Skip to main content Skip to secondary navigation

This Spring and Summer, find something that stands out.

Enroll in Spring IntroSems with Space Available now through the end of Week 1. Apply for Summer IntroSems starting March 8th!

Main content start

MED 73N: Scientific Method and Bias

Group of swans on a river. Nick Fewings on Unsplash.

Meet the Instructor | General Education Requirement

Course Description

This seminar examines theoretical considerations and practical examples where biases have led to erroneous conclusions, as well as scientific practices that can help identify, correct, or prevent such biases. We will also examine appropriate methods to interweave inductive and deductive approaches.

Over the past 50 years, remarkable advances in biomedical science and other scientific fields such as genomics have been steered by hypothesis-driven research. But there have been many setbacks and false conclusions because experiments were not properly designed or data were misinterpreted or improperly analyzed. For example, researchers observed the same19 e abnormal gene expression in a 50-year old obese man and some of his biological relatives who were also obese. They concluded that the abnormal gene expression causes obesity. This example demonstrates that inference cannot lead to valid conclusions about causation, as we do not know the social, environmental, or other biological reasons that may influence the family’s predisposition to obesity.

The seminar will cover the following topics: Popper’s falsification and Kuhn’s paradigm shift; revolution vs. evolution; determinism and uncertainty; probability, hypothesis testing, and Bayesian approaches; agnostic testing and big data; team science; peer review; replication; correlation and causation; bias in design, analysis, reporting, and sponsorship of research; bias in the public perception of science, mass media, and research; and bias in human history and everyday life.

At the end of the seminar, you will have an understanding of how scientific knowledge has been and will be generated; the causes of bias in experimental design and in analytical approaches; and the interactions between deductive and inductive approaches in the generation of knowledge. 

General Education Requirement

Meet the Instructor

John P.A. Ioannidis

John P.A. Ioannidis

"I have a long-standing interest in understanding how one can optimize the applications of the scientific method. Science is the best thing that can happen to humans, but performing scientific research can be like swimming in an ocean at night. Different disciplines within science use different methods and face different challenges, but biases and errors can be ubiquitous and the ability to minimize bias and error is instrumental in getting eventually reliable and useful knowledge. I have worked in different scientific fields, and I have made errors or fallen prey to biases probably in all of them. Much of my current research at the Meta-Research Innovation Center at Stanford (METRICS) is focused on mapping research practices, finding their weak and strong points, and making research practices more rigorous, efficient, and transparent."

Learn more about John P.A. Ioannidis

More to Explore

Prof. Ioannidis discusses large meta-analysis of antidepressants

More News Topics