Lecture 8 - Uri Simonsohn
One of the figures who popularized the term "p-hacking" , Professor Uri Simonsohn will explain why and how the p-hacking, a form of data dredging, achieving an artificially false significant result has transformed research and science.
Conference overview
P-hacking is the practice of conducting many analyses on the same dataset seeking to obtain a reportable, statistically significant, result (p < .05). P-hacking can lead researchers to believe in, and publish, findings that are false (i.e., not replicable), and it is likely the major reason why, in many fields (e.g., psychology, medicine), too many published findings do not replicate. The recognition that too many published findings do not replicate has led many different researchers to advocate for making many changes to the way we conduct and report our research.
In this talk, Pr.Simonsohn will discuss which of these proposed changes will effectively improve the way we do science and which will be ineffective or harmful. For example, he will conclude that pre-registration will do a lot of good, going Bayesian wouldn't do much of anything, and meta-analytic thinking would do a lot of harm.
The lecturer
Social psychologist, Uri Simonsohn is Professor in the Operations, Information & Decisions Department (OIDD) at the Wharton School of University of Pennsylvania.
Pr Simonsohn obtained his undergraduate degree in Economics from the “Universidad Católica de Chile”, and his PhD from the Social and Decision Sciences department at Carnegie Mellon University.
Pr Simonsohn has two main research streams at the intersection of psychology and economics, one focuses on how people think, make judgments and decisions, the other on how researchers do (methodology). He has published in many leading psychology, management, marketing, economic and business journals.
He is also associate editor of Management Science, reviewing editor for Science and the International Journal for Research in Marketing, and he is in the editorial board of Journal of Marketing Research and Journal of Behavioral Decision Making.
He teaches decision making process related courses to undergraduates, MBA and PhD students and how to motivate behavioral change, how to improve our intuitive and not-so-intuitive understanding of data based on psychological research.
Pr Simonsohn co-created the pre-registration website AsPredicted and co-hosts the DataColada blog.