Skip to main content
Ed. Magazine

Lecture Hall: Assistant Professor David Deming

David Deming

[caption id="attachment_11877" align="alignleft" width="300" caption="Photo by Kathleen Dooher"]David Deming[/caption]

Economist David Deming has always been interested in long-run outcomes — how one thing affects another over an extended period of time. These days, he's jumping into a new research project based in Texas and Massachusetts that looks at the impact of high-stakes testing on outcomes other than the actual test scores. For the most part, he says, the past decade of research on the accountability movement in education has focused on two things: whether or not the tests increased academic achievement, and how high-stakes testing has led to certain behaviors such as teaching to the test or manipulating the data. The problem is that neither of these really answers the question that may, in fact, matter the most: Does school accountability improve how students fare long after they've graduated (or not graduated) from high school? Deming recently spoke to Ed. about test scores, a five-year award he was given to conduct the research, and digging deeper.

Why not just look at test scores? Test score gains are hard to interpret. Do they really represent gains in student learning? Do they help students in the long run? Are students more likely to attend college or earn more as adults? NCLB and state accountability policies mandate standardized testing, and there are all sorts of benefits to that. But we shouldn't only care about test scores. We need another way to judge.

Such as? I'm using administrative records from K–12 schools through the Texas Schools Project, linked across the years to college records and quarterly earnings records. The records include every public school student in the state from 1991 to 2011, so many of those students are now in their 20s and early 30s. We are looking at the impact of raising high school students' test scores on their attainment and earnings, later in life.

That's a lot of data. It's complicated, but it's what I know. For me, the data is the more comfortable part of the project.

What's the more uncomfortable part? The qualitative work. I'm trying to push myself. I've been interested in this question for a long time, and I think I can answer it, but I can't do it from 30,000 feet, only looking at data.

What will you do? I have only started the work in Texas, but I will eventually visit schools in Massachusetts to get information for the second phase of the research. The goal is to find pairs of high schools that have similar student bodies but produce different outcomes. One may be good at raising test scores, but only average at sending children to college. The other may be good at sending kids to college but not good at test scores. Or maybe they have both: high test scores and high college attendance.

What are you trying to figure out? I want to see if you dig deeper, can you figure out if the difference is because of how they responded to test pressure? High schools face a fundamental tension between preparing students for the high school exit exam and preparing them for college. You might think this would force difficult tradeoffs. Yet some schools have found ways to have their cake and eat it, too.

Could there be other reasons for the differences — better guidance counselors at one school or outside mentors for students? Yes. That's why I will start with many schools and then eventually hone in on the pairs that look like good comparisons. Sometimes what looks like a good comparison in the data turns out to be less than ideal.

Why go into the schools in Massachusetts? Why not just analyze the data like you're doing in Texas? I know how to do that already. This is an adventure, and it's going to be fun for me. Even if I don't do a very good job (distinctly possible), I am going to learn a lot, and it will help me ask the right kinds of questions in my future research.

As someone interested in long-term outcomes, what's your prediction for the long-term outcome of this research? Prediction is very difficult, especially about the future.

Thanks to the William T. Grant Foundation, you'll be spending the next five years working on this. By the time you're done, your oldest daughter, Maia, will be in the second or third grade. Does that seem crazy? Yes, and she loves to tell me how long it takes to count to my age. My wife gleefully plucked my first gray hair yesterday. Why is everyone around me reminding me that I am getting old?

Ed. Magazine

The magazine of the Harvard Graduate School of Education

Related Articles