Where do bananas go to get educated? Sundae school. It was good enough to get a guffaw from my nine-year-old. Here’s something funnier: A recent Colorado State University (CSU) study on the social factors of educational outcomes tested among its 43 variables the price of a banana. If you don’t believe me turn to pages 8 and 15 of the study. You should not be terribly surprised to learn they found no significant correlation. (That said, check out the Spurious Correlations website for examples of meaningless statistical relationships.)
But the study is not completely bananas (sorry). Its data “are consistent,” it says, with widespread “findings that students from higher income and more educated households perform better academically.” This goes in the “no kidding, Sherlock” file.
The obvious conclusion to draw from such findings is that public school just doesn’t matter very much in terms of educational outcomes. Wealthier, highly educated parents will make sure their kids get a good education regardless, while the public schools generally do a poor job of helping disadvantaged students overcome their challenges.
Many parents like school because it’s free babysitting; kids often like school because it’s a place to hang out with their friends. Education often has little to do with it.
The exception is high-level classes such as Advanced Placement. Kids actually often learn useful things in such classes. But, generally speaking (with exceptions), you’ll find that the students in advanced classes come from wealthier, highly educated families, so we’re back to the original issue.
The CSU study reaches the opposite conclusion of the obviously correct one. It concludes, in essence, that because it’s practically impossible to change parental income and education levels in the short term, we should instead focus now on tweaking the public schools with reforms that we know in advance will have little effect.
In the words of the study: “[T]hree variables [are] particularly highly correlated with academic performance[,] those [that] measure the income and education level of the student households. In this analysis, the most highly correlated variable was the share of students NOT eligible for free and reduced lunch (a clear proxy for income), closely followed by the share of adults with a higher education degree and median family income. These ‘pick your parents well’ variables will emerge as highly significant in the regression analysis as well and serve as a continuing reminder that working to improve socioeconomic conditions in all households remains a powerful strategy for improving academic outcomes. However, . . . these variables require multigenerational approaches to poverty alleviation and largely are beyond more immediate policy responses focused on the school, individual family or student. For that reason, it is important to explore other more immediately addressable factors.”
The children currently failing to learn much of anything in the public schools might be forgiven for thinking that this “kick the can down the road” strategy is a bad one.
The failures of the public schools
Let’s quickly again review the 2022 Colorado results from the National Assessment of Educational Progress for fourth grade. In math, while 49% of white students and 65% of Asian students scored proficient or better, only 18% of Black students and 17% of Hispanic students did. In reading, while 49% of white students and 50% of Asian students scored proficient or better, only 21% of Black students and 20% of Hispanic students did. In both cases, the gap between students on the National School Lunch Program and other students was comparable to the difference of the racial gaps.
Results in Denver were better for white students but worse for black and Hispanic students. In math, while 62% of white students scored proficient or better, only 12% of black and 13% of Hispanic students did. In reading, while 63% of white students scored proficient or better, only 14% of black and Hispanic students did. In this case, the racial gap was worse than the school-lunch gap.
These disparities are shocking, but according to Denver Public Schools, there’s nothing to worry about. Everything is fine! Here is the headline for a June 25 DPS media release: “Denver Public Schools Celebrates Spring READ Act Results, Demonstrating Growth and Closing Achievement Gaps for K-3 Students.” DPS is “thrilled to announce” the results, the release states.
The release states, “Of all the students tested, an impressive 61% were reading at grade level or above, marking a 3% point surge from Spring 2023 and a remarkable 12% point increase since Fall 2023. Additionally, the data showed a 4% increase for Black students, 7% increase for Hispanic students.”
Let me do the math. DPS says it is “impressive” that 39% of DPS students are failing to read at grade level. I would not call that result impressive. I would call the fact that so many children are failing to learn how to read well a moral crisis.
Notice something strange here: The DPS release gives the overall percent reading at grade level, but it gives only the percent increase for black and Hispanic students.
Melanie Asmar’s article for Chalkbeat offers some context. Despite DPS’s bragging, “test scores are still below pre-pandemic levels,” Asmar writes, adding: “The district did not release a breakdown by race of the spring 2024 early literacy test scores.”
I asked DPS about this, and here are the numbers the district provided: While 78% of white students and 64% of Asian students were reading at or above grade level, only 52% of black students and 54% of Hispanic students were. It’s pretty bad that over a fifth of white students are not reading at grade level. It is catastrophically bad that nearly half of black and Hispanic students are not.
We can ask why the new results seem better than the NAEP results. I have not checked out the test in question. NAEP has a lower level than “proficient” that it calls “basic”; this seems closer to what Denver considers “grade level.” The time lag is important; students are doing better now than they were during the pandemic.
Teaching children how to read
A revolution is coming to the public schools. Educators, pushed by the legislature, have this newfangled idea that, to get students reading, educators should teach students how to read. I know it sounds incredible.
Shockingly, for many decades, the standard in the United States has been for teachers not to teach children how to read, then act surprised when many children cannot read. For details, see the podcast series Sold a Story.
Asmar observes that the relative improvements in Denver results follow the district “switching its elementary reading curriculum to one aligned with the science of reading.”
Obvious to everyone but professional educators for several decades, phonics is an essential component of learning to read. I learned to read phonetically on my mother’s knee, sounding out words. I helped teach my own son how to read partly using the phonics-heavy “The Reading Lesson” by Michael Levin and Charan Langton.
