Just Because Kids Have Bad Parents or Bad Luck Doesn't Mean They Should Be Stuck With Bad Teachers Too
As the debate over releasing teacher performance data continues, many in the anti-data dump camp (read: unions and friends) have settled on an argument that can quickly and somewhat unfairly be characterized as: "It's not our fault!"
Here's the longer version: Value-added data, which compares students' scores at the beginning of the year and the end of the year for each teacher, isn't a fair measure of teacher performance, because student home life, economic resources, language skills, and other factors outside of the classroom have a bigger impact on achievement than anything teachers can do.
In today's New York Daily News, former reformer Diane Ravitch repeats that claim, writing:
Certainly teachers are very important - the most important cause of student success within the school. But scholars agree that factors outside the school are even more important than the teacher, especially family income. Most economists estimate that teachers are responsible for about 10% to 20% of the variation in student scores, and that outside-the-school factors influence about 60% of student learning gains.
The first sentence explains why the argument in the rest of the paragraph is irrelevant to the debate. Teachers may not be the be-all-and-end-all of every student's life, but they are an important variable that should be easily tweakable. Changing the socio-economic status, immigration history, or disability status of most students isn't feasible. There's not much a middle school principal can do about the education attainment of a student's parents or the amount of time they have to devote to supervising homework either. But middle school principals do (or could) have the power to hire and fire teachers.
Ravitch notes that the data New York wants to release would be based on a single year's gains, unlike the Los Angeles Times data dump, which spanned 7 years. Fair enough, any teacher can get a crap batch of kids for one year. But consistent failure to add value across many years and many groups of kids—especially within schools where other teachers are producing bigger gains with kids who are demographically identical—is a sign that something is wrong with the teacher. A brilliant teacher can't fix everything, but that doesn't mean bad teachers shouldn't be exposed and fired.
Editor's Note: As of February 29, 2024, commenting privileges on reason.com posts are limited to Reason Plus subscribers. Past commenters are grandfathered in for a temporary period. Subscribe here to preserve your ability to comment. Your Reason Plus subscription also gives you an ad-free version of reason.com, along with full access to the digital edition and archives of Reason magazine. We request that comments be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of reason.com or Reason Foundation. We reserve the right to delete any comment and ban commenters for any reason at any time. Comments may only be edited within 5 minutes of posting. Report abuses.
Please
to post comments
That's an awesome empty chair argument.
"You can't judge just us because everything else in the world contributes also! But you'll just have to take our word for it that we're super important! Oh, and all of us are good!"
I too thought it was a bold move to argue teachers aren't really that important to a child's education.
Here's the longer version: Value-added data, which compares students' scores at the beginning of the year and the end of the year for each teacher, isn't a fair measure of teacher performance, because student home life, economic resources, language skills, and other factors outside of the classroom have a bigger impact on achievement than anything teachers can do.
Fair enough. They you're all overpaid.
Agree.
"Most economists estimate that teachers are responsible for about 10% to 20% of the variation in student scores, and that outside-the-school factors influence about 60% of student learning gains."
Wow, I actually agree with that. So let's not go on about how teachers are underpaid, and how we need a teacher for every 1.25 students.
If you WANT to learn, or if your parents MAKE you learn, than you will probably have some academic success.
Nothing like failure to finally admit the truth.
Even more so, it is possible that teachers matter so little precisely because we DON'T concentrate on their value add.
There's an old saying: "If you want it to get better, measure it."
I guarantee you that there are teachers out there that have more than a 10 - 20% impact on a student's improvement. And there is no better way to find those teachers than to measure the shit out of them.
This isn't rocket science. If there were more transparency in these schools, we could easily control for socio-economic conditions and compare teachers of similar students to one another. (Teachers Unions do this all the fucking time when they try to prove that public schools are on par with private schools.)
So, I believe you teachers. And I want to help you get better and more impact on your kids. Let us measure you, and then clone the practices of those teachers that we find are doing well.
Let us measure you, and then clone the practices of those teachers that we find are doing well.
Absolutely. And let us fire the ones that underperform year after year.
Workin' on it!
If teachers can't be held responsible, they are vastly overpaid. If they can be held responsible, the gains, or lack thereof, in students in their classes seem like a reasonable, if not fail proof, way of doing so.
