President Obama loves merit pay. So does Arne Duncan. Editorial writers from coast to coast support the idea proposed by Gov. Arnold Schwarzenegger that “teacher employment be tied to performance, not to just showing up.” Dan Willingham wanders into the fray with his latest video, “Merit Pay, Teacher Pay and Value-Added Measures,” and offers six reasons why “value added measures sound fair, but they are not.”
The political winds certainly seem to be very much at the back of merit pay plans. Months or years hence, there may be a temptation to describe the “unintended consequences” of such plans. Call them unintended, but not unanticipated.
It’s difficult to discern the finer points of a complex argument from a 3.5-minute YouTube video. If the producer is saying that we need to think a little deeper than creating a heavily-weighted merit pay plan based strictly on value-added measures, he has a point.
If, however, he is arguing against performance-based compensation programs altogether, he is misguided. Policy reforms must be compared to the status quo, not to some idyllic standard of perfection. His six objections taken in sum can hardly be considered weighty next to the utterly flawed structure we currently use to pay most public school teachers. Put simply, don’t let the perfect be the enemy of the good.
If you want a better picture of the promise of pay for performance, check out this report, or at least the 5-page introduction:
http://www.hks.harvard.edu/pepg/PDF/Papers/Winters_et_al_PEPG08-15.pdf
It will take a little longer to read than it does to watch this YouTube video, but it will provide some very critical balance.
The video doesn’t focus on whether merit pay works. That’s an empirical question. The video focuses on whether they are fair, whether the right teachers end up getting the bonuses. I’m arguing that value-added measures are not reliable enough or valid enough to be used for this purpose.
Dan,
Ben clearly believes PFP “has promise.” (I’ll read the paper to which he points us before I make call on whether he’s supporting that view effectively.) But for the moment, he has a belief, a concept, at least, of where he’d like to go.
In principle. I presume many would support this belief or concept.
But, as you correctly point out, details matter. Not just principles. So if we do not get the details right, our well-intended beliefs can lead us astray. Inaccurate tools, however scientifically designed, can’t get us to more exact measurements.
The question with which I struggle is, “so what do we do next?” Doug Harris and Rick Hess seem at times to have written “get the framework in place, and we can sort the details later.” (if you build, it, they will come?) Folks like yourself, and Dan Koretz, point out specific metrics matter, and if they are not right, the system will not work.
But as a parent of new public school kids in NYC, where close to $7 billion of our $21 billion in annual school spending goes to teachers’ salaries and benefits, I can’t help but think that the effectiveness of this spending can be measured. Maybe not the way that the VAM proponents would have you believe.
But then how? There has to be a more effective answer than the “more research is needed” that seems to end Koretz’ book.
Over at “Bridging Differences” Diane Ravitch had the following thoughts on merit pay just last month:
“There are several reasons why it is a bad idea to pay teachers extra for raising student test scores:
* First, it will create an incentive for teachers to teach only what is on the tests of reading and math. This will narrow the curriculum to only the subjects tested.
* Second, it will encourage not only teaching to the test, but gaming the system (by such mechanisms as excluding low-performing students) and outright cheating.
* Third, it ignores a wealth of studies that show that student test scores are subject to statistical errors, measurement errors, and random errors, and that the “noise” in these scores is multiplied when used to make high-stakes personnel decisions.
* Fourth, it ignores the fact that most teachers in a school are not eligible for “merit” bonuses, only those who teach reading and math and only those for whom scores can be obtained in a previous year.
* It ignores the fact that many factors play a role in student test scores, including student ability, student motivation, family support (or lack thereof), the weather, distractions on testing day, etc.
* It ignores the fact that tests must be given at the beginning and the end of the year, not mid-year as is now the practice in many states. Otherwise, which teacher gets “credit,” and a bonus for score gains, the one who taught the student in the spring of the previous year or the one who taught her in the fall?”
Before I read her oppositional expose I was convinced merit pay was coming and that it would be a good thing for public education. Now, I see it has a number of hurdles to clear before it can be helpful.
Since we have no effective way of measuring success in schools, merit pay is nothing more than the same system Lehman Bros. used to hand out bonuses. We’ll create a system of cheating for short term gains which produce bonuses separated from any real accomplishment.
