Workshop: Improving Inference in Evolutionary Biology and Ecology-Charlottesville, Virginia; November 12-13, 2015

Whitman College, Walla Walla WA

Investigators

Abstract

A hallmark of effective science is reproducible results. If a result cannot be reproduced, it cannot be considered a robust scientific finding. Unfortunately, obstacles often slow progress towards these robust, reproducible observations in ecology and evolutionary biology as well as in many other disciplines. Scientists are rewarded for publicizing exciting results, but scientists have much less incentive to publish unexciting results or to test the reproducibility of previously published findings. Scientists therefore often search their data for the most exciting findings that will allow them to publish in the most prestigious journals, and they often do not invest in seeking to publish their full set of results. This and various related practices leads to an inflated rate of published false positives - results that are the outcome of chance rather than real biological phenomena. Because attempts to replicate prior findings are often rare, the error of these false positives often goes unrecognized for long periods. Even when replications exist, drawing conclusions from them is hindered by a lack of standards to promote effective syntheses across studies. Reducing bias and promoting replication and effective synthesis will require changing the institutions that control the incentives currently guiding scientists' decisions. One of the major classes of institutions shaping incentives is scientific journals. Journals in some disciplines, such as psychology and neuroscience, have begun introducing innovative editorial policies to reduce bias and facilitate replication. Now ecologists and evolutionary biologists are seeking to develop ideas appropriate for their own disciplines. This project is a workshop in which prominent journal editors in ecology and evolutionary biology will join with researchers interested in improving scientific inference to develop incentive structures in these disciplines that will reduce false positive rates, promote replication, and facilitate research synthesis. By the end of the workshop, the organizers hope to have drafts of editorial policy templates and other ideas for shaping incentive structures that make sense for these disciplines. The organizers also hope that journal editors and others will then implement policies that are right for their disciplines and journals, and that the discussions that follow in the wake of these initial policy implementations will promote more widespread adoptions of policies that improve empirical progress in their disciplines.