Eduwonkette offered me the opportunity to respond to Skoolboy’s postings on class size –which I originally rejected since I wanted to at least try to spend a few days over the holidays in relative peace with my family, and not have to argue about this topic.
But as you can see, I changed my mind, since it proved difficult for me to resist responding to his observations on a subject so close to my heart.
1. Let’s start with his account of the STAR studies on Tennessee – which have been rightfully described as one of the best – and one of the few convincing large-scale experiments in the history of education research. The STAR studies demonstrated conclusively that class size does make a difference – though even earlier, many analyses showed a convincing correlation between smaller classes and better academic outcomes. Skoolboy writes:
“… since STAR involved a contrast between a class of 22 to 25 and a class of 13 to 17, you might not want to speculate about the consequences of reducing class size from 28 to 21, since that wasn’t actually observed in the STAR study.”
Actually, Alan Krueger of Princeton analyzed the results for the control group in STAR of students who were in classes from 22-25, and found that within this range, the smaller the class, the better the outcomes as well. Indeed, class size researchers believe that the benefits from class size reduction are roughly linear, and no threshold needs to be crossed in terms of class size to improve learning opportunities for students.
2. Now let’s consider Skoolboy’s comments on California, and clarify what the research actually showed about the results of that state’s efforts to reduce class size, which began in 1996-7:
“Studies of the initiative show that California districts scrambled to reduce class sizes by hiring teachers with intern or emergency credentials, and many of these teachers wound up teaching in large urban schools serving poor, minority students who were English language learners. (In contrast, all of the teachers in STAR were fully certified.) Moreover, districts had to cannibalize space for small classes that otherwise would have been used for other purposes, including special education, arts and music, and athletics. Perhaps as a consequence of how class size reduction was implemented in California, researchers were unable to conclude that it had positive effects on student achievement.”
Actually, every controlled study of the California class size reduction program – and there have been at least eight so far- showed significant gains from smaller classes. (For some of these citations, see this CSM fact sheet).
One analysis, by the Public Policy Institute of California, showed that in the five largest school districts other than Los Angeles, namely San Diego, San Francisco, Long Beach, Oakland and Fresno, class size reduction raised the proportion of third graders who exceeded the national median by l0.5 % in math, and 8.4 % in reading, after controlling for all other factors. And though this report didn’t find significant gains in LA, others that looked at student level data – generally considered more reliable -- that did find positive results.
Overall, the research on CA is more ambiguous than the STAR studies, if only because the state undertook several major reforms at about the same time, including adopting a new math curriculum, eliminating bilingual education, and implementing new accountability measures that make it difficult to isolate the effects of class size.
To further complicate the question, the state introduced a new test which made it impossible to compare scores before and after classes were reduced. Most schools also lowered class size so quickly it was difficult to find a control group to which results might be compared. Yet even the most equivocal study — that from the RAND/AIR consortium — showed a significant gains in test scores for the one control group they could identify: third graders who remained in larger classes as compared to those in smaller classes, though many of the latter group had been in a small class for only one year or less.
And although the researchers said these gains were “small”, the effect size they found was almost exactly as would be expected from the STAR studies, which showed negligible gains from third graders who spent only one year in small classes, with the largest gains, as might be expected, for students who were in smaller classes in Kindergarten and remained in a small class for several years thereafter.
In addition, the RAND/AIR Consortium made several other methodological errors. For example, in order to eliminate the “unobservable” differences between schools that reduced class size and those that didn’t, the researchers subtracted the difference between fifth grade student test scores at both sets of schools, assuming that fifth graders would have been unaffected by the program. Yet as Fatih Unlu demonstrated in his paper on CSR in California, 18% of the fifth graders in those schools with smaller classes in the third grade had actually been in smaller classes themselves two years before.
As Unlu points out, if the CSR program had positive effects, adjusting the test score differences of the third grade CSR participants and non-participants by subtracting the test scores of these higher-achieving 5th graders unfairly lessened the estimated effect of class size reduction.