Now Dan Willingham and I have battled over stuff before, but everything he says here is right. So here’s my suggestion, if Arne Duncan really believes in merit pay we start letting students vote on their educational experiences once they reach age 30. The students at that point will know whether their education worked for them. And they will get to vote on bonuses for all in education at the time they were involved, from the Governors down through local and building administrators, and teachers. If things are really bad, we’ll claw back salaries, starting with that of the US Secretary of Education.
After all, accountability should begin at the top.
Dan,
Ben cites a paper by Marcus Winters, looking at a theoretical labor model that might account for teachers’ varying amount of effort in determining the tradeoffs they make between work and leisure. Then Winters looks at the merit pay system that was implemented in Little Rock and assesses that kids whose teachers were in the merit pay program had higher test scores than kids whose teachers did not.
The paper is in draft form, so presumably not yet peer reviewed, but it does seem to suffer from the flaw of equating outcomes on the Iowa test of Basic Skills general test as indicative of successful learning in the broader sense. There’s a slew of information on the ITBS website that suggests to me these tests were not designed to be used in “high stakes” testing. But that’s a battle for a different day
Accepting then that Ben’s evidence for his belief in merit pay may not be as strong as it could be, the question remains: Is the idea of value added metrics for evaluating teachers simply unworkable, and therefore as Ira suggests, it should be opposed on principle? Or can a workable alternative be crafted?
How about a merit pay system that takes into account multiple years of data.
For instance, assume:
Teacher A raise students test scores over a period of five years as:
24
28
18
27
26
24.6 average
Teacher B over five years has gains of:
17
19
22
27
19
20.8
Now all we have to do is design a system that both weights recent years (to take into account improvement) while at the same time including an average of say 5 years to take into account sustained performance.
Daniel is right in arguing that in any given year a host of variables could affect the scores, but I think most people would agree that over a 5 period of time, good teachers will more than make up for any bad luck years.
Of course this system means that teachers wouldn’t be eligible for merit pay until they hit a certain amount of time… but that shouldn’t be a problem since it allows new teachers to build up to proficiency (or excellence).
Another bonus would be adding an incentive for new teachers (especially good ones) to stay in education since they would anticipate a significant increase in pay when they hit the 5 year (or three, four, choose a number) mark.
Also, for the first five years of teaching there would be traditional pay steps which mirror the system in place now… ideally, the five year starting period would be tied with tenure.
My whole merit pay idea is at http://parentalcation.blogspot.com/2007/05/my-teacher-merit-pay-plan.html (posted back in 2007)
The Air Force has long had a promotion system that weights performance reports and over a period of 5 years.
Brilliant, no?
Hi Dan,
Good video. I have a question about your first point.
You say “Kids who move are more likely to be poor, and poor kids get lower test scores.”
But wouldn’t the question be whether poor kids, on average, have lower value-added gains?
Ie, if poor kids in a school routinely scored, say, 10 points lower, but had roughly the same VAM, then departing students would not be a problem.
Matthew: What do we do next? I would say that there first needs to be a discussion about what constitutes success for a teacher. Is it only teaching a lot of content and skills? How about a teacher who inspires, etc.? I’m really leery of using test scores as the only (or major) metric of evaluation for teachers because it clearly means that there is nothing else that teachers do of value. My own value system says that teaching kids skills and content is necessary to be a good teacher, but that teachers who do more should be valued and rewarded for doing more—inspiring kids, teaching children to love learning, teaching children to be good citizens.
Next, I would hope that teachers’ professional organizations would take ownership of this issue and put the time, money, and human capital (i.e., the measurement experts) into the problem to figure out how to reward good teachers and how to get incompetents out of the classroom. The fact that there is not a reasonable system to do this in place already is, I believe, why merit pay has gained such momentum. As I’ve written about before, if teachers don’t do this themselves, they are inviting others to do it for them. More research is doubtless needed. It’s a terrible problem because you know that you have multiple inputs which interact, and you have a number of outcomes that you care about, most or all of which are difficult to measure.
Ira: I’ve actually heard this idea suggested, although only for teachers—ask kids when they are adult to name the teachers they thought were most effective. The objections I’ve heard raised were: (1) there would be a bias to pick teachers from high school, because they are better remembered and because older kids can better appreciate what the teacher is doing; (2) teachers might still try to “game the system” by doing dramatic things to be memorable and later favorites; (3) the reward would be so distant that it wouldn’t affect the teachers much; (4) terrible logistical problems in keeping tracking of students and teachers to make the system work; (5) probability that kids would overweigh surface, personality qualities of the teacher and underweigh how much they learned from the teacher. I haven’t thought about this carefully, so I don’t know how serious these problems are. . .just throwing these thoughts out there.