That poor kid in the cartoon got taunted into suicide for being gay, didn't he?
Thats not funny/Yes it is
No, it's not/Yes it is
OK, maybe a little.
Angelogical/demonological jocularity-quotient casuistry-dilemma skirted (you should pardon the expression).
But consistent failure to add value across many years and many groups of kids ... is a sign that something is wrong with the teacher.
It seems there should be a lesson here concerning the economy, but I can't quite put my finger on it.
Most teachers don't want to be honestly evaluated for the same reason that most students don't want to be honestly graded.
Not that many are going to all As and Bs.
Most teachers don't want to be evaluated for the same reason most of their students don't want to be graded.
Not a whole lot of either group are going to get all As and Bs.
Is there nothing that a powerful union won't enable?
Progress?
Before any discussion of teacher performance, I need a showing that people involved are actually teachers. I have seen way too many illiterates and innumerates waving transcripts full of A's and B's to mistake people who had issued those grades for teachers.
Look, Ravitch has been the poster person for conservative critiques of education for a while now. When even she has bailed on the sense behind this indict the teacher stuff then that ship is sinking.
"The first sentence explains why the argument in the rest of the paragraph is irrelevant to the debate."
That's just goofy. As she says while teachers may be the most important single factor that factor is responsible for only 10-20% of variation. So Super-teacher could be the teacher but if all the other factors are breaking against him that teacher's scores would look relatively worse than worse teachers with more favorable conditions in the other areas. In other words the problem with the value-added measure is that it is alone a bad measure of whether the teacher is good or bad.
Right, minge. Then we should abolish schools and let people fend for their own children's education.
Good point if that's where you were heading with it.
But it wouldn't explain why that teacher consistently got the same results as other teachers at the same school, and KMW pointed out.
Sure it could. If the teacher, for whatever reason, got virtually unteachable kids every year while his peers got better kids, then equal scores on value added would be what you would get, even though teacher A was doing a better job than his peers...A great teacher might add, say, 5 points per kid with unteachables where an average teacher could get the same value added with better kids. The scores would simply say they were equal. But without knowing what they had to work with we such a conclusion is goofy.
No, unless you believe that teacher A was not just getting difficult to teach kids, but difficult to teach kids that lucked into getting equally great teachers at each grade before and after teacher A had them.
If the kids require a great teacher to get 5 points added per kid in 5th grade, then presumably unless they had great teachers in 1st through 4th, they'd average learning less in their prior grades, and they'd average learning less after they went on to different teachers.
The whole "value added" metric used in LA looks at what those students accomplished before reaching teacher A, and accomplished after getting other teachers. Since the LA data used seven years of analysis, an enormous portion of these concerns are addressed.
You're acting as though these "unteachables" appear out of nowhere, and aren't taught by other teachers in previous and future years.
It's a fine complaint for why one-year data isn't as good-- which is why LA addressed that complaint.
Right- as I said above- the key is for schools to release the data. We can easily control for a lot of this data- especially since many public schools have multiple classes at each grade level, with kids moving in and out of the same classes.
Also, shifting metrics around to understanding whether the kids' performance is increasing can have several beneficial side effects on behavior. Today, unteachable children continue to shuffle through school. But teachers would have a huge incentive to break through with such kids, as we would see a spike in the value add. Today, the incentive is to get those kids shuffled through- because it costs too much to kick them out, and is too much work to intervene.
There wouldn't be "whatever reason" for that - it would mean the administration was deliberately trying to fuck over a teacher.
Probably because they didn't join the NEA.
Or because they didn't have seniority.
You misunderstand how the data was analyzed.
Rather than think of teaher A, think of students A through Z and Teacher Smith. The students were tracked individually. If, during the previous four years these individual students showed no, or minimal improvement, and during their year with Teacher Smith they showed a 5% improvement, then Teacher Smith did a good job.
But if, on the other hand, the individual students showed a 15% improvement in the previous four years and a 5% improvement with Teacher Smith, then Smith did a poor job.
The data tracked student performance over time versus student performance with each teacher. "Unteachable" students for four years would most likely be unteachable for the fifth year and Teacher Smith would not be rated poorly as a result.
There are certain to be some students that are dealing with home issues (divorcing parents, illness etc.) and individual performance over time would vary as a result, but a teacher that showed consistent low performance over a span of years would not be too difficult to identify.
