Posted: November 20, 2017

Using a Survey of Graduate Student Skills to Guide a Novel Course in Experimental Design and Analysis


Until 3 years ago, graduate students in our PhD programs were required to take a course in biostatistics at the end of their first year, as they transitioned from rotations to dissertation labs. Unfortunately, most students disliked the course and reported that it was irrelevant to their scientific training – most examples were taken from problems in public health and differed from the research done by our students. In response, we took a different approach by designing a new course entitled, D2K:from data to knowledge – Biomedical experimental design and analysis. The course has now been taught 3 times to a diverse class of around 36 students per year drawn from 8 programs. Our survey found that most students (85%) have prior formal training in statistics, yet have difficulty applying concepts, even after one year of laboratory experience in graduate school.


Data Analysis, Reproducibility, Rigor, Experimental Design, Statistics


John P. Horn, PhD, University of Pittsburgh School of Medicine

Deepti Ramadoss, PhD, University of Pittsburgh School of Medicine

Richard A. Bilonick, PhD, University of Pittsburgh School of Medicine