When you initially presented the study, you explained the finding as “entirely sensible and logical and to be expected” by anyone not making the “whopping” and supposedly fallacious assumption of other things equal, because..

in fact other things must be *hugely* unequal.

Every school system has a finite budget. Big or small, it doesn’t matter. So you can’t just reduce class size — you have to do it **at the cost of something else**.

That was your argument, not that, holding other resources equal via additional funding, reducing class sizes actually worsens outcomes.

Only now do you say that “these studies were not of budget cut results — they were tests of reducing class size, many providing full additional funding needed to do so while keeping other things equal.” First of all, that could mean a mix of observed situations, some in which smaller classes came at the expense of other resources, some not, and some various degrees of proportions, which is NOT the same as a study finding that other things equal (other resources constant via additional funding), smaller class sizes worsen outcomes (or even that they don’t improve outcomes). Second, and moreover, given what you argued (repeatedly), it’s unreasonable for you to think it unreasonable of me to think your point regarded only optimal allocation of resources within a given budget, i.e., with smaller classes requiring reductions in other resources, as opposed to the other things equal scenario of a larger budget providing for smaller classes without cutting other resources (or a preserved budget enabling avoidance of larger classes without cutting other resources).

And my point was simply (1) that you weren’t really saying that it’s some myth that larger classes hurt outcomes, other things equal, (2) you were only saying that avoiding larger classes hurts outcomes when it means greater cuts in other resources amid budget cuts, and (3) many who warn against the adverse effect of larger class sizes are advocating against a budget cut in the first place, not advocating against larger class sizes in the context of a given (reduced) budget.

If you can’t simply acknowledge that you were only referring to a “given budget” scenario and that it was misleadingly broad of you to use labels like “greatest urban legend” and “fallacy” without that qualifier (or at least acknowledging, after having it pointed out to you, that such labels don’t apply at all in many cases of people warning of the adverse effects of larger class sizes), I don’t know what to tell ya’.

Yeah, obviously with a given budget to purchase a variety of resources that are factors in any outcome, it’s silly for anyone to overlook the fact that purchasing more of one means less of another, and the negative effect of the latter should be considered along with the positive effect of the former. That’s the point you applied here, and you claim that many either fail to consider that negative effect, or erroneously (per the meta study) think class size generally trumps other factors in outcomes (what the study says about “other things equal” is unclear based on what you’ve said). I’m not disputing any of that. My point is that it just doesn’t apply to the point you are ridiculing as a myth, the “other things equal” point made by those advocating against budget cuts.

As Vivian pointed out and I agreed, there certainly can be a dilution of quality if more teachers are hired (or layoffs avoided, if we’re assuming inferior teachers are the ones laid off — a big assumption in the current union/system environment), and if we assume that directional dynamic, then we can’t assume “other things equal” even if non-teacher resources were unaffected (meaning budgets cut or remaining teachers paid more), but that is a separate argument vs. the one you presented, and it is still an open question how that nets out vs. the (presumably) positive effect of more teacher time per student.

Norquist has always been a paper tiger. How can you take seriously tough talk from someone who shares a name with a lovable muppet from Sesame Street?

Recent experience has taught me that there can be a lot of reasonable misunderstandings among bloggers and commenters. It is far less important that this misunderstandings occur, than how one reacts to them when they do. It sounds like we are in general agreement there.

You’re right about the discussion when you “stepped in”; however, I think Brooks changed the direction a bit when he raised his quite theoretical issue of *all things equal* what about class size? That caught me a bit off guard, but it also got me to think a bit more about the issue—that’s one of the advantages of these exchanges.

I gave an example previously that I think is still valid. When considering budget cuts, the point of departure is the general budget, and then sub-categories within that budget. Thus, a locality might be faced with deciding how to allocate funds among, say, roads and education. So, before one makes *that* choice, one might ask, I think, “what would be the consequences of the latter”? That leads to different questions about class size, teacher numbers, etc. So, it is not only a question of *if*, but *whether*.

