Interview with Larry Hedges

Larry Hedges is chairman of the department of statistics and a professor of education and social policy, psychology and medical social science at Northwestern University in Chicago. One of the most influential applied statisticians in the world, he is credited with developing sound statistical methods for meta-analysis. Last year he was awarded the Yidan Prize for Education Research, which comes with HK$15 million (£1.5 million) of prize money and the same amount in project funding

When and where were you born?
In 1952 near Fresno, in a region of brutally hot summers and miserably cold winters – the Central Valley of California, which is famous for having seven of the 10 poorest counties of the state. That was true when I was growing up and it’s still true.

Your mother washed dishes at a dormitory at California State University, Fresno. Did you have a feeling of being outside looking in?
As a child I was interested in being a scientist. It was something I saw on TV and it sounded romantic. My mother would say, “People like us can’t do that.” She said that often. I came to understand it later as a loving gesture. She didn’t want me to experience the pain of being slapped back down to my status. That’s helped me, as a social scientist, to understand class in a much more visceral way. It’s not just how much money you make; it’s how you see the world and how you learn what you can expect.

How did you respond to your mother’s expectations?
I was a bit of a rebellious kid and didn’t always believe my mother. When it came time to do things that moved me towards going to college, she reluctantly helped me. If I ended up getting into a good college, it wasn’t clear how we would pay. Then a kind of miracle occurred. I got a letter from the University of California, San Diego offering me a Regents Scholarship, which paid the entire cost of education, room and board and a stipend to boot. The same day the university called to say, “We want you to come down and see the university so that we can persuade you to come here, and we’ll pay for your trip down to the university.” This was an unimaginable thing to my parents. At that point they certainly realised that, if I wanted to go, I could.

When you started in education research it was described as having an ‘awful’ reputation. Why?
Within the field there was despair that some major intellectual thrusts seemed not to have been successful. A lot of research had been done but it seemed puzzling and contradictory. There didn’t seem to be any patterns in the findings. Distinguished people were saying that maybe systemic knowledge about education was impossible and we should abandon the quest to learn anything durable or transportable about education. I began to see that the problem wasn’t that education research was impossible, but that we didn’t know how to put it together in ways that would allow us to discover patterns. Things weren’t as contradictory as they seemed if we just had the right methodology to take on this task. Essentially, I was trying to develop mathematical statistical formulations for subjective ideas, using a systematic methodology and drawing on a set of technical principles that could be validated.

How did this so-called ‘voodoo statistics’ gain credibility?
I was very conscious that I was swimming upstream. I would talk to people about the work I was trying to do, and they would look at me as if I had three heads and quickly extricate themselves from the conversation. A dean said something like, “Larry my boy, you’re a smart kid. Why don’t you stop this meta-analysis stuff and do something useful? I feel certain you could if you just put your mind to it.” As often happens with new ideas, within a few years the tide had changed and the very same people began saying, “I always knew this was important.” Now it’s become the standard method.

Can meta-analysis tell us everything we need to know about education?
When I was young and naive I thought all we needed to do was rearrange the research that had already been done, and major new insights would come. I now understand that we haven’t done enough of the kind of research that is going to be needed to ultimately definitively show which kinds of education innovations, products and services are having a major impact on student achievement or other outcomes. Serious synthesis has helped to illuminate the shortcomings of some of the primary research.

How can these shortcomings be overcome?
There has been a great deal of interest in doing randomised experiments in education. Education is complicated. There are lots of confounding factors – differences in student background, teachers, schools, communities. Discovering the effects of interventions or products or services, with all those things going on, is just very difficult. In medicine we ultimately discovered that randomised controlled trials were the only way, in many cases, to understand whether a treatment worked or not. We have begun to realise that we have to do that in education too. It’s not the only thing we have to do, but it’s an essential step if we’re to discover things that really work – and to have confidence that they really work.

What would you most like to see changed in university education?
For a century, mathematics was part of what was thought of as a liberal education. Calculus was the first step into higher mathematics, and training people to know a little calculus was thought of as an important thing. These days, learning about statistics – how to manipulate data, and how to understand when people are trying to pull the wool over your eyes about data – is much more important for being an educated citizen. I would hope that universities would think about including that not as a specialty, but as part of a liberal education.

john.ross@timeshighereducation.com

Appointments

Paul Mazerolle has been chosen as the next vice-chancellor of the University of New Brunswick. The Canadian criminologist will return to his home province in July after almost two decades in Australia. He joined the University of Queensland in 2000, before moving to Griffith University in 2006 as a professor and director of its Violence Research and Prevention Program. In 2009, he became Griffith’s pro vice-chancellor of arts, education and law. Larry Hachey, chair of New Brunswick’s board, said Professor Mazerolle “possesses a strong track record in both academia and management, with proven experience leading growth-oriented change and restructuring in a multi-campus environment at Griffith”.

Rajkumar Roy has become the new dean of City, University of London’s School of Mathematics, Computer Science & Engineering. The Indian computer engineer, who has held research roles with Rolls-Royce and BAE Systems, joins City from Cranfield University, where he was director of manufacturing. Sir Paul Curran, City’s president, said Professor Roy would “bring a formidable combination of research excellence, successful enterprise collaboration and visionary higher education leadership to his new role”.

Emma Flynn has been appointed pro vice-chancellor for research and enterprise at Queen’s University Belfast. She will join the Northern Irish institution in May from Durham University, where she is associate provost and a professor in the School of Education.

Neil Brownsword has been appointed professor of ceramics at Staffordshire University. The artist and academic will join from Bucks New University, where he has worked for 23 years.

Vanessa Wilson, the current director of commercial and communications at UK Sport, is the new chief executive of the University Alliance mission group.

Tanya Monro, deputy vice-chancellor (research and innovation) at the University of South Australia, has been named as Australia’s next chief defence scientist.