Presenting NPI Research at McCahan Education Day
Presenting our near-peer instruction research at Boston University’s annual medical education conference — and what it felt like to stand next to your own work and explain it to the people who were there when it happened.
“The purpose of research is to inform action.”
- McCahan Education Day
- Standing Next to the Work
- The Conversations That Followed
- Why Education Research Matters
- Read the Full Paper
- Definitions
McCahan Education Day
McCahan Education Day is Boston University Chobanian & Avedisian School of Medicine’s annual celebration of medical education — a day dedicated to faculty, staff, and students who are thinking carefully about how medicine is taught and learned. Posters, presentations, and workshops span everything from curriculum design to clinical skills assessment to educational technology.
It’s a fitting venue for work about near-peer instruction. The questions at the center of that research — does structured peer teaching actually help students? does program design matter? how do you evaluate a voluntary program without fooling yourself? — are questions that medical educators across the country are still working through. McCahan Education Day brings together exactly the people who care about getting those answers right.
The keynote speaker opening the day’s sessions on innovation in medical education.
Standing Next to the Work
There’s a particular dynamic to poster presentations that you don’t fully appreciate until you’re doing one. You’ve written the paper. You know every methodological decision, every limitation, every time you ran a model three different ways to make sure the result wasn’t an artifact. And then you stand in front of a printed version of it and wait for people to walk up and ask you to explain it in two minutes.
The NPI poster summarizes the core finding: apparent negative effects of voluntary tutoring programs are a selection artifact, not evidence of genuine harm. Students who seek tutoring are already struggling before they show up — and naive comparisons between them and students who didn’t seek help capture that pre-existing gap, not the program’s effect. When you apply progressively rigorous causal controls, the apparent harm attenuates to null, and structured TA sessions trend consistently positive.
Walking through the methodology and findings with conference attendees.
That finding is more actionable than it might look. A lot of voluntary academic support programs get evaluated with naive comparisons, show null or negative associations, and either get cut or get quietly de-prioritized. Some of those programs are probably doing exactly what they’re supposed to do — they’re just being evaluated with the wrong method.
The Conversations That Followed
The best part of presenting research in a room full of medical educators is that the questions are good. People who design and run academic support programs have a practical interest in this question, and several of the conversations at the poster session pushed into territory the paper doesn’t fully address — program fidelity, tutor training, how you measure the quality of an individual session rather than just attendance.
A few people asked specifically about the structured TA format: the 30/30/30 design, the faculty review process, the competitive hiring. There’s genuine interest in whether those structural features are what differentiated the outcomes, and what a well-designed replication study would need to look like to answer that cleanly.
Those are exactly the right questions. The current study shows the selection artifact clearly and shows that TA sessions trend positive under rigorous controls, but it can’t pin down which specific design features are driving that signal. Future prospective work with larger samples, pre-course measures (GPA, test anxiety, study habits), and fidelity tracking would get closer to the mechanism.
The poster session in full swing — a good mix of faculty, students, and staff across the school.
Why Education Research Matters
I came into this project as a TA and tutor coordinator trying to understand whether the program I was running was actually helping. What I came out of it with is a much clearer sense of how easy it is to misread the answer — and how much the method you choose determines the story you tell.
That matters beyond near-peer instruction. Voluntary programs in medical education — tutoring, mentorship, wellness initiatives, supplemental review — are almost always evaluated under conditions that favor selection bias. The students who need them most are the ones who use them. If you evaluate those programs with naive comparisons, you will systematically conclude they don’t work, even when they do.
Medical education has the tools to do better. McCahan Education Day is full of people who want to. Presenting this work here felt like putting it into the hands of exactly the audience that can actually use it.
Read the Full Paper
The full NPI paper is published in Physiology (American Physiological Society). Find it on my Research page.
Definitions
- McCahan Education Day
- An annual conference at Boston University Chobanian & Avedisian School of Medicine celebrating research and innovation in medical education, featuring poster sessions, presentations, and workshops from faculty, staff, and students.
- Confounding by Indication
- A statistical bias that arises when the condition prompting treatment (e.g., academic struggle leading to tutoring) independently affects the outcome (exam performance), causing programs designed to help struggling students to appear ineffective in unadjusted comparisons.
- Poster Session
- A conference format in which researchers display their work on printed boards and engage directly with attendees who stop to discuss the research.
