by Citizen Stewart

Poverty and student achievement, as measured by standardized test scores, are highly correlated. How do I know this? Because people who think I put too much faith in education reform constantly tell me so.

It’s not that I feel poverty has no impact on school outcomes. It just that causation and correlation mean two very different things for me. Test scores are also highly correlated with white skin, yet, I doubt anyone believes white skin causes high test scores.

Save Donald Trump. He hasn’t weighed in on the issue, but you never know with him.

Here’s my question, if test scores and income are so tightly correlated, shouldn’t we see rising and falling scores when the economy falters, then recovers?

It’s a serious question. I honestly don’t know. I feel like Denzell Washington in this clip:



I truly need someone to explain it to me like I’m a toddler.

First, look at this chart from the Economic Policy Institute. It shows steady decline in real median household income for blacks between 2000 and 2010.


Screen Shot 2015-09-21 at 4.16.53 PM


This chart from Business Insider shows a longer view. Black median household income rose between 1960 and 2012, but not by much.

Screen Shot 2015-09-21 at 4.18.04 PM

Now, look at this chart that shows steady growth in black NAEP test scores. It’s actually fairly steady growth, even through seven economic recessions.


Screen Shot 2015-09-21 at 4.17.14 PM

Two points are weird to me: 1) there were sharp gains between 1978 and 1985, a time period when black median household income was depressed. And, 2) NAEP scores were fairly flat during the 1990’s when black income rose quickly.

So what gives?

If the correlation between poverty and test scores is so impregnable, how is it that scores went up during times of economic turmoil, and fell flat during times of prosperity?

Surely this there is a simple answer.

Since I don’t know, I’m going to ask people who spend a lot of time thinking about this stuff for their easiest answers.

Stay tuned.

Chris Stewart is the Chief Executive Officer of Education Post, a media project of the Results in Education Foundation. He is a lifelong activist and 20-year supporter of nonprofit and education-related causes. Stewart has served as the director of outreach and external affairs for Education Post, the executive director of the African American Leadership Forum (AALF), and an elected member of the Minneapolis Public Schools Board of Education.


  1. Thanks for reaching out about this question. The most important thing to note here is that a strong correlation between two variables does not imply that they always track each other perfectly. Anyone who claims median income is a perfect predictor of average NAEP scores is wrong (I hope nobody actually makes that argument). What we do know with more confidence than just about anything else in the realm of education research – and I believe you and I will be discussing the evidence for this idea more soon – is that non-school factors, economics being a primary one, play more of a role in differences in student outcomes than do school-based factors.

    Relevant economic indicators might include median pre-tax incomes (as shown in EPI’s chart) and might also include things like the poverty rate, features of our tax and transfer system, and levels of inequality, among other factors. Though it’s harder to study directly, various forms of both explicit and implicit racism also surely play a major role. Neighborhoods and schools are important as well. In addition, there’s a fair amount of unexplained variation in test score data (which appears to be about as influential in score differences as all school-based factors combined).

    Unless someone has studied a particular trend closely, I’d be skeptical of an explanation that claims to tell you exactly why scores went up or down during a given time period among a given subgroup, especially if that explanation asserts certainty. I unfortunately don’t have too much to offer at this time beyond the general points above, but the questions you ask are good ones and I’d at some point love to look more deeply into them.


Please enter your comment!
Please enter your name here