In his study of the impact of smaller classes in California, Unlu avoided some of the limitations encountered by other researchers who were stymied by the fact that the state tests were new. He instead analyzed pre-existing NAEP scores for California 4th graders in 1996 and compared them to the results in 2000, since the earlier group had no experience of smaller classes, while the latter group had experienced smaller classes for several years before entering the fourth grade.
The NAEPs are also considered more reliable than state standardized tests, since no high stakes are attached. By using two different statistical methods, Unlu finds very substantial gains from smaller classes -- .30 of a standard deviation.
Though Skoolboy and others emphasize that the number of uncertified teachers rose as a result of CSR, none of the studies – including the analysis done by the RAND/AIR group, could find any evidence that these teachers were any less effective. In fact, more recently, Robert Gordon et al. write that “When the Los Angeles Unified School District needed to triple its hiring of elementary teachers following the state’s class-size reduction initiative in 1997, the district was able to do so without experiencing a reduction in mean teacher effectiveness, even though a disproportionate share of the new recruits were not certified.”
And while many anecdotal reports to this day claim that certified and experienced teachers fled high needs schools to go teach elsewhere when class size was reduced, there is no evidence for this; in fact follow-up studies from California showed that after rising temporarily, teacher migration rates fell dramatically in all types of schools to much lower levels than before class sizes were reduced, and most sharply in those schools with large numbers of poor students. In fact, after CSR, migration rates began to converge in all these schools, for the first time.
See this chart from Lawrence P. Gallagher, “Class Size Reduction and Teacher Migration: 1995–2000,” in: Technical Appendix of the Capstone Report, Part C, 2002, p. 47 (click to enlarge):
Indeed, several studies have shown consistently lower levels of teacher attrition after class sizes were reduced, in CA, NY state, and elsewhere. Here is a synergistic effect that would be expected to enhance teacher quality over and above the direct impact of a smaller class itself, as lower rates of attrition would increase the overall experience level and effectiveness of the teaching force over time.
3. Skoolboy points out that at higher levels of education, such as law schools, large lecture courses are popular with professors who engage in the Socratic method –as in the movie “The Paper Chase” – one of his favorites. He writes:
“Using the Socratic Method, an instructor can engage virtually all of the students in a class of 95 or larger.”
But there are two major problems with this example. One obvious is that the extremely high needs students that populate our NYC public schools are very different from the college graduates who attend our elite law schools, as NYC Educator pointed out in a comment to the blog. The other problem is that even in college and graduate school, class size still matters, as shown by a surprising wealth of research.
It is simply untrue that all the students in a college or law school class of 95 or larger will be engaged, or have their academic needs satisfied, to be able to receive adequate feedback on their written work or verbal contributions. It is even doubtful that real Socratic dialogue can take place in classes this large.
It is no wonder that Socrates himself engaged in debate and discussion with smaller groups of followers – and that as many former law students have described it, their lecture classes turned out to be more a way for professors to terrorize and humiliate them rather than engage with them in a truly intellectually productive manner. In fact, research indicates that large lectures in college contribute to high rates of student attrition, and that many college students find small classes to be far superior opportunities for academic growth and engagement.
See, for example, this paper by Joe Cuseo, “The Empirical Case Against Large Class size: Adverse Effects on the Teaching, Learning and Retention of First Year Students”, with a wealth of citations. Here is an excerpt:
Research on the lecture method strongly suggests that student attention and concentration tend to drop off dramatically after 10-20 minutes of continuous instructor discourse (Penner, 1984; Verner and Dickinson, 1067.) This attention “drift” occurs even among highly motivated postgraduate students (Stuart and Rutheford, 1978) and learning-oriented (versus grade-oriented) undergraduate students (Milton, Pollio, Eisen, 1986.) Among undergraduates in general, it has been found that about half of the time during lectures, they are thinking about things unrelated to the lecture content, with up to 15 percent of their class time spent fantasizing (Milton, Polio & Eison, 1986.)
In fact, one of the few good effects of the US News rankings is that colleges and universities are now rated in part on what percentage of their classes are 20 students or less – which has created an impetus to reduce class size even at such highly regarded but (some would say) overrated universities as Harvard and MIT.