Matthew Levey: I wasn’t familiar with the Winters paper and only looked at it briefly. As I noted in my previous post, it asks a different question than the one I asked. It claims that merit pay makes kids learn better taken as a group. Could you imagine other “treatments” that might get a group of teachers to improve teachers test scores? How about assigning bonuses randomly within the group? How about threatening all the teachers within a group that they will be fired if kids scores don’t go up? These might make scores go up, but that doesn’t mean that they “work.” To me, if a merit pay plan is not fair, it’s self-evident that it ought not to be used. Can a workable alternative be crafted? I sort of suspect that that test scores can be useful in some way at the level of the individual teacher, but that’s a mere intuition on my part.
Rory: taking five years at a time certainly helps with the reliability problem (how much depends on the correlation between fall and spring), but it doesn’t help with the other five because those sources of variance are probably systematic, not random. The helpfulness of taking five data points instead of one will be attenuated to the extent that you weight recent years, though, for obvious reasons.
GGW: Yes, if there is firm answer to exactly how the missing data problem affects scores, I don’t know it. (And if it were known and consistent, you could probably correct for it.) What we do know is that the missing data points aren’t random, and that’s always a problem. How big a problem, I don’t believe is known. (There are other troubling sources of non-randomness that I didn’t mention in the video; notably kids are not assigned to classrooms randomly.)
There’s a little more info on this stuff on my website on this page:
http://www.danielwillingham.com/readings-valueaddedmeasures
Why are we even having the discussion when the idea of merit pay is nothing more then another weapon to bash teachers with. Make enough public outrage and eventually force teachers to be nothing but minimum wage hourly employees. Use LOGIC here..
Was Doc Rivers a bad coach when the celtics finished dead last two years ago? Was he suddenly a great coach after the team acquired Garnett and Ray Allen? What was his value added? Look its simple teaching and coaching are similar in that we judge them based on results whether its passing tests or wins and losses. However both outcomes are determined by the talent the teacher or the coach has to work with.
Proponents of basing teacher “merit pay” on difference scores (honorifically termed “value added testing”) on instructionally insensitive standardized achievement tests are obligated to refute Willingham’s 6 objections and Ravitch’s related set of 6 objections. They haven’t done so and they can’t.
The Winters et. al. paper cited is based on 3 schools. The statistical gymnastics performed overwhelm the data. No descriptive statistics (Mean and SD) are provided for student test scores, but it’s evident that there is wide variability in the performance and that a goodly proportion of students at all grades are being “left behind.” There was also considerable variability within and among schools in payment received. The comparison group was nowhere near comparable to the 3 schools. The comparison differences are statistically significant, but of dubious practical significance.
This is not a study we can believe in as a basis for “merit pay.”
Jeff,
I like your Celtics analogy.
I’ve often thought the same of Phil Jackson. Where would the Bulls have finished without Michael Jordan or Scotty Pippen? Where would the Lakers have finished without Shaq or Kobey? What value did he add to either team? Based on his performance without the superstars I’d have to conclude his value added was not terribly significant.
Your comparing teaching and coaching looses a bit in the translation. Teachers have NEVER been judged under any quantitative rubric. It’s all been subjective “BS” by administrators, many of whom were paranoid of being sued if they ever put something negative in a teacher’s evaluation. I also believe all kids can learn. It might take some longer but they can all learn and teachers should be judged on proposed “growth models” tracking individual student progress from year to year.
Jeff: Merit pay is not necessarily, “just another idea to bash teachers with”–if teachers take charge of the conversation and lead in the developing of performance pay plans that are designed by us and for us, rather than another education reform plan that is done to us.
I agree with Dan that teacher organizations need to take ownership on this issue as they did in Denver; or as those of us in Teacher Leaders Network did when we wrote our own report on Performance Pay for Teachers (http://www.teachingquality.org/legacy/TSP4P2008.pdf). Rather than a bonus add-on, I would like to see a complete restructuring of teacher pay scales based on what highly effective teachers actually do (student learning, planning, assessment, collaboration, mentoring, etc.) instead of just on years and degrees.