Perhaps I’m not expressing the pont well. In any event, I think we’re closing in on a distinction that approaches semantic silliness, so I won’t belabor the point. Also, I’m a firm believer in the 80:20 rule and the law of diminishing returns. I think you, I and Brooks (and, who knows, perhaps Arne, too) are in greater than 95 percent agreement on this topic, so there is little point in expending a lot of effort to go that final mile, much less the last few meters. When you reach that point, one is better off breaking out the champagne to celebrate.

Here, David Leonhardt interprets the data as justifying $320K per year for outstanding kindergarten teachers (note the all things equal phrase two-thirds down the column):

http://www.nytimes.com/2010/07/28/business/economy/28leonhardt.html

But, here, Catherine Rampell, follows up with an alternative take by Hanuschek:

“A teacher one standard deviation above the mean effectiveness annually generates marginal gains of over $400,000 in present value of student future earnings with a class size of 20 and proportionately higher with larger class sizes. Alternatively, replacing the bottom 5-8 percent of teachers with average teachers could move the U.S. near the top of international math and science rankings with a present value of $100 trillion.”

Note that “proportionately higher with larger class sizes”.

http://economix.blogs.nytimes.com/2010/12/21/when-good-teaching-pays-off/

And here’s Richard Rothstein taking issue with Leonhardt for suggesting that education alone is going to fix the problem of inequality:

“Policies based on exaggerating school reform’s ability to ameliorate inequality leave most working families and their children unprotected. We need educational improvement, including better kindergarten, but also economic reforms — more job creation, greater protection of union organizing rights, higher minimum wages and more generous earned income tax credits — if we want disadvantaged children to have a fighting chance.”

And, David agrees!

“I agree with Mr. Rothstein that education, by itself, is not sufficient to guarantee good living standard for all workers. His recommended list of other policies has some important ideas, including greater protection of union organizing rights.”

http://economix.blogs.nytimes.com/2010/08/05/does-education-stop-at-kindergarten/

Yes, raising teacher salaries to $320K and providing greater protection of union organizing rights seems just the right combination of policy actions to improve our childrens’ educations and solve the long-term budget problem!

Jim, The “shock and awe” comment was intended to be a major compliment. I always find your comments well reasoned, well written and strongly supported by both logic and facts.

Aw, you make me blush. Thank you for kind words, they are so rare on the internet. Back at ya’.

I know you, and took the comment as friendly.

I was thinking of those poor Iraqi’s under a massive and precise barrage.

Yes, I do overdo it sometimes, pile it on. I just did again I fear.

I just spent the long weekend all alone in an empty large house in the country, no family, no friends, no neighbors, nobody at all to talk to … just blog comments for companionship … a sad case.

I’m pretty sure I’m compensating for something here, so forgive.

That said, I think your reply to Brooks’ query might be talking around the issue, particularly if you understood it completely the first (and second) time around. “If there is no budget cut, there’s nothing to talk about” is not entirely true. There would be plenty to talk about before arriving at that “if” point of departure. One first needs to decide *whether* there is to be a budget cut…

Maybe I’m dense, but I’m still missing the point.

When I stepped in people were talking (I thought) about the merits of saving teachers via stimulus funds, etc, from budget cuts. My point was to question any *assumption* that there is any special merit to using stimulus funds to protect teachers.

I assume that budget cuts are forced. Who cuts budgets but of necessity? *Given* the cuts, should stimulus be used to save teachers as a priority above other things?

OK, politics may evolve to avoid cuts altogether. But if there are no budget cuts, then teachers aren’t at risk and the budget isn’t being changed and class size isn’t being changed — nobody is doing anything, so what’s the point of talking about what to do?

Of course one can always talk about the merits of class size as a separate issue on its own — but I wouldn’t do that here. I’ve *done* at lot of that in the past on usenet in the misc.education newsgroup (never was one so appropriately named!) and in education forums. I would *not* bring the fury storms that erupt from that over here to Economistmom’s polite and friendly home.

Off topic here. And inadequate fire insurance. From the flame wars I went through in those discussions I still have scorched asbestos suits hanging in my closet.

So my thinking has been: consider class size as related to state budget cuts and using stimulus funds to save teachers, not in terms of general principles or anything else.