4. I also have to respond to eduwonkette who says that there’s no problem with stuffing students into large classes at a high school like Bronx Science. Though it may be true these kids don’t face the same likelihood of failing their classes or dropping out of school as low-performing students at other high schools, many students at selective schools feel thwarted by not getting the feedback they need, on their writing, their ideas, or even being able to ask questions in class.
I had a couple of Brooklyn Tech interns who volunteered for me as part of their “Participation in Government” class, who felt very frustrated with the fact that they were unable to fully participate in class, because of its large size. Or read Frank McCourt’s book about teaching at Stuyvesant, who writes that at first, he thought he would get to know to know his students, but quickly gave up on the idea because of the class size.
Also see what Seth Pearce, a student at La Guardia, has written about the need to reduce class size in the NYC student union blog. Seth asked his younger brother, who now attends a private school, what the biggest difference was between his new school and the public school he used to attend:
His answer was quick and simple: the adults in the building have time to care about the students. In the NYC education system, the first step to improving schools is creating a situation in which educators have time to care about the students. This can only come for significant reductions in class size and teacher load.
Teaching loads in a typical NYC high school are about 150 students --almost twice what they are nationally. Let’s say that each teacher has one assignment a week to grade. If it’s 2-5 pages, it would have take at least five minutes to read it; to give meaningful comments, at least ten minutes. That’s more than 25 hours a week right there. Then if each student got a chance to speak one on one with the teacher for 5 minutes a week, that would be another 12 hours or so. So now the teacher has to spend 37 extra hours a week to do the bare minimum necessary to provide an adequate education – a whole second job. This is nearly impossible; as a teacher wrote me,
“The only students who can succeed in the current system are above-average intelligent children from a high supportive family environment. The others fall through the cracks.”
Which doesn’t mean that we shouldn’t reduce class size first in our failing or low-performing schools. But perhaps I’m naïve enough to believe that all public school students who live in NYC deserve the sort of education that those in the suburbs get as a matter of course – with teachers who actually have the time to care about them, and a classroom environment in which they can thrive.
5. I must say that I was relieved that Skoolboy decided not to do the sort of “back-of-the-envelope” calculations he originally envisioned, to make an estimate of the cost-benefits of class size reduction. Why? Because it’s already been done. In fact, the aforementioned Alan Krueger–former chief economist for the US Labor Dept – found that the economic benefits of smaller classes twice outweighed the costs, in what he calls a conservative analysis.
More recently, Peter Muennig in an article in the American Journal of Public Health, showed that class size reduction is not only a great investment in terms of educational benefits, but one of the best in terms of public health as well –rivaling vaccinations, with medical savings of $150,000 or more and almost two more years of life for each student who attends a smaller class in the early grades.
6. Finally, I have to note that Skoolboy’s critical musings on class size are far from untypical – and yet they still perplex me. There seems to be a real double standard in much of what is written on class size by many academics, for reasons I don’t fully understand. Why focus so much energy in trying to knock down class size reduction, given that it has a far stronger research base than nearly anything else attempted in our schools?
Indeed, the Institute of Education Sciences, the research arm of the US Dept. of Education, has concluded that class size reduction as one of only four, evidence-based educational reforms that have been proven to increase student achievement through rigorous, randomized experiments -- the "gold standard" of research. (FYI, the only other three reforms shown to work are one-on-one tutoring by qualified tutors for at-risk readers in grades 1-3 ; life-skills training for junior high students, and instruction for early readers in phonics) .
This combined with the fact that actual practitioners in the field, including teachers and principals, overwhelming respond in surveys that class size reduction would be the most effective way to improve the quality of teaching, I find it hard to understand why there continues to be so much intellectual effort expended in combating any efforts to achieve this. Why does class size attract criticism that other proposed reforms, including expanding the number of charter schools and small schools, implementing merit pay, school grading systems, or high stakes testing, seem to avoid -- indeed most of the strategies favored by this administration - none of which have the same evidentiary support?
It can’t be primarily that class size reduction is too expensive. Perhaps this is why CSR is targeted by conservative academics such as Caroline Hoxby or Eric Hanushek – because they oppose spending more money on anything in public education. Since class size has the strongest support in public opinion and in the research, they must feel driven to slay this dragon over and over again, to keep the threat of rising education expenditures at bay.
