Home » Meetings

In Anticipation of ICHPS … A Conversation with Long-Term Excellence Award Winner Don Hedeker

1 July 2019 No Comment

The Polkaholics at the May Fest in Lincoln Square in 2014. From left: Blitz Linster on bass and vocals, Action Jackson on drums and vocals, and Dandy Don Hedeker on guitar and vocals. Photo by Vera Gavrilovic

Throughout the past 20 years, the International Conference on Health Policy Statistics (ICHPS) has played a vital role in the dissemination of statistical methods in health policy and health services research. In preparation for the next conference, which will take place in January 2020, we’re running a series of interviews with previous Health Policy Statistics Section award winners.

The following is an interview between Juned Siddique and Don Hedeker, conducted in 2015 in honor of Don being awarded the Long-Term Excellence Award.

Juned: Your degree is in quantitative psychology from The University of Chicago. Could you tell us a little bit about the Department of Behavioral Sciences Committee on Research Methodology and Quantitative Psychology, which granted your degree?

Don: It was a small committee within psychology. Darrell Bock taught courses on multivariate statistics, on loglinear modeling, on item response theory. And Don Fiske taught courses on research methodology—designing studies, evaluation research—so I took courses like that. Steve Shevell taught a class on mathematical psychology, and also one on experimental design. But there weren’t that many, because it was a small group, so we were encouraged to take classes in statistics, in econometrics, in other areas. And so I took a lot of courses, actually, in statistics at that time.

Juned: Do you feel like it was a good choice to study quantitative psychology versus something more traditional like statistics or biostatistics, or even economics?

Don: You know, I’ve often thought about that. What was really good about quantitative psychology was that because it was a small committee, we took classes in other areas, whereas if I were strictly in a biostat program, I might not have done that. So, the benefit was I learned how sociologists applied statistics, how econometricians applied statistics, how statisticians thought about statistics, how psychologists did, and I got exposed to, in particular, latent variable models in psychology.

At that time—this was in the 1980s—there wasn’t that much exposure to latent variable models of any type in statistics, per se. But in psychology, you can’t measure IQ or personality with a blood test, right, so it’s all latent variables, really. And so is ability in taking an educational test; that’s a latent variable. So, I think that exposure really helped.

Juned: Music has been a big part of your life for a long time. Could you talk about your interest in music, how it began, and your current band, the Polkaholics?

Don: I was one of those kids who grew up in the ’60s, and I saw the Beatles on the Ed Sullivan Show when I was five or six years old and I thought, wow! It just blew my mind. It was like I want to get a guitar. As soon as I saw that, I wanted to sign up. And then when I was in high school, I started playing in bands. We played at sock hops and high-school events. I played in bands here at The University of Chicago. My biggest musical claim to fame is in spring of 1980, when I was a senior here at U of C. They had the Ramones playing at Ida Noyes Hall, and we got to open for the Ramones.

And how I started the Polkaholics was this. When I was in high school, I played in a band where we played a couple weddings. And if you play a wedding in Chicago, you have to play a polka song. I remember learning “Beer Barrel Polka,” and I was like, wow, this is kind of cool, actually. That was my impression back when I was in high school. And then it kind of stopped there.

And then, you know, in the ’90s, me and my wife, Vera, we went to thrift stores a lot to get outlandish clothes, and eventually my closet got completely filled. I couldn’t fit any more clothes, so I started looking at the records. And I saw these polka records and it just kind of intrigued me, so I started buying them.

You know, Iggy Pop, he once said he had to find his own blues, and I felt that way, too. I’ve got to find my own music. And so, to me, polka connected with my ethnic background and growing up in Chicago, and so I thought, wow, maybe we can play this music with our standard rock instrumentation. So, when we started, it was just like, can we even pull this off? And we did, and it was like, wow, this is so much fun. So, 18 years later, here it is, and we’re still going.

Juned: Returning to statistics, it seems like a lot of your ideas for statistical work come from your collaborations.

Don: Exactly. I think collaboration, for a biostatistician, is really important because, throughout my career, I’ve always tried to develop methods that can be used, and not some statistical thing that nobody’s going to use. So, when you deal with actual research in some substantive areas, you come to learn how they’re doing things and how you can improve upon that. Much of my statistical developments and software have been through collaboration and trying to improve upon the methods my colleagues are using, and trying also to make methods they can use.

Juned: You and Robert Gibbons wrote a book on longitudinal data [Longitudinal Data Analysis]. How did you decide to do that?

Don: When we first started it, there really wasn’t too much out there in terms of good books on longitudinal data. Now, of course, there are several very good, excellent books on longitudinal data in different areas.

Again, the idea is to disseminate and try to present the methodology in a way that’s not overwhelming technically. It’s a fine line you’ve got to walk. You don’t want to make it so simple that you lose meaning. So, trying to strike that right balance.

Juned: What does the future hold for you? Do you have any ongoing projects underway or new methodological areas?

Don: Yes, definitely. I have a grant from NHLBI [National Heart, Lung, and Blood Institute] with Genevieve Dunton from the University of Southern California to further develop the software we’ve been working on for mixed location-scale models. What I’ve developed is to allow multiple random effects in terms of the mean, so intercept and trend, things like that. You have multiple random effects there, and then to build a second model where the random effects from the first model influence a subject’s outcome. So that was the nature of the grant, and we’re working on methodology for that. And that’s going well.

The full interview is available as a PDF.

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)

Comments are closed.