CAS Workshop

Prediction, Registration, and Replication of Scientific Findings

June 1 - 2, 2021

Overview

The past decade has witnessed a rapid movement toward transparent and reproducible science. In many fields, researchers started to adopt policies and practices such as de-emphasizing statistical significance to discourage a dichotomous interpretation of the statistical evidence, requiring preregistration of research hypotheses and declaration of analysis plans, and launching a new type of publication called registered reports. What have we learned and achieved, and how should we move forward? This interdisciplinary workshop aims to bring together researchers from several disciplines, including economics, political science, psychology, neuroscience, and statistics. Topics discussed in the workshop include, but not limited to:

Poster

Logistics

Confirmed speakers

Schedule

Talks will take place over the course of two afternoons in Europe.
05:00 LA | 08:00 NY | 13:00 London | 14:00 Munich | 21:00 Tokyo

June 1 (Tuesday)
14:00 Taisuke Imai
Opening remarks
14:10 Felix Schönbrodt
Thoughts on quality monitoring in research - Who, how, and when
15:00 Tom Hardwicke
Preregistration: A pragmatic tool to reduce bias and calibrate confidence in scientific research
15:50 Break
16:10 Christopher Chambers
The past, present and future of Registered Reports
17:00 Break
17:10 Anna Dreber
Prediction markets on replication outcomes
18:00 Eva Vivalt
Uses of forecasts in research: Evidence from the Social Science Prediction Platform
18:50 Open discussion
June 2 (Wednesday)
14:00 Taisuke Imai
Day 2 kickoff
14:10 Macartan Humphreys
Reconciliation of preanalysis plans using design declaration
15:00 Sabine Hoffman
The multiplicity of analysis strategies jeopardizes replicability: Lessons learned across disciplines
15:50 Break
16:10 Maximilian Kasy
Pre-analysis plans and mechanism design
17:00 Abel Brodeur
Is peer review biased toward statistical significance?
17:50 Break
18:10 Gideon Nave
Two decades of Human Oxytocin Research: Evidence for low replicability and publication bias
19:00 Open discussion