I have a long article in Slate (with Chris Chabris) on the importance of replicating science. We use a recent (and especially bitter) dispute over the failure to replicate a social psychology experiment as an occasion for discussing several things of much broader import, including:
- The facts that replication, despite being a cornerstone of the scientific method, is rarely practiced (and even less frequently published) not only in psychology but across science, and that when such studies are conducted, they frequently fail to replicate the original findings (let this be a warning to those of you who, like me, cite empirical literature in your scholarship);
- Why replications are so rarely conducted and published, relative to their importance (tl;dr: it’s the incentives, stupid);
- Why it’s critical that this aspect of the academic research culture change (because academic science doesn’t only affect academic scientists; the rest of us have a stake in science, too, including those who fund it, those who help researchers produce it (i.e., human subjects), those who consume and build on it (other scholars and policy-makers), and all of us who are subject to myriad laws and policies informed by it); and
- Some better and worse ways of facilitating that cultural change (among other things, we disagree with Daniel Kahneman’s most recent proposal for conducting replications).