GGW: According to William Sanders and others VAM experts, value-added gains cannot be reliably determined for mobile students, or for schools with highly mobile student populations–which tend to be the children and schools of the poor. It is not uncommon in high needs schools for students to move several times during a single school year. Wherever the child lands at test time gets held accountable for his/her test performance.
Rory: Your idea is interesting. Still, it relies on the quality of available test data. As many of the posts here have indicated (thanks Paul for sharing the one by Diane Ravitch)there are still many technical, pedagogical, and ethical issues surrounding the collection and use of VAM, even for those teachers whose students are tested.
Renee, you missed the point that Dan and I were discussing. We’re not interested in the VAM of departed students. We agree we want to toss them out. We’re interested in whether, when you do throw them out, it makes it impossible to measure the teacher effect on the “stable” kids.
Hoss and Jeff,
The NBA question you wonder has been answered! By economist D Berri. In his forthcoming book, he DOES find that some coaches have an econometric, measurable, controlled-for-the-roster effect on the game. And as much as it pains me to say it, Phil Jackson indeed leads the pack of the 19 coaches he evaluated.
http://dberri.wordpress.com/2008/11/18/coaches-and-deck-chairs/
Dan,
Randi and Sandra’s failure to define a positive (rather than reactive) program on this issue has, to me at least, always served as a reminder of where the UFT/AFT think their true interests lie. You can been a tireless advocate for your members’ interests, as Al Shanker famously was. Or you can undertake thoughtful and open-minded experiments to try to identify new policy options, as many in academia do. But I can think of few organizations that do both jobs, and do it well. Certainly not the UAW, the CWA or the Teamsters, to think of a few.
While it is easy to bash the union for doing the first job well, as a parent I have no real issue with them protecting their members’ interests. I can’t imagine why you would want to join a union if they did not do that well.
On the research side, equally disappointing is the eagerness of policy makers to embrace whichever of the latest “solutions” is put forth by the most recent eager-to-make tenure post doc with a year’s worth of data. While I really bear no ill will towards Winters, I think as consumers we need to be aware of his predisposition and his data set before we interpret his findings.
You may be aware of the kerfuffle about whether Roland Fryer’s draft paper on Harlem Children’s Zone “proves” that “quality schools” alone – and there’s no merit bay at HCZ, so far as I know – can close the achievement gap. If you go way back (to 2004), Fryer and Steve Levitt wrote a similar paper based on the ECLS-K data that the US DOE organized. Contradicting claims they made in 2003, the authors said they could not determine whether school quality explained the difference in student outcomes. And then to confuse matters further Eric Hanuchek wrote a paper for NBER that same year, using the same ECLS-K data, saying that Fryer had made a math error, and that in fact quality schools do matter. Makes you feel like a parent separating your three kids from some silly fight, with each one of them insisting that it was their sister or brother who did “it” and no real chance of arriving at the truth.
It is such a shame that with all their money the AFT couldn’t fund some really good high-quality long term research with a diverse panel of investigators. Put Kruger and Hess and Hanushek and you and Dan K. in a room, lock the door and say “no one comes out ’til you have a collaborative plan for some meaningful metrics and end points.” You could probably even get Fordham and Broad and Gates to contribute, and make it really non-partisan, or post-partisan
Alas I’m probably too idealistic.
Various plans for implementing merit pay for teachers have got to be:(1)easy to understand,(2)fair,(3) consistent over time and (4) independently funded. By #1 I mean that developing complicated statistical data models, etc. where the recipient is unsure what is going on and cannot verify the results independently, will just cause people to be cynical. By #2 I mean it has to apply to everyone, not just core subjects like math. By #3 I mean that the criteria doesn’t change from year to year. By #4 I mean that the pay has to come from an independent ongoing sourse so an X dollar amount of merit pay for the same quality of work will not be reduced the following year due to “lack of funds”. These are enormous problems to solve.
Thanks for an interesting video once again.
Some reactions and questions:
Missing data. But how does this lead to overall bias? Do you think that within in a given school, 10 poor kids will leave all from the same classroom, while no poor kids will leave from any other classroom? Do you think that poor kids will all leave one particular school while no other poor kids will transfer into that school? What about the possibility that the overall effect would be a wash, if poor kids move around in all sorts of directions and then it all balances out, on average?
