15 February 2010

No Merit Against Merit Pay

Here's why teachers, by and large and as a profession, suck like Hoovers.

From Education Week, where veracity and non-verisimilitude would be best served by judicious editing to Education Weak, comes this "attack" on merit pay for teachers. The salient points of the puny popgun wielded by Kim Marshall are presented below, with My responses between brackets.

Set up: Here's what the author says immediately after firing blanks: "These concerns, to my way of thinking, demolish the argument for individual merit pay."

Okay, now find out Marshall's "way of thinking" and his ignorance of the meaning of the word "demolish": (Don't worry: I'll educate him on that as well.)

Is Merit Pay the Answer? ...Here's why (it isn't):

It undermines teamwork. Teachers who are rewarded for their own students’ test-score gains are less likely to share ideas with their colleagues. [Oh, like what We have now is the epitome of "We're really a family" teamwork. Your contention is that teachers are more interested in their own results and would avoid collaboration to serve their students better. Here's an idea: Let's fire those who do that.]

The best teachers are already working incredibly long hours, and there’s no evidence that extra pay will make them work harder or smarter—or that it will motivate mediocre teachers to improve. Quite the contrary: Merit pay will steer all too many teachers toward low-level test preparation. [Hmm, hundreds of years of experience across several hundred professions indicate that the chance to earn more money motivates many people. So if the best teachers are already working long hours to be the best, what the hell is wrong with the others? Lack of motivation? Okay, then how about merit pay?]

Standardized tests are often “instructionally insensitive”—that is, they’re better at measuring students’ family advantages and disadvantages than the school’s or the teacher’s value-added effect. [True, but Teacher A gets great results and Teacher B in the same school is a sack of crap. Blame the students? The parents? The school board? The janitors? Mountains of evidence show that great teachers can appear in any school, regardless of its socioeconomic environment. And standardized tests are not--should not ever--be the end-all and be-all of education.] 

Standardized tests in many states don’t put enough emphasis on writing and critical thinking, so raising the stakes for teachers creates an incentive to shortchange these important life skills. [Literacy rates in the U.S. of part of A. have dropped in the past 35 years. In fact, the U.S. of part of A. could have as much as 35% functional illiteracy. An emphasis on standardized tests may be a factor, as well as TV, video games, iPods, cell phones, El Niño, Tupac's death, Ritalin and sightings of the Virgin Mary on taco shells. Literacy has dropped because teachers are no longer trying as hard to teach these basic skills, maybe because they lack them (all too common) or because teaching them is never anything but the result of hard, persistent effort.]

To address the last two problems, it’s been suggested that schools should use higher-quality, before-and-after tests in September and May to measure each teacher’s contributions to student learning. Nice idea, but experts say it takes at least three years of data to produce a fair value-added measure of individual teacher effectiveness. [First point: Three years are going to go by anyway. Second point: To get to three years of data, start with one, then keep going. Third point: Suspend all judgment on the data until you have the three years' worth. Fourth point: You have got to be kidding. Is this what passes for reasoned discourse in your brain?]  

Raising the stakes on tests increases the urge to cheat. Most teachers are scrupulously honest as they proctor their test-taking students, but higher stakes will result in more thumbs on the scale. [It's already happening and will continue to do so until the utter stupidity of "No, Child: Left Behind" is dropped in favor of actual education.  By reducing the inane emphasis on standardized tests, the "incentive to cheat" would dissipate. This is not a merit pay problem, it is a standardized test problem. Try, really try to keep your criticisms straight.]

A good many students are pulled out of regular classes for small-group help with other teachers. How could we figure out a fair way to dole out merit pay for these children’s achievement? [How do you measure the value of a bunt that moves the runner up, of a pass that sets up a dunk or a thunderous block that springs the fullback for a long gain? It's called statistics, based on data and measurements, worked on over time until a framework is developed that assesses values to different actions. In this specific case, both teachers would have to collaborate to achieve great results so that both could qualify for merit pay. No other scenario benefits both or either one. Think it through. Have someone help you.]

Good scores in one 4th grade class (for example) would boost that teacher’s pay—but what about the 3rd grade, 2nd grade, 1st grade, kindergarten, and preschool teachers who helped those students along the way? Don’t they deserve some of the loot? If so, how would we calculate their share? [Jesus F. Christ in a bottle! Are you really this stupid? "Loot"?! If the students improved significantly in the 4th grade, it's very likely it was because their 4th grade teacher helped them do so. If the other teachers did equally good work in earlier grades, they deserve higher pay as well. But if the students had "average" results for their K-3 years and then "overachieved" in the 4th grade, pay that teacher for that merit and let the other teachers work better to earn their merit. This is the same "regression to the bottom" that plagues this profession so unwilling to be measured for results. Let Me reverse the argument for you to show how moronic it is: If you did high-quality work in the 4th grade and the same students bombed in the 5th, should you give back part of your "loot"? I thought so.]