Value-added is just one tool to evaluate teachers, but it is a valuable one. I find it more interesting that when test results go up, teachers and their unions crow that this shows how dedicated teachers are making a difference, but when scores go down, it is the fault of everyone, parents, society, video games and not the fault of the classroom teachers.
You're having trouble with the "value-added" part, aren't you? The whole concept of value-added is to attempt to measure and correct for those "other factors" instead of simply measuring the actual achievement which, yes, is affected by those other factors.
I'll give you the benefit of the doubt and assume that you're simply arguing that value-added measures are impossible.
Are'nt value added measures like the one involved simply pre-post testing? I understand the logic that this is supposed to control for these factors because you're just looking at the pre-post test for each class, and while, yes the overall score for class A might be lower than that for class B we don't compare that, we compare the improvement in A and B. But doesn't that assume that each set is equally "teachable" (likely to make the same amount of improvement though starting and finishing at different places)? I just don't buy that assumption...
Look, no, it doesn't work if you compare two classes. But LAUSD has 700,000 students and 45,000 teachers. Let's say those students are taking around 16 courses a year, and you have 11.2 million courses taken, each with its own set of data points. Using that information you can adjust for factors like a school's resources, socioeconomic changes.
If a kid downtown improves by 5% when someone with the same background uptown improves by 10%, you might say the kid uptown has a better teacher. But maybe the school downtown lacks the materials in his curriculum. Maybe the school had a tragic accident and kids are adjusting. Maybe the teacher had to take leave for some time and the substitute was inadequate. You can adjust for those. Or, maybe the kid had home problems which are beyond the knowledge of the school. Well, there's nothing you can do about that, but over time such factors will balance out.
We don't even have to find the perfect formula; the point is more that certain aspects of a teacher's job can be quantified, and to that extent we should include objective data when evaluating their performances. A good example would be how we look at athletes or companies (who seem to have no problem with statistics being kept). There are many ways of rating players, like PERS, or teams, like RPI (or W/L). Investments are judged based on P/E, bond ratings, ROE, etc. These numbers are not the end of evaluation, but they do provide a very useful context. I can't tell you whether JPMorgan (P/E 10.3) is cheaper than Morgan Stanley (P/E 11.4), and I can't tell you whether Michael Jordan (the highest career PERS rating, 27.9) was the best ever, but numbers do provide a basis for comparison.
If anything, given that we have a large number of teachers teaching to the same material, it makes quantitative analysis that much easier.
Awesome. So we can eliminate all public elementary and secondary schools, save ourselves over $600B a year, and suffer, at most, a 20% decrease in student achievement. Many/most parents would pay for private school (especially with their reduced property and state income taxes), so we probably would see considerably less than that.
especially within schools where other teachers are producing bigger gains with kids who are demographically identical
One problem with this, however, is that even intra-school student populations can be greatly segregated to such a degree that the classroom make-up can vary greatly from teacher to teacher. Consequently, teaching at the same school does not necessarily mean that the teachers within that school are dealing with the same groups of students. Further, it is not at all uncommon for teachers to teach similar courses over a number of years such that a "crap bunch of kids for one year" may occur every year.
I know many teachers who really support the idea of merit based incentives, etc., but rarely if ever are apples being compared to apples where they would feel that their performance is being fairly measured against their peers.
But wouldn't that "crap bunch of kids" scores at the beginning of the year vs the end of the year still indicate the progress they made?
Apples to apples means progress for each group based on their group's starting point, not an overall starting point for every kid in that grade. Therefore, I still think rating teachers this way is a fairly effective method.
Which is why the value-added data discussed in LA addressed this issue, by looking at the specific group of incoming students for specific teachers, not for the school as a whole.
All those concerns are reasonable. But teachers unions tend to ignore any efforts to address those concerns, sparking suspicion that they really don't want a good metric that address them, that they'd rather just throw their hands up and say it's impossible.
Ya know. In the private sector, this discussion wouldn't even take place. Goals and criteria would be set every year, and if you couldn't meet the expectations you would be out the door. End of story. They're basically saying that they're not smart enough to do what gets done in the private sector every day. The critics of these attempts to measure teachers performance seem to me to be making a really good argument for school vouchers.
