Uploaded November 2018 | Updated September 2026, 11 hours ago
The past few years has witnessed much debate regarding research practices that can potentially undermine the accuracy of reported research findings. A strong case can be made that, for example, that the Type I error rate is in fact much higher than the nominal a level of .05 because of study design and data analytic flexibility. Combined with the typically low levels of statistical power (~50%) in the published research, and the fact that over 90% of published findings are reported as statistically significant, a non-trivial number of published research findings are false-positives. But which ones? The most effective way to reliably distinguish true positives is to conduct high quality close replications of published research in order to derive more precise effect size estimates for presumed effects. In this presentation I will discuss how such a hands-on replication approach to teaching replication and open science practices is an excellent opportunity to teach how to do good science at an early stage in training new scientists. The goal is to make these teaching resources openly available to encourage widespread adoption as well as solicit feedback for improvement.
Lorne Campbell, Western University
Rotman 2018 Annual Conference: Understanding Replication Across the Sciences
October 14, 2018
Visit the Rotman website for more information on applications, events, project descriptions, and openings. rotman.uwo.ca
Follow The Rotman Institute on Twitter: twitter.com/rotmanphilo
Like The Rotman Institute on Facebook: facebook.com/rotmanphilosophy
Subscribe to our channel: youtube.com/user/rotmanphilosophy
The past few years has witnessed much debate regarding research practices that can potentially undermine the accuracy of reported research findings. A strong case can be made that, for example, that the Type I error rate is in fact much higher than the nominal a level of .05 because of study design and data analytic flexibility. Combined with the typically low levels of statistical power (~50%) in the published research, and the fact that over 90% of published findings are reported as statistically significant, a non-trivial number of published research findings are false-positives. But which ones? The most effective way to reliably distinguish true positives is to conduct high quality close replications of published research in order to derive more precise effect size estimates for presumed effects. In this presentation I will discuss how such a hands-on replication approach to teaching replication and open science practices is an excellent opportunity to teach how to do good science at an early stage in training new scientists. The goal is to make these teaching resources openly available to encourage widespread adoption as well as solicit feedback for improvement.
Lorne Campbell, Western University
Rotman 2018 Annual Conference: Understanding Replication Across the Sciences
October 14, 2018
Visit the Rotman website for more information on applications, events, project descriptions, and openings. rotman.uwo.ca
Follow The Rotman Institute on Twitter: twitter.com/rotmanphilo
Like The Rotman Institute on Facebook: facebook.com/rotmanphilosophy
Subscribe to our channel: youtube.com/user/rotmanphilosophy
