knows" type of statements. I think there are actually very few things that everybody knows.

I still haven't heard you articulate what you think the criteria are for judging teacher effectiveness. You must have some criteria because you have stated that you KNOW who the good teachers at Wildwood are. My guess is that if you listed and prioritized your criteria, it would turn out that other people have different criteria. That's what makes it difficult to come up with a formula for compensation which all stakeholders consider fair.

Even in an occupation where the output is reasonably well-defined, like creating software, I bet there are still significant differences in people's judgments of value. What about the person who writes code quickly but doesn't comment the code adequately because he or she thinks everything is obvious? That calls for a judgment.

Those judgments are even more complicated in the world of education where the products are not things but people. You have mentioned "student performance" a few times but I am not sure what you are referring to with that term. Did you just mean student performance on a certain kind of test? That's pretty narrow. We want much more than that for our students. Some of what we want for them won't really become apparent until they are adults. I would love to see some kind of longitudinal study that followed students through college and into adulthood to see what happens to them. I think it might be very revealing.

Gavin Andresen said...

RE: criteria for evaluating teachers:

How about either:

a) Ask the students to rate their teacher at the end of the year.


b) Ask parents to rank which class they'd like their children to be in next year (or if they have no preference).

Again, I think either of those would be better than the system we have now (he says, without knowing all the details of the system we have now-- it IS seniority and education based, right?)

Anonymous said...

There's no action on this blog until you start asking questions about teachers.

What do you make of that?

You've entered the No Standards of Performance Zone.

Anonymous said...

You can't agree on what you want from education.

You can't really evaluate us.

So, by default, just let us keep doing what we're doing. If it works, it works. If it doesn't, well,......we're all temporary until your kid moves on to someone else.

Nina Koch said...

There is a big difference between evaluation and merit pay. Of course there should be evaluation according to rigorous standards and every teacher should be working on ways to improve his or her practice.

Evaluation Plan

We all have things we want to get better at. There has never been a time when I have felt like "okay I'm done now; I'll just keep doing what I'm doing."

Merit pay is something different than evaluation. It is based on two premises, both of which I think are difficult to defend:
1) That you can quantify teacher performance with a single number;
2) That teacher behavior would change significantly given the presence of financial incentives

The Nashville Study suggests that premise 2 is not supported. Of course that is only one study which focuses on one narrow measure of student achievement, so I wouldn't call it definitive. It is just a piece of information.

I am only one person but I can tell you that the promise of a bonus would not be a source of motivation for me. I am a strong believer in the importance of intrinsic motivation, both for teachers and for students. The work needs to be meaningful. In fact, one of the things on my top ten list of ways I want to improve my teaching is to find a way to get students to focus more on a desire to understand and less on what their grade is. I want them to be life-long learners.

Gavin Andresen said...


I agree that just paying better teachers more won't motivate them or make our schools better.

I suspect that if you paid them more and gave them more flexibility and control that WOULD make a difference, but maybe not. I think that depends on whether effective teaching is a creative endeavor (I think the best teaching is) or more of a "you get the best results by following these exact best-practices techniques" type of job.

The research I've run across seems to show that the way to get good teachers is to be ruthless about firing teachers early in their careers. Hire 50 young teachers and expect to fire 25 of them in their first 3 years of teaching (all numbers made up, I am not an expert on any of this, maybe teacher quality is obvious after 1 semester of teaching).

Get a higher degree? Doesn't seem to matter.
Professional development? Doesn't seem to matter.
Small class sizes? Doesn't matter beyond 3rd grade.
Support from a teacher's aid? Doesn't seem to matter.

Great teachers DO matter, but, from what I've read, it seems great teachers are born, not made.

Anonymous said...

We havnt been throwing enough money at the system if teachers are still being paid in the twenty thousand dollar range. In mississippi, trash collectors are paid more than teachers. I wonder why we have the highest rate of illiteracy in the country.

There are 6.2 million teachers in the country. If you added 20 thousand dollers to every teachers salary, your problem would be solved, even though you probly dont need to do that for every teacher. In todays ecconomy, 60k should be enough for a teacher, especially provided that they get all those summer months off. In total this would cost 100 billion dollars, but this isnt much when you consider all the bullshit we spend money on. 600 billion on defence is just too much. 40 percent of the worlds military spending is too much. Let the rest of the civilized world take care of itself. And stop selling guns to everyone abroad for crying out loud. JEEEZE the world is F*****ED if we cant pay our teachers more than tax collectors and when all we car about is research on "AIRBORN LASERS" and god knows what other sci fi nerds wet dream of a defence project.

Gavin Andresen said...

Where are you getting the data that trash collectors in Mississippi get paid more than schoolteachers?

According to the Bureau of Labor Statistics, elementary school teachers in Mississippi are paid an average of $41,770 per year.

Refuse and Recyclable Materials Collectors are paid an average of $23,880 per year.

Larry Kelley said...

Geeze, I guess this demonstrates where teachers are gonna go to get out their positive message once Catherine Sanderson pulls the plug on her blog.

Anonymous said...

Their positive message of change or their positive message of "either love our work or take your kid elsewhere"?

Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...

As someone who's seen the internet, I'm all for better teachers. Maybe there's a genetic component to "great" teachers? Has there been a twin study?