Good teachers trump class size

Hiring (and supporting) better teachers is more important than keeping classes small, argues Jay Mathews on his Washington Post blog.

He suggests a thought experiment: Would you want your child “transferred to the school’s best teacher, an imaginative and energetic motivator” with 30 students in the class?

. . .  when the Center for Public Education examined 19 studies of class-size effects that met its research standards, it reached two interesting conclusions. First, most of the studies focused on kindergarten through third grade, and most of the beneficial effects of smaller classes seem to occur in those years, when students are learning to read. Spending money on class-size reduction for those kids makes sense, as several local school systems have shown.

Second, the studies showed little effect from class-size reduction unless the number of students was 20 or fewer, and little effect in middle or high schools.

High-quality schools in low-income neighborhoods typically focus on improving instruction, not on offering small classes.  Classes small enough to make a difference are too expensive.

Tennessee’s STAR study found significant, long-term benefits for students, especially blacks, in K-3 classes of 14 to 17 students.

However, when California reduced K-3 classes to 20 students, teacher quality declined, especially at schools serving low-income, minority students. That wiped out any benefits. A research consortium “found no relationship between statewide student achievement and statewide participation in class size reduction.” To keep classes small, schools are cutting funds for libraries, after-school programs and teacher training.

About Joanne


  1. “A great teacher can teach 60,” Rafe Esquith tells Mathews. “A poor teacher will struggle with five.” That’s no doubt true. But the great teacher is even better with fewer students. I has as few as 18 5th graders and as many as 34. I was far more effective with 18. Like so many debates in education the choice between better teachers and smaller classes is not an either/or proposition. If I had to choose between putting my daughter in a class of 30 with the schools best teacher or the worst with 15, of course I’d take the former. But I know she’d be better off with that great teacher in a class of 20-22.

  2. I agree with the previous comment. Right now schools are making hard decisions and those decisions include larger class sizes. Unfortunately these decisions have nothing to do with teacher quality. If smaller class size makes an already great teacher better than larger class size makes an already weak teacher weaker.

  3. This blog leads the reader to believe that smaller class size only matter in the primary grades and/or only when class sizes are very small. Research on single classes supports this theory, but research on overall teacher load at secondary levels does not support your conclusion. Here’s an example of how it’s misleading.

    High School Teacher A teaches 5 classes with an average of 22 students in each class.

    High School Teacher B teaches 6 classes with an average of 27 students in each class.

    Teacher A’s total student load is 22×5 or 110 students and teacher B’s total student load is 27×6 or 162 students.

    If teacher A and B assign the same 4-page paper which takes 10 minutes to read and correct, teacher B will spend 9 more hours assessing and giving feedback. Teachers are normal people and have fixed amounts of time each day to work with students, design curriculum and give assessments. Therefore, the quality, quantity and timely response of feedback to the classwork/homework will vary widely. Research clearly links timely, high quality feedback to successful student learning.

    This one example shows the complexity surrounding the class size issue which you oversimplified in your post. Teacher contracts, class schedule, student demographics and subject area make a huge difference in teacher effectiveness relative to class size. Any administrator or policy-maker for education institutions are poorly served by these types of research “summaries”. Research based policy-makers and data-driven educators continue to bring the desire for corporate-styled objective performance indicators to the education realm. Not all areas of education can be placed into such neat and simple statistics.

    Schools don’t get to choose whether to provide students with good teachers OR small classes. They simply take the good teachers with the bad and give them as many kids as the current political environment will financially support. These discussions provide nothing but anecdotal evidence for politicians and anti-tax groups to use for support of their fiscal policies.

  4. Mark Roulo says:

    If I had to choose between putting my daughter in a class of 30 with the schools best teacher or the worst with 15, of course I’d take the former. But I know she’d be better off with that great teacher in a class of 20-22.

    What do we do with the other 8-10 kids? Would you be okay with a school where the best teacher got 20 kids, the worst teacher got 10 and the kids were assigned based on lottery (including your own, of course)? Or would you prefer all 30 kids to be assigned to the best teacher? Or what?

    Picking “great teacher, small class” doesn’t appear to be an option for everyone, especially in the current financial climate.

