We've moved!

Social Icons

twitterfacebooklinkedinrss feed

Thursday, December 2, 2010

Spend More on Kids and Teachers, Get Higher Math Scores? Maybe!

In the conversation about our schools not producing enough smart kids, I mentioned Stanford economist Eric Hanushek's contention that per-pupil spending doesn't correlate with educational outcomes. Former Madison Superintendent John Sweet challenges this contention, suggesting we look at spending in Massachusetts and Minnesota, where the percentages of kids rating "advanced" on math tests are highest in the nation. (Dr. Sweet also happens to now supervise the tutelage of smart kids in Minnesota.) But what, retorts neighbor Linda McIntyre, of high cost and low results in DC and the rez?

O.K., let's check some numbers.

The report from Hanushek, Peterson, and Woessmann (in PDF!) offers a nice table giving percentages of kids getting advanced math scores in each state. The Census Bureau offers data on per-pupil education spending in each state. I paste those numbers into Excel and run the CORREL function. The results:

Correlations with percentage of students scoring "advanced" in math
Total 0.37
Salaries and Wages 0.40
Employee Benefits 0.30
Instruction Total 0.39
Instruction Salaries
and Wages
Instruction Employee Benefits 0.33
Support Services Total 0.33
Pupil support 0.17
Staff support 0.25
General administration 0.18
School administration 0.27

A correlation of 1 means an absolute postive relationship (when one number increases, the other one always increases). A correlation of 0 means absolutely no relationship, or complete randomness (the numbers have no more connection than a list of results from rolling dice).

A correlation of 0.37 means there's some positive relationship, but that other factors can wash out that relationship. That fits what we see when we line up the states by per-pupil expenditures:

State per-pupil expenditures (2008) % adv. math scores
United States
New York 17,173 6.3
New Jersey 16,491 8.7
Alaska 14,630 5.8
Vermont 14,300 8.8
Connecticut 13,848 7.8
Wyoming 13,840 3.5
Rhode Island 13,539 3.3
Massachusetts 13,454 11.4
Maryland 12,966 6.8
Delaware 12,253 5
Pennsylvania 12,035 5.7
Hawaii 11,800 2.5
New Hampshire 11,619 6.5
Maine 11,572 5
Wisconsin 10,680 6.7
Virginia 10,659 7.9
Illinois 10,246 5.4
Ohio 10,173 6.6
Minnesota 10,140 10.8
Michigan 10,069 5.5
Louisiana 9,954 1.7
California 9,863 4.5
West Virginia 9,852 1.4
Georgia 9,788 4.3
North Dakota 9,675 4.8
Kansas 9,667 5.2
Montana 9,666 5.6
Nebraska 9,577 6
Oregon 9,558 7.3
Iowa 9,267 5.7
Missouri 9,216 4.1
South Carolina 9,170 6.7
Alabama 9,103 2.3
Washington 9,099 8.7
Colorado 9,079 6.3
New Mexico 9,068 1.4
Indiana 9,036 5.1
Florida 9,035 4.6
Kentucky 8,686 3.4
Arkansas 8,541 3
South Dakota 8,367 6.5
Texas 8,320 6.2
Nevada 8,285 3.1
North Carolina 7,996 7.1
Mississippi 7,901 1.3
Tennessee 7,739 2.9
Oklahoma 7,685 2.4
Arizona 7,608 4.6
Idaho 6,931 4.5
Utah 5,765 4.7

Among the top ten big spenders, you have six states above average on the advanced math score and four below. Among the bottom ten low spenders (that includes South Dakota!), you have three states above average and seven below. That's far from a perfect correlation, but it does appear that, all other things being equal, if you make more money available to schools, they'll have a better shot at turning out smarter kids.

Now look again at the first chart, which breaks down the correlation by how the money is spent. The strongest correlation in the batch is the 0.42 you see between advanced math scores and teacher pay. The weakest correlations are the 0.17 for pupil support and 0.18 for general administration. Again, none of these correlations are guarantees of results, but if you have money to bet on improving student performance, these numbers suggest the best place to put that money is right into your teachers' pockets.