Fully half of teachers work with grades and subjects that don’t have standardized tests—kindergarten, 1st and 2nd grades, art, music, and physical education, for example. Is it fair that they aren’t eligible? [Wait. A fucking. Minute. Didn't you just expend 4 of your blithering points arguing against standardized tests and now you are saying that it is unfair--unfair you say?--that "fully half of teachers" won't get the sublime benefit of being measured by them? And your "point" that only by standardized tests would a teacher be eligible places--once again--the same notion you argue against as the criterion for merit pay? Look, you may have had a wisp of a shadow of a sliver of a notion of an idea possibly maybe floating somewhere near a neuron in your head, but it faded quickly into pure vapidity. The only reason I'd give you a D- instead of F on this piece of numbskullness is that your handwriting in crayon was quite neat and legible.

The Jenius Has Spoken.

[Note: 10 March 2010: From Newsweek, Why We Must Fire Bad Teachers. Short version: accountability and children's futures.]


The Insider said...

How about de-merits? All teachers enter a two level, 10 year tenure program.

After 5 years, they are evaluated and achieve what we're refer to as level one tenure, more specifically a pay raise with a designation making them harder to fire in the future, and more qualified to move between jobs (including qualification for school management, etc).

After the next 5 years, they are evaluated again. If they meet the qualifications, they achieve tenure level 2 (same principles as tenure level 1, including a raise and high qualifications achieved).

If they fail either tenure level, they are reviewed again in 3 years.

As for wage increases, they receive them only as adjusted for inflation, localized cost of living standard for their region, with economic factors considered.

Failures to achieve tenure delay their pay increase, and are marked permanently on their record as failure to meet the standard. For instance, a teacher who failed level 1 or level 2 tenure multiple times does not have the credentials to move into a faculty leadership or school management role, and should not be teaching special needs or honors programs.

Repeated poor performance and failure to achieve tenure levels one and two are grounds for potential dismissal by the school's administration.

If the administration keeps a failing teacher on staff, it will be known on record and subject to the scrutiny of the school board and parent teacher groups, etc.

Any teacher that fails tenure and is kept on staff should be required by their own school administration (if kept on) to take additional continuing education courses to correct particular issues identified in a peer review.

So rather than merit based pay, with year to year evaluation, based on standardized tests, we take a more hybrid approach.

We make merit based gains in salary and qualifications more of a long term commitment, less focused on year to year standardized tests that are subject to manipulation and highly controversial in the first place. If a teacher wants to accelerate their career progress, then they need to better prepare their students to meet the challenges of the next grade they enter, year after year on a consistent basis - which is what they should desire to do, and be qualified to achieve in the first place!

Now before, we get onto how they are actually evaluated, let's cover how tenure might be decided.

Basically, we run the performance criteria against all teachers eligible for tenure each year. 75 percent of teachers performing above the bottom level achieve their tenure level 1. Only 50 percent of those eligible achieve tenure level 2. It is only run for teachers who have met their 5 year date for evaluation (or their 3 year reevaluation).

Now the criteria that we use to evaluate them is the final part. This is going to be a new approach - and it is going to require technology to achieve. Basically a new method of crunching data.

Instead of running the numbers on the standardized tests year to year, we *follow* each student who has a particular teacher throughout the life of that student...


The Insider said...

(part 2)

Once a teacher has been responsible for a year (or semester) of a student's life, they bear some associated responsibility for their progression through the rest of the students studies. That is, if they prepare them adequately for the next grade, then that student should perform better.

Since the students will take multiple different paths (i.e. different schools, different teachers, different states, etc), a statistic model can be developed that attempts to identify correlations between students and their previous set of teachers.

Example: Mr. John Doe teachers 3rd grade. He's an amazing teacher. Committed to the students, and highly effective.

His students pass grade 3 to go on to grade 4. Some of them go on to teachers A, B, C, D, E, etc. Those teachers vary in their effectiveness as teachers. However, statistics indicate over 5 years that Mr. Doe's students perform at an improved level in each of A-E grade 4 teachers regardless of the relative effectiveness measured by each of those teachers.

We can track Mr. Doe's students as they move through all the grades.

I'll leave the statistics up to the stats masters. However, we might think of one model that has applicability on such a large data set. That is the "Google Page Rage" algorithm that was the foundation of the world's most successful search engine... successful for being able to have a high degree of accuracy in finding true correlations between "your search query" and "the info that matches your query". Stats geeks - you figure it out.

Now, if such a system that can correlate a student's achievements with earlier teacher contributions, it will be a much better way of finding out which teachers are successful at doing the job, i.e. preparing students to go on to the next level of academics successfully.

The algorithm allows us to pinpoint their success. The tenure system allows us to reward the best of the lot and creative a competitive environment for professional improvement. It also provides us with a way to dismiss (or not hire) teachers who consistently fail to meet the standards.