    -Mark Roulo

  5. To be honest, Mark, I don’t see the linkage between teacher quality and class size. It’s not as if schools are sitting around saying, “Hmmm, let’s see. I’ve got X dollars. Should I hire great teachers, or hire extra bad teachers and reduce the number of students in each room?” Neither is there a magic number that each teacher has beyond which he or she is no longer effective. Every class is different and takes on a different dynamic.

    As for what to to with the other 8-10 kids, what about the kids in that “worst teacher” classroom? If we follow your logic to its logical conclusion, we’d put EVERY student in the good teacher’s room and give the bad teacher none.

  6. @Steve:

    Sounds to me like you’re the one providing the anecdotal evidence.

    And riddle me this: Why would a highly effective teacher do something so monstrously stupid as assign four-page papers due in all six of his/her classes on the same day?

    Funny you should allude to Marzano’s findings re: timely, effective feedback, then ignore his findings that teacher effectiveness–not class size–directly correlates with student learning.

  7. Who said the teacher assigned it on one day? Or that the teacher was highly effective? It could be one week or one month – it’s still 1/3 more time (9 more hours). Read before you so rudely reply

    Again – the premise is the problem – you cannot rid our schools(or our businesses) of average people. What is the impact of learning when class size goes up with a bad teacher?

  8. Bill Leonard says:

    I wonder whether class size — to a point, obviously — counts as much as homogeneity, racial/ethnic group cultural standards, cultural norms, and quality of teachers (which is important, but not decisive).

    I went thru the first six grades in the Des Moines school system of the late 40s and early 50s. Every kid was from what now would be quantified as a “poor” family; all of us were from semi-rural families, and many were fronm farm families. This was just after the tail end of WWII, and a signficant minority of kids were the children of displaced persons (DPs, or the “refugees” of later years) from Europe.

    But we were, overwhelmingly, from the same western/northern/eastern European background. Overwhelmingly Protestant; there may have been 15 Catholic kids in a school of around 400, K-6.

    I think it didn’t hurt that the Des moines school system in those days was structured so that no kid moved from second to third grade unless he/she was reading on a second-grade level. Easy to do when you can start kids in kindergarten at mid-year, and promote kids into second, third or whatever grade at mid-year.

    In retrospect, some of the teachers were spectacular; some were not. Most were competent. But we all learned.

    And no class (I started in a 2-room school with outdoor plumbing, in 1948) was less than 30 kids, minimum.


  9. Margo/Mom says:

    It is true that no single variable can account for everthing that comes out of a classroom. This ought not, however, remove focus from the issue of varying teacher quality–which comes up again and again. Based on such crude proxies as experience and certification we can observe that teachers with more of these things tend to migrate towards schools with more resources and students who are more socio-economically advantaged. The existing classroom size research does in fact support an impact–but in the lowest grades and the class sizes must be quite small. A reduction from twenty-eight to twenty-two at the fifth grade level is very expensive and is not likely to have significant results.

    Steve’s discussion of student load is something different than classroom size. I would note that teachers in Singapore elected to have larger classes in order to have a bigger block of time during the day for planning, grading and collaborating. In other words, three classes of forty rather than four classes of thirty. Until fairly recently, class size in Asian countries overall has tended to be quite large, but I believe that actual “face time” with students has tended to be smaller–with an emphasis on teachers working together to plan and evaluate lessons.

    Hanushek suggests that the factors that affect teacher quality may be more complex than we can account for in looking for linear relationships between such things as teacher test scores, certification and licensure and years of experience. Yet, there are observable and measureable differences between teacher effect. He suggests that a certain amount of building autonomy is needed in such things as teacher assignment, perhaps class size, in order to make the kinds of matches that make best use of all teachers. While most would consider it anathema to construct the kind of classrooms that Robert suggests (much smaller number with a weaker teacher and a greater number with a strong teacher), or to match the skills of a teacher with the needs of students–some research suggests that this is effective (Hanushek looks at this as a factor in high performing countries; a California study compared high performing and low performing schools of similarly low SES). Some high performing countries have a reduced teaching load for teachers just entering the classroom–we tend to do the opposite. The new teachers get the left-overs and left-outs. New teachers are likely to have more challenging and low performing students and to have more preparations.

    We really need to move in the direction of applying some of this knowledge. Instead we keep recycling the mythology that the only thing that matters is classroom size.