If this has caused me to miss a point, I apologize, I may be being dense but I’m not trying to be obstinate. If somebody wants to dummy-down to me what I’m missing, and it’s still worth discussing…

I don’t think there should be any argument that *all things being equal* smaller class sizes are better than larger ones. I’m surprised there would be much argument about that as a very general theoretical proposition.

…and…

In my view, though, all things are *never* equal.

That’s right, which makes the two statements contradictory - and the second trumps the first.

The first is “Assuming the impossible…”, which is a bad idea in any policy discussion. I’ll try a comparison:

*All things being equal* increasing the minimum wage by $100 per hour would make most everybody far, far better off by receiving a big increase in pay. But wait — even the most liberal of liberals know that would be a fiasco because other things can’t be equal. People priced out of jobs, unemployment, mass business bankruptcies, etc. There is no way to keep the other things equal and everybody knows it.

Yet, curiously, when the minimum wage increase is limited to say $3 an hour, suddenly masses of people believe other things can be equal. The associated costs disappear. If enacted on a big scale it has a big cost, so in enacted on a small scale it should have a small cost — but psychologically that somehow becomes no cost at all, so millions of people assume the increase is good.

Well, maybe it is or it isn’t. But given the known costs that come with it and all the other things that can’t be kept even, that surely is something not to be assumed but to be proven on the facts in each particular case.

ISTM that it’s the exact same thing with the class size argument, only even more so.

Harvard’s Ec 10 taught by Mankiw has 640 students. Say a public high school has an average class size of 25, using the best teachers it can get. Now they both decide to go to a 1-to-1 teacher student ratio, Harvard hires 639 more teachers for Ec 10, the high school goes out and hires 24x more teachers.

*All things equal* surely they both greatly increased the quality of the education they produce! No argument possible. But … wtf? How could other things possibly be equal??

The money cost would be … nuclear, the schools would be destroyed. But forget that — what would the quality of the 649th instructor hired by Harvard be compared to Mankiw? The high school would be recruiting teachers from the unemployment office and behind the Burger King counter. Put together, the money cost plus the quality-cost of hiring fries-sellers to teach, the result would be beyond a disaster. And everyone sees it.

Yet make the change in class size from 26 to 23 and somehow almost nobody sees the cost that comes out of resources and teacher quality — somehow, everything else now can be equal and near everyone assumes the result will be good. Go figure.

But it is just the same. There is zero reason to assume smaller class size helps rather than hurts. No ground for that assumption at all, and the data back that up.

Look at examples to the contrary. Harvard classes up of up to 640, in law school I had huge classes, NYC’s top selective public high school (Stuyvesant) has classes of 40+, the Gates Foundation just revoked its endorsement of small schools, etc.

With a moment’s thought it is at least as obvious that increasing class size can improve educational results. If students are competent and capable of learning on their own from capable classroom instruction, then larger class sizes that result in better teachers and far more resources per student can pay off hugely. While giving closer personal attention to these students who *don’t need it* is a hugely expensive waste. How many students like this are there in America? I say, a great many.

But without that moment’s thought, even for those students, 90% of people assume smaller class size would be better, when it would be worse. That’s how Urban Legends work.

BTW, all this is giving the benefit of the doubt to public schools — that they actually are managed competently and allocate resources efficiently. There’s ample reason to question that: Catholic schools that outperform public schools with 5 times the per-student budget, endless increases in public school expenditures with no matching rise in outcomes, the general historic record of how politically managed organizations operate so economically inefficiently. They give an entirely different grounds to reject any assumption that smaller call sizes improve outcomes.

I set up the easy case of eliminating 5 percent of the worst teachers as a trade-off for larger class sizes

Yes, that is entirely plausible and sensible. I’ve seen serious reform proposals much like that making exactly this point: the *quality* of teaching is very important — bad teachers can be *very* damaging — and simply eliminating the *worst* teachers significantly improves average quality of teaching for all. (E.g., Stevenson & Stigler saying “fewer, better teachers.”)