Of course, the problem is that to understand this argument, we need more kids who get good math scores....


  1. Michael Black12/02/2010 9:16 AM

    I wonder what the numbers say about participation in athletics giving an edge in student performance. Do you think that players do better in school because they are part of a team?

  2. John Thune seems to think so. And I have seen numbers saying that at the collegiate level, South Dakota athletes maintain higher GPAs than the overall student population. But as with the above money numbers, that's correlation, not causation. Do college athletes do better in school because they play sports? Or does the same ambition that drives them to play sports also drive them to hit the books harder?

    And if we can find numbers on sports participation, I want them placed side by side with numbers on arts participation. I'll bet any academic gains from sports are swamped by academic gains from music, debate, drama, visual arts, and other such activities.

  3. Michael Black12/02/2010 4:02 PM

    I see coaches as great motivators to a student's success. The athletes must maintain a certain level of academic performance or they cannot participate in their chosen sport.

  4. Compare Texas and New Mexico, two states that spend similar amounts on education, that lie next to each other geographically, and that differ drastically in math scores. How can we explain this vast disparity? (Note: New Mexico has a personal income tax; Texas does not.)

    It appears to me that South Dakota does pretty well in math as a function of its education spending, relative to states in general. How can we explain that phenomenon?

    The real question, methinks, is "How can we get the most 'bang for our bucks' in mathematics performance?" and then we can go on to ask "How can we optimize our mathematics performance in the absolute sense?" Finally, we might ask ourselves "How important is good mathematics performance to us, really, in relation to other factors?"

    We might look to Asia for answers. They're killing us in mathematics proficiency today, just as they've been doing for decades. They're starting to swamp us economically, too. I suspect that it will all come down to good old fashioned hard work and dedication on the part of the individual, but of course I cannot prove that. Maybe Robert Reich can, if he's-a mind to.

  5. We always have to optimize/make do with what we've got. But South Dakota schools and teachers are being asked to make do with awfully little... and I worry the Russ Olson Senate is going to ask them to make do with even less to balance the budget. Even if money isn't the magic input, it is the moral input—i.e., we should pay people a good wage for good work.

    Stan, when you mention Asia, you hit a big point behind Hanushek's research. He opens with this:

    "Maintaining our innovative edge in the world depends importantly on developing a
    highly qualified cadre of scientists and engineers. To realize that objective requires a
    system of schooling that produces students with advanced math and science skills."


    "In a prior study, two of the authors of this report
    demonstrate that growth in the economic productivity of a nation is driven
    more clearly by the math proficiency of its high school students than by their
    proficiency in other subjects."

    I'm thinking the best route to better math scores is better math teachers. Maybe we need to change the teacher certification process to focus less on traditional teacher training and more on having a major and experience in the field. (I'm not sure how we make that work at the elementary level, though!)

  6. Maybe we need to change the teacher certification process to focus less on traditional teacher training and more on having a major and experience in the field.

    Interesting -- you less training on how to teach and more focus on the subject matter would make for better teachers? My gut (and my husband, who maintains that is the biggest problem with teachers at the college level) would tell me the opposite.

  7. I'm hoenstly not sure, Kelsey. Teacher training may not be the issue at all. Maybe it's too much TV and texting, not enough parental involvement.

    If teacher training is part of the issue, then we either need less, mroe, or different. I was not terribly impressed with what was taught in my education classes: my best teacher training was practice. I didn't feel much scholarly rigor in my teacher training classes. I could have done with replacing those classes with... something. Maybe more subject area classes? Maybe very different teacher training classes? Maybe more teaching internship experiences?

  8. Hire math teachers to BE math teachers, not coaches who happen to also teach math. There is a big difference. I do not know the math teachers currently in Madison HS, but this was a problem years ago. Same thing with some other subjects.


Comments are closed, as this portion of the Madville Times is in archive mode. You can join the discussion of current issues at MadvilleTimes.com.

Note: Only a member of this blog may post a comment.