Our 2021-2022 Distinguished Dissertation Award winner is Tony Mangino. Tony earned a PhD in Educational Psychology in July of 2021. Tony’s advisor, Dr. Jocelyn Bolin said, “From my perspective, Dr. Mangino’ s dissertation was an obvious choice for this award. His dissertation required understanding of complex statistical theory, practical application of statistical analysis, computer simulation coding and, most importantly, ability to translate the findings into something usable for the applied research community. All of these elements are very serious undertakings and Dr. Mangino navigated them masterfully. Not only was his dissertation complex, it was well executed. The writing was clear and well edited. Everything was presented in a logical manner. It was the only dissertation I have seen pass the defense with ZERO requested edits or changes from the dissertation committee in my entire 13 years at BSU. In sum, an incredible effort, an incredible product and exceptionally well deserving of this award. Dr. Mangino is exceptionally personable and excited to use his knowledge to help others. He is a wonderful and engaging teacher. He is an excellent statistical consultant and researcher. He is an enjoyable colleague and will do well wherever he goes in the future.”
While a graduate student at Ball State University, Dr. Mangino also won the 2020-2021 Doctoral level Excellence in Teaching Award.
Here is a Q&A with Tony
What made you select your research topic?
This topic was a logical extension from a couple of the other projects I’ve worked on with Holmes Finch and Jocelyn Bolin and served as the intersection of two analytical paradigms: predictive classification and multilevel modeling. Having a topic that was so closely aligned with my existing work made the jump to a dissertation-level project much more tractable, far less intimidating. Knowing just how detailed a dissertation should be, I needed to be cognizant of choosing a topic and a research question that was not monolithic (taking on too much is a very real possibility), but that still had practical applicability.
Please tell us about your research?
My dissertation study focused on whether we need to use multilevel classification models for prediction tasks just because we have a multilevel data structure. I wanted to understand whether these more complex models were necessary just because our data were more complex.
The motivating example I used was, from the PISA dataset, predicting whether students would be held back in school. The PISA dataset has information both on students themselves and on the schools that students attend, so the question became whether we need to account for school membership in making our predictions of student retention. The short answer is: not necessarily… but far more research needs to be done yet before we have anything resembling a universal answer.
By understanding how these different variables affect our model decisions, we can perhaps better inform research practices and simplify the already complex task of prediction. Similarly, if we know which factors to exclude from our models, we may be able to mitigate potential problems that arise from overly complex models. It is not sufficient just to understand complexity, it is even more important to understand where this complexity is situated, when to make something simpler.
In my work now, I focus more on the application of novel statistical methods to problems in both the biostatistics/epidemiology and psychology/education domains, among others. That’s one of the best things about being a team scientist: investigators from all different fields and with different types of research questions bring me interesting problems and I need to find an effective way of solving those problems. It ends up being a constant flow of new research ideas!
What brought you to Ball State University?
I happened to see the Master’s Program in Educational Psychology online when I was looking for an avenue to change my career to a more research-oriented direction back in 2016. It was after only a semester that I decided I wanted to pursue my PhD in Educational Psychology in the same department, particularly after seeing just how supportive my professors were during the Master’s program. Despite not having a distinct research area at the time, I knew that I wanted a career in research.
What are your future goals/career goals?
I’m currently working as a Biomedical Data Scientist in the Biostatistics Consulting and Interdisciplinary Research Collaboration Lab (Biostat CIRCL) at the University of Kentucky with a team of several other collaboration-focused biostatisticians. Much of my work in the Research Design Studio at Ball State prepared me to be an effective and flexible collaborative researcher, which made this position as a collaborative biostatistician a logical (if somewhat unexpected) step forward. I’m thoroughly enjoying my collaborations studying anything from surgical procedure outcomes to understanding longitudinal trajectories of opioid overdose rates to improving medical students’ educational experiences. Ultimately, I would like to start my own lab, but whether it has a focus on consulting or a focus on the intersection between computation and cognition (the inevitable tension between my loves of both psychology and statistical consulting is ever-present) remains to be seen. I do know that collaboration will be a continual theme.
Anything else you would like us to know?
Research, like teaching, should not be an isolating profession! As I wrote in the acknowledgment pages of my dissertation, the journey to finishing my doctoral experience and beginning my career as an independent and collaborative researcher was supported by a wide range of influences, to some of whom I was never quite able to articulate the full impact of their mentorship.
Thank you to Drs. Jocelyn Bolin, Holmes Finch, Lisa Rubenstein, Linda Martin, Jerrell Cassady, Matthew Stuve, Gerardo Ramirez, Serena Shim, Kathryn Fletcher, Robert Johnson (University of Northern Colorado), and Amy Eppolito (University of Southern Maine)!