Example: If Jane Doe has failed to achieve her level 1 tenure after 3 tries (5 year review, 3 year reevaluation, 3 year evaluation = 11 years), and she is unable to rank above only 25% of her peers (who include brand new teachers) they we have a valid reason to question whether or not Ms. Doe should be responsible for the education (and hence the whole lives) of the students who would be (unfortunately) placed in her care.

I guess I'm beating the hell out of this idea. But I do want to add that the top 10% of high performance teachers can also indicate their "best practices" in achieving their success which can lead to understanding how a truly effective teacher is made, what other teachers may do to improve, and how we can make incremental improvements in the entire education system, year after year.

I guess I should have blogged this separately. Alas it will become one of Gil's longest comments. ;)

Apologies for any inconsistencies. It was a fleeting idea I felt like recording here. I understand the model would be complex. Fortunately, it would be the computers handling the problem (not the people), therefore making it more viable.

Gil C. Schmidt said...

Holy Mother of Merit! Okay, Insider, the major strengths of your ideas are that you are looking to measure teachers over time (allows for gains in experience and added knowledge), extends job security without making it impossible to dump the ones who slack off or screw up and that you threw in a very successful data-crunching aspect of Our new Overlords, the GoogleTubes, Long May they Reign!

Weakness? You give teachers up to 11 years (roughly) to do a sub-par job. That's practically a generation of students (K-12) and I feel that We should be able to dump the failures after less time, such as 3 years.

But again,(sucking up to Our new Overlords) your Google algorithm idea is brilliant! I'm glad I thought of it and will be sure to remind the GoogleTubes Kommisat about it at the next pogrom, er, meeting. Thanks!

The Insider said...

Ok, rapid refinement. =)

Google-style correlation algorithm, theoretically, can give real time feedback on performance. Of course, validity of data increases teacher's time in system increases.

However, that does allow for individual teachers to view their performance in a near real time fashion (although data after each year completes will be most useful). It also allows faculty to review the same. And theoretically, it also allows every citizen of the United States to review the data publicly.

The faster feedback helps to show individual teachers whether or not they are on track without waiting for 3 to 5 years. It also allows management to do the same, and pinpoints those that need work earlier. The extra help should be welcomes since it relates to a long term goal that has big impact in their career.

I was identifying a bunch of flaws while flushing out the brainstorm... although it would not have been as fun to list those. ;)

However, you're definitely right about it taking a little too long to flush out the bad ones. The levels could broken into more than 2 steps, and the time line compressed.

However, another thing to note is that (I believe) it is more important to have the strongest teachers in K-6, the formative years, while 6-12 the students are better able to build off their own strong elementary foundations (even in spite of having to sit through sessions with sub-par teachers).

Therefore, we modify the program to protect those K-6 students. How? Eligibility for those positions goes as follows:
- Tenure Level 2 + Masters
- Tenure Level 1 + Masters
- Tenure Level 1
- Masters

This would make these positions less available to teachers without greater levels of qualifications, although it would ship some of the sub-par teachers into the 6-12 end of the program. If we have to have them in the system at all, at least we should have their impact distributed to students who have solid foundations and are at a level that they can make some level of progress in the classroom, even despite the teachers.

I would like to say that competition for these levels could also be increased by boosting the salary in the startup years (of the data collection), but I doubt there is any budget for that.

Now - we can also work in another way to weed the sub-par ones out of the system:

Provide an additional signing bonus for Armed Services sign-up, especially for teachers. If they are not good at leading (the classroom), then perhaps they might be more suited to following. They'll already have university level education behind them, and may do very well in the forces. This bonus will serve as encouragement for them to make the career transition.

Another way is to provide scholarships for those who choose to leave the teaching profession and pursue education in a new area, perhaps technology (since the USA wants to be a long time leader in this sector).

For those sub-par teachers in families with enough income to allow them a career change, the scholarship will be an additional incentive to take them out and keep them away from the kids.

For teacher's failing to achieve two consecutive tenure levels, let's make them ineligible to teach at the K-6 level unless they achieve their masters. If they have a masters beforehand, they are simply ineligible.

Gil - knock a few holes in my revisions if you'd like. I challenge your readers to do the same, or provide some alternative solutions that are outside of the existing spectrum of proposed solutions. ;)

P.S. Google says: "All your O are belong to us."

Gil C. Schmidt said...

Insider, I like your refinements and additions, especially in terms of "life experiences" and focusing on strong K-6 teachers. However, My mom, a 30+ year veteran of grades 7-9 would say that those are the "tough transition" years and that really good teachers there can help overcome the common decrease in parental supervision and stave off desertion. Teachers in grades 10-12 will also pitch for their "group" to be favored with the best and the brightest. I'd settle for simply having a rise in the average competence of teachers.

As you will see, My latest post (17 Feb 2010) gives you--aaarrrgghhh--credit for the idea while adding some justification and methodology of My own.