Nevertheless, there are many ways to reduce class size without hiring a whole new cadres of teachers, by redeploying existing staff that currently serve in out-of-classroom positions or hiring teachers in place of teaching assistants or administrators, as schools in Burke County NC, New Jersey, and Wisconsin have done, at little added cost.
This should be especially true in NYC, where the teacher/student ratio has grown sharply in recent years, with scant reductions in class size. Indeed, non-teaching teaching positions have proliferated like mad, including literacy and math coaches, lead teachers, now data coaches, and on and on and on. As one teacher wrote to me yesterday:
“There is no lack of teachers already in the schools and the pay roll; however, Klein has encouraged the principals to use these licensed teachers in non-teaching positions. The schools are full of teachers serving as deans, coordinators, staff developers, coaches, administrative assistants and who knows what else….In addition, in many schools these non-teaching teachers are taking up regular classrooms.”
Preliminary calculations of the new contract for excellence submission by the city show that while the DOE claims to have hired almost 2000 teachers this fall to reduce class size, fewer than 300 extra classes are projected – which suggests there is even more extra staff floating around than usual. Not to mention the estimated 700 teachers sitting in the rubber rooms, and 700 more in ATR (absent teacher reserve), who, in many cases, are doing little but acting as occasional subs.
Is this cost-effective, do you think? Why don’t the economists focus on all these inefficiencies, do you suppose, instead of spending their time arguing that class size reduction is too expensive?
Moreover, even if lots new teachers would need to be hired, is CSR really so expensive? Compared to what? It seems to me that if this reform actually works to improve outcomes, and leads to lower rates of disciplinary problems, dropouts and teacher attrition, as the research shows, it is far less expensive than other reforms that may seem on the surface less costly, but have little or no positive effect– or even have negative impacts, as for example, merit pay has been shown to have in North Carolina.
I suspect that if there were only four medical procedures which rigorous evidence had shown been shown to improve health outcomes, and that in survey after survey, hospital administrators and doctors themselves said that adopting one of these would be the most effective way to raise the standard of medical care and life expectancy, there would be little or no controversy in the halls of government or in medical schools about the need to implement it.
The debate would be only how fast, and how to finance this procedure, not whether it was “too expensive” or should be done at all.
But overall, there seems to remain a great deal of undue resistance. I’m no psychoanalyst, but I will speculate nonetheless. I think it may be that no one can make any money or personally profit from class size reduction. It’s a reform that seems awfully simple, and indeed it is.
Teacher colleges can’t market it, as they can a new method of professional development. Textbook companies can’t sell it, as they can a new curriculum. The research think tanks or consulting groups can’t make any money analyzing it, since the studies have already been done. Management consultants wouldn’t dare propose it, since the advice would seem too obvious. The testing companies and tutoring companies can’t use it to pitch their products.
Most of all, it doesn’t seem “trendy” or exciting to many academics who are eager to write their own best-selling versions of “Freakonomics”, or become famous on their findings, unlike a catchy experiment that rewards students with high test scores by giving them cell phones.
There’s also the administration’s control over the data, as Skoolboy himself points out in a comment, as well as its impressive power, wealth and popularity – especially with the business and foundation community. Given all this, perhaps it’s not surprising that criticism of most of the DOE’s unproven and often incoherent reforms is muted, that many academics feel inclined to go with the flow, and instead spend their time sitting in silence until after Bloomberg and company are gone.
There are notable exceptions to the rule; wonderful people in the academic community who have been unafraid to support class size reduction. For example, there’s of course Diane Ravitch, intellectual warrior supreme, but also see this terrific letter to the Governor and the State Legislature, organized last spring by Celia Oyler of Teachers College, asking that some of the additional state education aid be spent in NYC lowering class size in all grades. (More recently Celia put together this quite critical quiz on the new school grading system – which Teachers College has been unafraid to post on its website.)
So why aren’t more academics like Celia, and/or the other signers to her letter? Is it really because of the reasons I’ve cited above, or others yet unmentioned? Please enlighten me, or correct me if I’m wrong -- about any of this.