Scores correlated means that difference scores are unreliable. Could you point me to a more detailed explanation of this point?
Same score gains might not be equivalent. There should be a way to collect data on how much students at various points in the distribution tend to improve from year to year. Thus, if students in the bottom decile average an improvement of 20 scale score points next year, while students in the top decile tend to improve by 30 scale score points, then you would know that it’s easier to get scale score improvements in the top decile. And then the cut scores for any rewards could be set accordingly.
Other people affect what teachers do. This problem would be ameliorated, wouldn’t it, if a merit pay plan rewarded everyone on a school-wide basis for school-wide improvements? Let’s not limit the concept of “merit pay” as only consisting of rewards given to reward individual teachers standing alone. There are other ways to structure merit pay.
Should teachers worry about short-term gains? I’m not as sure what this means. “Broad knowledge of the world” may in fact be quite important to one’s success on reading comprehension tests, as Hirsch often points out.
Effects of peers in different classrooms. See above about a school-wide plan.
GGW,
I think we understand your point, but I think you are overlooking a major point. The burden of proof is on advocates of merit pay. You argue that it may not be more difficult to raise the value-added of lower performing students than higher performing. IN THEORY you may be right, but value-added models fall apart if your theory doesn’t hold.
But, if I understand Dan correctly, researchers would probably find patterns of evidence for your theory if evidence was available.
On the other hand, you are ignoring the preponderance of evidence from practical experience. I’d think that the overwhelming majority of neighborhood school teachers would agree that it is much more difficult to add value to lower performing test scores. The basic reasons should be obvious. With so many troubled kids, you can’t address the academic deficiencies until you address the emotional needs of those students.
I’d also suspect that the “Matthew Effect” is crucial. Once your learn to read, and then read to learn, and once you have the background knowledge, and all that feeds into more motivation, students with those advantages can “take off.”
This is a distinction that should be obvious, and most likely explains a big difference between charter schools, even those that try to keep creaming to a minimum, and neighborhood schools.
I suppose the key question to be answered is what is more fair… an imperfect value added system or a system that pays all performers on the same scale, even though some might be far superior to others.
Additionally, I read two of the studies that Dan had linked to from his website, and while they did acknowledge that there were varyiations in consistancy in performance from year to year, it was consistant enought to compare to other private sector rating sytems.
Additionally, in one of the studies the author did mention using multiple years to get performance data.
Stuart Buck The real problem is non-random assignment. Value added models implicitly assume that students are randomly assigned to teachers. They aren’t, obviously. A principal might believe that Teacher A is especially skilled in handling discipline problems, so she gets more than her share of disruptive kids. The differential attrition is the same problem in a different guise. It’s not obvious how it would affect the data==if it affected the data in a consistent way, you could probably account for it in the model. Sorry, I don’t know of a good explanation of the reliability problems. . . the best way to think about it (I think) is that if A and B are highly correlated, changes in ranks from A to B will mostly be due to random error in the tests. (The reliability problem isn’t that bad unless A and B are pretty highly correlated. If you had tracked ability classes, for example, the problem would probably be attenuated.) Norming merit pay advances by average gains doesn’t solve the scaling problem–scale intervals could still be different–but it’s a nice end-run. Off the top of my head, I think that would work. I agree that school-wide merit pay would solve some problems. It wouldn’t solve the principal leadership problem though. (I’ve heard this suggested as a way of solving another problem–that merit pay would encourage teachers *not* to collaborate. I’m not sure what other effects it would have. . .it might make teachers bitter that lousy teahcers in the school freeload on the good ones and still get merit bonsues. . . or the peer pressure might encourage the poor teachers to really buckle down and try harder. ) “Short term gains” is one part of the you-can’t-test-everything-that’s-important-but-teachers-will-end-teaching-only-what’s-tested argument. I was suggesting exactly what you raised–that broad knowledge of the world is crucial to teach in K-3, is difficult to test, and is vital to reading comprehension after grade 4.
Rory, if the system is lousy, I suspect it won’t work long because teachers will come to perceive it as unfair. If the main way merit pay works is by motivating teachers, how long will they remain motivated under a system that they think is arbitrary?