I was thinking (somewhat) along these lines. I understand quality teaching is a very hard thing to measure. But in the private sector there are hard things to measure that, well, get measured. I guess you could argue the measures in the private sector are always based on what people are willing to buy and perhaps what teaching is supposed to do is poorly measured by such a metric (for example, very few master-pieces in literature have been the highest selling books for their time).
I was thinking, what private sector job is close to what teaching is supposed to do. Maybe coaching? If you get the Clippers you can't be expected to do what Jackson does for the Lakers, but if you are a good coach then you should be able to improve the record of the team...
I dunno...
Plenty of private sector jobs are pretty close. The difference is in the organization.
In the private sector, people lose jobs because of factors out of their hands (whether luck or bad decisions above them), but get second chances.
The advantage of the free market doesn't lie just in competition, it lies in the ability to shutdown failures and start over, something that the government is terrible at.
If you try to completely eliminate Type II errors and never fire someone who did good work but had bad luck, you'll have to live with a lot of crappy employees. The free market way is to give people second chances.
alt text: You niggers stay in school!!!
Clearly all they need is more money...
I thought they had agreed to give all kids in California schools B's this year to boost their self-esteem.
Most economists estimate...
Isn't this generally a signal that what follows is not to be confused with reality?
For whatever bearing this may have on the topic - possibly none - school counselors sometimes shoehorn "troubled" kids into a particular teacher's class because he has good rapport, good class room management skills, coaches a sport in which some of the kids are involved, etc.
I think principals will be firing people every two-three years as they fail to raise the educational standards of useless black/mexican kids in the ghettos that lack any fundamentals and manners.
Look, no one is going to fix the social economic situations. Tt's very un-libertarian-like and conservatives will soon take over the government again to spend the taxpayers money on more noble causes than education like Iran.
We should go with an Idea I had years ago: consider the sterilization of high-risk welfare groups. Soon they'll die off and we won't have to worry about this.
Brilliant! Eugenics is progress for the working class. Now, all you filthy non-whites line the hell up and spread your legs. This might sting a shitload..
Pretty good post. I just stumbled upon your blog and wanted to say that I have really enjoyed reading your blog posts.
Certainly teachers are very important - the most important cause of student success within the school. But scholars agree that factors outside the school are even more important than the teacher, especially family income.
IOW, the concept of "failing schools" is bogus. Only "scholars", in the Orwellian sense, could be surprised at that.
"Family income" is just a proxy for IQ. PC "researchers" aren't allowed to use that term even though it's a much better predictor than income.
The best part of measuring kids performance is that it reinforces the idea that the prime goal of public education is to educate kids. It isn't $/pupil or class size or one of any other things that I hear about so much from my kids' schools.
If you pay teachers for performance, I think you will get a lot of good things happening in the school. For one, teachers won't put up with much nonsense from the kids. They will realize that their paycheck is in the hands of those kids and they will push them to meet their goals.
Smart teachers would also realize that if parents are an important factor in learning that they should make sure to call them in for meetings and get them to cooperate too. In fact, if the bonus was big enough, they might think of bribing the parents. I would. "If Jonny improves by 5% I get an extra $1000. I'll split it with you guys."
I'm trying to think of the headlines if grades increase dramatically at a public school:
"Teacher union declines to take credit for grade increase - 'this was simply a matter of parents getting more involved with their kids' education - we had nothing to do with it,' says the President of the local NEA."
"Local Teacher of the Year returns award for helping increase student grades - 'I didn't do a thing for these students; they and their families did it all by themselves,' said the teacher in returning the award."
"Teachers' union declines salary increase for its members - 'since teachers don't have much effect on what students learn, and since it's the home environment which is more important, we urge that this money not go to us, but instead be used to fund local social services and improve these kids' home lives,' said the head of the teachers' union local."
Tt's very un-libertarian-like and conservatives will soon take over the government again to spend the taxpayers money on more noble causes than education like Iran.
"Local Teacher of the Year returns award for helping increase student grades - 'I didn't do a thing for these students; they and their families did it all by themselves,' said the teacher in returning the award."
plus
ray ban aren't different. You could do this, by removing the contact lenses you possibly can consider considerably off the out the door cost. contemplate purchasing a couple of
ray ban sunglasses sale.