Recently I had the opportunity to take a look at a very impressive chart. The chart is from cato-at-liberty.org and it appears to indicate that spending more on public schools doesn’t increase students’ academic performance.
Indeed if you look at the graph you see a steady increase in per pupil spending since the 1970s and basically no change in student test scores as measured by NAEP test scores. Wow. This means that you can spend more and more on students and you get no similar increase in test scores. Naturally, this means that all of the increases in student spending have not done anything to improve test scores, right. Wrong.
Statistics are like prisoners. If you torture them enough they will tell you what you want to hear.
A more interesting question would be to look at students with good test scores and students with poor test scores and see how they differ. It would also be interesting to see a graph of the cost of living relative to the cost per student in this chart.
Of course, because the author at the Cato Institute wanted to make a statement about spending rather than how to improve test scores they do not mention these other variables which may impact their analysis.
In discussing this issue on line I issued a challenge: “Show me a graph of Iowa test scores based on parental income or property taxes or per capita student spending that shows no difference and I will eat my hat!”
An astute observer noted that “There is a difference between parental income and per-pupil spending/tax rates/etc. Two totally different matters. Higher parental incomes do tend to correlate with test scores….There ARE factors that correlate very strongly with test scores.”
Very well said. My conservative friend actually proved my point although he did not know it.
The fact is that there are many other variables *other than* per pupil spending which affect test scores. (Tax rates, parental income, family types, etc, etc). Trying to address test scores by addressing per pupil spending may actually accentuate (and worsen) the trend of these other variables and result in worse scores in the student’s who face other structural (non-spending) related disadvantages. That’s pretty much the way multiple regression analysis works.
Here’s an example: A suburban school with 99% two-parent families pays 25% less per student than an inner city school with 10% two-parent families and the suburban school gets 25% better test scores. Would any reasonable person say that the solution would be to cut per student spending by 25% in inner city school???
I think not.
Complicated problems are multifactorial. They have many different interdependent factors which determine outcomes. There is no data to suggest that school test scores cannot be improved by decreasing spending on students. There are (as my good friend said) other factors which correlate strongly with test scores.
October 2nd, 2009 at 6:42 pm
I see what you are saying – it is true that MANY factors influence test scores. But it is also very true that one of the main solutions clamored for over the past 30 years in order to improve educational outcomes has been increased funding. Given that funding has more than doubled (in constant dollars), and test scores have remained stagnant, that would mean that either (a) increased funding doesn’t improve test scores or (b) there has been some severely negative changes in the other factors that affect test scores and educational outcomes, that more than offset the gains from the increases in funding.
Given that I have not seen any clear data to suggest that (b) is true to such a huge extent as to offset any gains from increased spending, I still rest on the conclusion that increased spending is not yielding sufficiently positive results.
I leave the burden of proof on those who say otherwise, to find out and point to these severely negative changes in the other factors that affect outcomes that they believe are the “true” cause of test score stagnation.
Also the “cost of living” issue you mentioned is already taken into account. The data is presented in constant 2008 dollars, meaning the data is all inflation adjusted. The $5000 data point for 1970 is $5000 in 2008 dollars – which means the actual amount of money spent in 1970, in 1970 dollars, was something much less than $5000.
October 2nd, 2009 at 6:51 pm
Thank you for the thoughtful response.
Briefly, there is actually another explanation. Academic achievement is a bell shaped curve. I would not expect it to change in a linear fashion over time. Indeed, if I saw a graph with rising test scores for all students over time I would actually suspect (artificial) grade inflation rather than actual achievement gains.
Per pupil spending IS NOT guided by any bell shaped curve; and even if the cost of living is factored in with constant 2008 dollars, the increase could still be accounted for by new technology and other changes. How many computers were there in classrooms in 1970? How many Powerpoint slide projectors? etc. This (plus the other factors I’ve enumerated) could account for spending increases.
October 2nd, 2009 at 7:09 pm
Yes, I too have thought about the “technology” issue. Certainly there is much more expense to properly outfitting a classroom these days, and in fact those additional expensive gadgets likely will NOT help increase Math or Reading scores (hopefully they will help our children learn the technology and how to use it, and learn history, science, etc. better).
And you are correct – test score data should lag spending. Any improvements from increased spending shouldn’t show up until several years down the road.
But what concerns me is the trend line. Spending was increasing pretty much all throughout the 70′s and 80′s – a time when, well, there just wasn’t a lot of additional expense being incurred to outfit a classroom properly (sans the tiny Apple computer we had in 4th grade in 1982!). Yet test scores were stagnant.
And then when spending levelled off in the early/mid 90′s – a time when the computer age was just getting going – there didn’t appear to be a negative effect on test scores then or a few years down the line.
I’m all for digging into the numbers more – I agree that at a high up view the “truth” can be hidden in the statistics. But I still have my doubts on this one, that our increased spending is paying off unfortunately.