We all know how glacially slowly policy changes come about, and we all know how political opponents of a policy attack when a policy doesn’t work perfectly==there is no rescuing the good parts of a plan, the whole idea is just labeled BAD. If merit pay doesn’t work well the first time, there probably will not be a luxury of tinkering with it. So why not try a little harder to come up with a system that, going in, we *think* is pretty darn good? There are doubtless going to be angles on this we hadn’t thought of, so if our starting point is “flawed, but better than nothing” I predict we’ll be looking at an experiment that lasts a few years and peters out.
Thanks for taking the time to write such long and thoughtful replies. Much appreciated.
A few minor points:
I was suggesting exactly what you raised–that broad knowledge of the world is crucial to teach in K-3, is difficult to test, and is vital to reading comprehension after grade 4.
In my state (Arkansas), K through 2 aren’t tested. Just 3 through 8. So maybe it just needs to be made clear to teachers of earlier grades that classes should include lots of background knowledge, not just decoding.
For school-wide merit pay, I suspect that peer pressure would overcome the impulse to freeload, but that could obviously depend on the particulars of a given school.
John Thompson,
Your thoughts on the value-added by Phil Jackson? Would he have won all those championships in Chicago or LA without his cadre of superstars? I know you’re a big bball fan and participant.
Hey John,
I agree most folks would agree with you that the burden of proof should fall on those who want to use merit pay. Usually, once you have a default position, the burden of proof is on those who want change.
Just musing aloud, but I wonder if that the best thing.
Two thoughts:
1. Shouldn’t the burden of proof be on what compensation scheme is most likely to lead to gains in student achievement?
Ie, let’s say there were evidence that flawed VAM (for the reasons Dan describes) leads to kids learning more. I’m not saying that evidence exists. If it did, wouldn’t the burden of proof be on those defending the status quo?
2. Why do you and others hold merit pay to such a higher standard than other places where judgment is required?
You cannot perfectly measure the performance of your students. Do you throw up your hands and give them all the same grade? Of course not. You choose to give imperfect grades because it’s better than essentially giving no grades.
What about hiring? The decision to hire someone, whether you’ll give them a couple million dollars, is super subjective. Yet you do not advocate we should simply give everyone an equal chance (a random lottery for who gets teaching jobs).
So we’re cool with blocking people from a $2 million lifetime job based on an interview and maybe a sample lesson….
….but we’re against paying one person $2.1 million based on much better — though admittedly wildly imperfect — data? Does that make sense?
Paul,
Im agreeing with you on Phil Jackson.
GGW,
The issue is sustainable improvements. I have no doubt that D.C.’s heaviliy funded schemes could have produced more good than harm for the 40,000 students in that district, just like the breaking of unions and subsidizing the flight of jobs from our industrial belt in the North to the South, provided more short-term good to the Sun Belt. But it unleashed the forces of greed, and impoverished us all. And that rapid deindustrialization created much of our educational problems.
George Soros explained the financial collapse of last year was like a number of bottles filled with water, while only one is poisoned. But that makes all of the water worthless.
Who would invest their career in inner city schools if a flawed VAM put it at risk every few years? Would you become an inner city teacher if flaws in the model put your career at risk one out of ten years, or one out of four years, or even one out of 20 years? Let’s say the model leads to an injustice 5% of the time. For every victimized teacher, how many fellow teachers would be demoralized? For every teacher whose career was destroyed unfairly, how many would have to put up with how much more stress with the Sword of Damacles over them. At what point would self-respecting teachers say that thet won’t take it anymore and flee the profession in mass?
But I’d support TAP’s performance pay. The VAM would be for bonuses so imperfections would balance out. Plus most pay increases would be for teams not individuals.
The best thing about performance pay is that it could encourage collaboration and open discussions.
So, the burden of proof is balanced with incentive pay, but when it involves the potential death penalty for individuals’ careers, then the burden of proof must always remain on the VAM.
In fact, our legal system is fundamentally based on that burden. So, trade-offs for a flawed VAM must included our respect for contracts, due process, and the rule of law.
Which gets us back to self-respect. Who wants to be a member of a profession that is denied the fundamental rights of citizens? Any profession that would trade their rights for a boost in test scores, to paraphrase, is worthy of neither, and will lose both.
This is like paying the car salesman based on how well the customer drives the car. Eventually we will come to realize that there is NO bureaucratic way to ensure that teaching and learning takes place. The government needs to get out of the education business all together.
In a democracy, ideas are to flow from the people to the government, not the other way around.