Asmar explains: “In the 2022-23 school year, DPS rolled out a new reading curriculum in kindergarten, first, and second grades called Core Knowledge Language Arts. The switch was prompted by a 2019 state law requiring schools to use scientifically based early reading programs.
“DPS’ previous curriculum was rejected by the state for not following the science of reading, a large body of research on how children learn to read. One key belief is that teaching phonics in a direct and systematic way helps build skilled readers. DPS’ new curriculum is generally well-regarded both for teaching phonics and building students’ background knowledge, which helps with reading comprehension.”
Of course a curriculum does not teach itself. Success depends on the willingness and ability of teachers to implement it well. Hopefully, though, teachers will increasingly make an effort to help teach children how to read, and those efforts will bear fruit over time.
The weak finding for school spending
In evaluating the factors for reading success, the CSU study doesn’t say anything about the importance of teachers teaching children how to read. But, hey, when we have things like the price of bananas to discuss, who has time for such trivial matters.
When the study looks at 43 variables and their relation to third grade reading, it finds that what most matter, by far, are factors closely related to family income and education. Then the study works the statistics to come up first with nine and then with eight of the variables as positively related to a statistically significant degree.
One strongly related factor is so-called churn, referring to how much students come and go in a classroom. This also largely relates to household income and therefore stability. The study cites things including “housing instability,” “unstable relationships or behaviors,” and “employment mobility of the parents” (among other things) as contributors to churn. The study recommends a few things to try to reduce churn or compensate for its effects.
“Total override funding” and “average per pupil funding” make the initial list, but the latter is lower on the list than things like “tree canopy” and “banana price.” With the different modeling, override funding drops out of the narrower list, while total per pupil funding remains.
The study explains that, with the narrower model, “override funding for schools is no longer significant; however, total per pupil funding retains significance. . . . [M]uch of the variation in per pupil funding is a result of override funding approved in the district. The more restrictive specification retains the finding that funding levels are positively associated with reading performance even while the override component does not retain significance.”
The study points out, “It is reasonable to question whether it is the funding levels per se or whether the funding level is enough associated with household wealth and income capacity in the district to argue that the correlation is spurious—that the true association is with community income levels.” However, largely because of “the base formula which augments funding challenged districts . . . there remains evidence that . . . funding levels matter.”
Sure, it matters, a little. This says nothing about whether districts could use its existing resources more effectively to achieve better results.
The promise of more choice in education
The CSU study finds that what matters overwhelming to student success are family income and education levels. The study favors “multigenerational approaches to poverty alleviation” that do nothing to help struggling kids now.
But the CSU study totally ignores a straightforward policy idea for immediately increasing the household income of families with school-aged children: Convert existing school spending to means-tested educational vouchers or (better yet) no-strings-attached cash transfers. By means-tested, I mean give poor families more.
The CSU study does not consider a voucher or subsidy proposal because the study has an obvious leftist progressive bias. If it did not, it would consider conservative and libertarian policy ideas, not just leftist progressive ones. The study is partly an exercise in rationalizing policies that progressive Democrats and the teachers’ unions already want.
That said, a finding that higher household income is correlated with better educational outcomes does not demonstrate the causal flows. The same qualities that tend to make adults higher earners and better educated, such as being conscientious and long-term planners, also tend to make them better at ensuring that their children get a good education. Just handing people money does not automatically alter those underlying factors.
Still, if less-well-off parents have more resources, it’s plausible many will spend some of those resources on things like better schools and tutoring for their children.
If we look at the Colorado Department of Education’s handy financial transparency portal, we find that per-pupil spending in Denver is $17,591, higher than the state average of $14,845. If, instead of giving all that money to the public schools under the control of the teachers’ unions, government simply gave every Denver family $17,591 per school-aged child (or more adjusting for poverty), educational outcomes almost certainly would quickly improve. At least they couldn’t get much worse than they are now!
Notably, the CSU study finds it surprising that a working mother is significantly related to reading outcomes “but did not carry the hypothesized positive relationship.” The study plausibly says, “Working mothers in more economically challenged households, particularly those who are working multiple jobs with extended hours likely are associated with reduced academic performance.”
Again, an obvious approach is to give families the money directly, so that parents can have more flexibility to avoid high-stress, low-paying jobs and spend more time with their children.
A personal progress report
All children are different. Some children have an especially hard time learning how to read because of dyslexia or some other neurological issue. So, even in the best of conditions, we have to expect a lot of variance in the tested academic performance of children of a given age.
What I can tell you is that my homeschooled nine-year-old, who is officially a fourth grader as of this Fall (not that grade levels mean much when you’re homeschooling), recently finished the fourth Harry Potter book and scored 100% on a year-end assessment for the challenging Dimensions Math fourth grade program. He’s well ahead in math and reading, having never in his life spent a single day in public school.
Yes, our household fits the expected patterns, with a household income a little above the average and with both parents holding four-year college degrees. Of course, now that the state board of education has cut us off from the meager tax-supported program we were enjoying as homeschoolers, we’ll get zero of our education-directed tax dollars to use for our son’s education. (Other tax-supported programs do exist for homeschoolers, but they don’t work well for us.)
The public schools could do a lot better job teaching children the basics with the resources they already have. But the public schools have, at enormous expense, largely failed in their task of helping children become literate and numerate. I think it’s time to consider the admittedly radical alternative of instead empowering parents.
Ari Armstrong writes regularly for Complete Colorado and is the author of books about Ayn Rand, Harry Potter, and classical liberalism. He can be reached at ari at ariarmstrong dot com.
