Over the years, many people have inquired about updating the chart below, which plots change in academic achievement for the nation’s 17-year-olds—the “final products” of our education system—against change in public school spending. The chart shows that spending grew precipitously while performance was basically flat. There are a couple of significant problems with updating the chart, and today I offer a stab at a new version that attempts to deal with them.
The first problem is that the test on which the original chart is based—the National Assessment of Educational Progress Long-Term Trend Assessment (LTT NAEP)—has not been administered to 17-year-olds since 2012. It is impossible to take it any farther. The good news is that in addition to the LTT, there is the “main” NAEP. It does not extend as far back as the LTT, but it does have longitudinal results up to 2019 for high school students; in its case, 12th grade. Note that various iterations changed from previous administrations, reflected in two scores—the old and new—in 1998 and 2004.
The second problem is tougher. The original chart plots the percentage change in inflation-adjusted funding for students’ full K‑12 careers against the percentage change in test scores. The main problem is that while percentage changes in spending are easy to understand, it is not clear what a percentage change in a test score means. Is a scale score rising 2 percent meaningful change or not? Second, while possible increases in spending are unbounded—you could get a one million percent increase in spending if you tried wasting taxpayer money hard enough—once a test score hits the top of the scale, it has nowhere to go.
To deal with the end of LTT in 2012 and scale score equivalency, the new chart does two things:
Adds main NAEP scores for high school students.
Plots the change in the percentage of students surpassing “proficient” on the “main” NAEP and the middle-performance level on the LTT. This compares both spending and performance in percentage terms and with a more concrete measure than a scale score. Still, readers must decide for themselves if the change is good, bad, or indifferent.
I also added two trends not in the original chart: change in GDP per capita and federal spending per school-aged American. The reason for adding GDP is that generally speaking, scores should rise if Americans’ material well-being increases and decline if it falls. I added the federal line because the federal role in education is being hotly debated right now, and it is important to have some broad evidence of whether federal spending is helping. Finally, to make isolating lines a little easier, you can add or subtract them from the chart by clicking on the legend.
What does the chart suggest?
While spending and GDP have risen substantially, high school performance in reading has been on a slightly declining trend. That suggests, that regardless of how much a given amount of increased spending should translate into higher achievement, spending has not been having the effect many want, at least on its own.
For math, the results are a little tougher. Linear trend lines show that roughly 2.8 percent annual increases in per-pupil spending and 2.7 percent annual increases in GDP coincided with a .26 percentage point annual increase in test takers exceeding the middle performance level in LTT math, and a .02 percentage point annual increase in students hitting proficiency.
The federal spending line is a little tougher to deal with because of the huge 2009 spending surge during the Great Recession. The trend is about a 3.6 percent increase annually, with the outlier included. Without it, the change is about 3 percent annually.
As mentioned, there is no definitive answer about the “right” outcome per dollar, but to put a little more flesh on the bones, here are the endpoints for achievement and spending:
The LTT math results are perhaps encouraging—an 8 percentage point rise in students clearing the middle score band—but overall, the chart suggests that we have not been getting much bang for our buck in education spending.
Does the chart need more work? I look forward to hearing from you.