A Budget Director of the NYC Board of Education a few years ago created some furor in the ed world when he wrote a PhD thesis that ended by suggesting the NYC schools budget maybe should just take a good major whacking across-the-board. He documented how the ever-more money pouring in was so politically mal-spent that it only made things worse. He said maybe a bankruptcy-type major slashing of the budget would force pointless programs and the worst teachers to be cut, “rubber rooms” and all — literally improving the system by subtraction.

Alas, the assumption with both his idea and your “eliminate the 5% worst” is that the unions and politicians would allow the “worst” to be cut, instead of mandating via seniority that the juniors be cut while protecting all the senior worst who created the entire situation.

I cited a meta-study of hundreds of studies on reducing class size which reported that 20% found reducing class size worsened educational results, 67% found it made no difference, and a mere 13% found it improved outcomes.

Note well, these studies were not of budget cut results — they were tests of reducing class size, many providing full additional funding needed to do so while keeping other things equal.

The near universal assumption that smaller class size improves educational performance is the greatest Urban Legend in educational politics. The facts say so.

You were addressing those advocating against larger class sizes within the context of a given, reduced budget…

All budgets are given. I made it very clear that the analysis is marginal — repeatedly using the word “marginal” — and marginal analysis applies whether the budget a stable, being increased or being cut.

BTW, the marginal analysis I discussed is the explanation for the data results that is most sympathetic towards the public schools. Charitable, even. I use it when discussing the issue with public school teachers. There are other very plausible explanations of that data…

you’re not really disputing that larger class sizes yield inferior results, only that cutting other resources…

My citing the meta-study above doesn’t dispute that? My quoting Stevenson & Stigler proposing reform by increasing class size doesn’t dispute that? My statement that Harvard doesn’t hurt its reputation any by having classes with hundreds of students in them doesn’t dispute that?

Brooks, before declaring what I really wrote, read what I really wrote.

I am absolutely disputing each and every assumption that reducing class size will increase educational outcomes, and that increasing class size will hurt it.

An Urban Legend is something that on first thought so seems that it ought to be true that masses of people believe it actually is true — even though, on second thought both facts *and* reason clearly indicate it is not true.

This extremely widely-held assumption about the effect of class size on educational outcomes meets the definition of Urban Legend perfectly.

Again, I’m speaking of the argument — made by those advocating against…

I can only respond to what you actually wrote. That consisted of an example exchange which concluded, with my response…

Advocate: “But my point is that if we don’t make this budget cut, we won’t have to cut those other (non-teacher) resources or have larger class sizes. We won’t have to cut either. Those other resources stay the same as they are now — i.e., “other things equal.” And we won’t have larger class sizes, which I assert would harm outcomes, other things equal, regardless of whether that harm would be greater or less than the harm done by cutting only non-teacher resources to fit the smaller total budget. Even if increasing class sizes to fit a smaller budget were optimal within that new constraint (i.e., better than cutting only elsewhere), that outcome is still inferior to no-budget-cut (or lesser budget cut) scenario.”

Right, of course. If there’s no budget cut there’s nothing to talk about. If there is a budget cut then cutting teachers should be on the table on equal terms with cutting anything else.

What else do you want me to say? After “Advocate” gives that speech to the voters pleading not to cut the budget there are only two possibilities: (1) The voters decide not to cut the budget, or (2) The decide to do it. What else is there?

I’m sorry to disappoint you, but if you want me to give a different kind of answer then help me, give me a clue. “Jim, I imagined you might say something like ______ .”

Again, many (perhaps most) people warning against the adverse effects of larger class sizes aren’t advocating within the context of a given budget (i.e., implicitly advocating preserving class sizes at the greater expense of other resources), but rather advocating against the budget cut itself

Of course, and I never said otherwise. So what?

If they decide not to cut the budget because they don’t want to cut the budget then the whole issue of class size is moot.

they are advocating against the budget cut in the first place.

Yes. Point?

If Congress had implemented the Bowles-Simpson plan when it was first proposed then the tax code would already be simpler, fairer, and better suited to support economic growth. http://bit.ly/GVrWuY

A Bowles-Simpson type “grand bargain,” which leaves no sacred cows untouched, is the best way to get our economy back on track. http://bit.ly/noTDPF