  10. I think that discipline is a significant factor. A relative attended a Catholic school, in the 40s, which had 100 kids per class. One nun, without college degree or classes, per class. Lots of first-generation kids who spoke another language at home. Real education. No discipline problems.

    In today’s classrooms, especially at the lower SES level, LOTS of time is wasted on discipline issues. I think heterogeneous grouping contributes to disciplinary problems,since the yardstick, dunce chair and other public humiliations the nuns (and public school teachers) used to employ are no longer permitted and parents are more likely to complain about school discipline.

    Except for foreign languages, my kids’ AP classes often had 30+ kids, but it was never a problem, since they all had had the same prerequisite classes.

  11. This is an interesting discussion. I’d like to thank the people who took the time to comment; I have learned something from most of you.

  12. And to notice that dog that isn’t doing much barking; how is teacher quality to be determined?

    The STAR study doesn’t look like it’ll be of much use in differentiating between good and bad teachers on an on-going, day-to-day basis. Any other widely-used measurement instruments in use? No? Then the discussion’s academic the irony being that the academics in question aren’t all that interested in the issue.

  13. Lightly Seasoned says:

    I wish some of these studies would pass through a common sense filter.

  14. Andy Freeman says:

    > It’s not as if schools are sitting around saying, “Hmmm, let’s see. I’ve got X dollars. Should I hire great teachers, or hire extra bad teachers and reduce the number of students in each room?”

    Actually, that is exactly what they’re doing. The only question is whether they know that that is what they are doing.

    They have X dollars and they’re deciding how to spend it. Money that they spend one place is money that they can’t spend some other place.

  15. Well done Margo/Mom! Many, many high performing countries have students going to school more and teachers teaching less. They also, as Bill alluded to, have less diversity – culturally, socio-economically, and intellectually. Large class sizes, even multi-age classrooms, are much easier in a mono-cultural setting.

    Andy – X dollars can’t be spent on anything an administrator chooses. In fact, most building level leaders I work with believe that their budgets are completely tied up by the Board of Ed, Superintendent, Union, etc.

    Teacher tenure and union contracts – especially in states with binding arbitration – make it nearly impossible to simply decide that they want to spend X on better teachers. They can’t afford the law suit to get rid of a mediocre/bad teacher and they don’t control how much teachers are paid. To “buy better teachers” and increase class size is a non-issue because it’s not real life. There aren’t any schools I know choosing to hire bad teachers. Firing or non-renewing mediocre teachers is a crap shoot – the candidate pool in our system has been pretty poor lately. People can make a lot more money in other jobs. Sometimes the devil you know is better than the devil you don’t know.

  16. Margo/Mom says:


    The homogeneity argument looks good in Japan, possibly even in Finland. Much less so in Singapore, Hong Kong, New Zealand, Australia or Ontario.

  17. mike curtis says:

    Teachers don’t get to pick their students. Therein lies a corollary; is a good teacher more effective in a classroom with 15 behaviorly disruptive and apathetic dullards, or, when surrounded by 30 self actualizing learners who know why they are in a classroom?

  18. @Steve:

    So your stance, in essence: our system must employ teachers of low quality; get used to it, and compensate by reducing class size?

    I suppose there is plus side to such an argument: fewer students must suffer at the hands of the incompetent educator.

    Sorry, some of us aren’t willing to accept this “reality” as an inevitability.

  19. Bill Leonard says:

    Momof4 brings up a key point that has been rather overlooked in this discussion: classroom discipline. In the K-6 schools I attended, there was never a question about who was in charge in the classroom.


  20. Andy Freeman says:

    > Teachers don’t get to pick their students.

    No one gets to choose their inputs. Why should teachers be any different? And why don’t public school advocates know that no one chooses their inputs?

  21. Andy Freeman says:

    > Andy – X dollars can’t be spent on anything an administrator chooses. In fact, most building level leaders I work with believe that their budgets are completely tied up by the Board of Ed, Superintendent, Union, etc.

    > Teacher tenure and union contracts – especially in states with binding arbitration – make it nearly impossible to simply decide that they want to spend X on better teachers.

    I never said that each and every person in the public schools had complete freedom to spend as they saw fit. I said that the system as a whole makes spending decisions.

    Those decisions tell us what the decision makers actually value.


  1. […] teachers or smaller class size?  Joanne Jacobs launches the discussion with Good teachers trump class size posted at Joanne […]