Since 2018, the National Association of Scholars (NAS) has been beating the drum about the irreproducibility crisis. That’s the failure of an enormous amount of modern scientific research to meet an elementary criterion for reliability: for the results of experiments to be reproduced by other scientists. The failure is rooted in politicization, groupthink, ambition, statistical error, and methodological sloppiness, and this has rendered a vast array of scientific research untrustworthy, false, or deceptive. There are easy solutions to the irreproducibility crisis. We’ve called for scientists to preregister their research, for example, as a way to increase reproducibility. That undermines one of the major contributors to irreproducibility, Hypothesize After the Research is Known, or HARK. Without preregistering, you vastly increase the chance that your “finding” is a false positive, or a statistical fluke. Because most scientific research is funded by taxpayer dollars, we’ve been calling for the federal government to require researchers to preregister their research if they want their research to get federal support or to inform federal policy.
New research just came out that makes the case that preregistration really does work. A Who’s Who of the scientists working to reform psychology (ground zero for irreproducible research) have just published a fascinating experiment. They preregistered a large amount of research studies and then worked to replicate them. Their preregistered research replicates about 86 percent of the time—a far higher proportion than the un-preregistered psychology research (about 50 percent). From the paper:
This paper reports an investigation by four coordinated laboratories of the prospective replicability of 16 novel experimental findings using rigour-enhancing practices: confirmatory tests, large sample sizes, preregistration and methodological transparency. In contrast to past systematic replication efforts that reported replication rates averaging 50%, replication attempts here produced the expected effects with significance testing (P < 0.05) in 86% of attempts, slightly exceeding the maximum expected replicability based on observed effect sizes and sample sizes. When one lab attempted to replicate an effect discovered by another lab, the effect size in the replications was 97% that in the original study. This high replication rate justifies confidence in rigour-enhancing methods to increase the replicability of new discoveries.
This research already has critics. It’s possible, the critics argue, that the reproducibility reformers committed some methodological errors of their own!
However, the fact that the researchers chose which of their studies to put forward for replication gives some pause to Berna Devezer, a metascientist at the University of Idaho (UIdaho) who was not involved in the work. The 16 findings used in this study were chosen very differently from past replication studies—they ‘are not randomly selected from, or representative of, a well-defined literature,’ she says.
The authors also checked whether their results looked different when they used other ways to define a successful replication. Some of those methods put the replication rate as low as 71%. But highlighting higher estimates and burying lower rates in the details of the paper is “disturbing and ironic,” says UIdaho metascientist Erkan Buzbas, ‘because some of the authors are ardent proponents of not cherry-picking results.’
Quibbles over statistics clouds the basic issue, however. Preregistration is a sign of good faith, the research equivalent of escrow. Why wouldn’t scientists voluntarily embrace preregistration? Do they need to be given a statistical probability of proof that it makes scientists’ research more reliable?
Sad to say, science suffers from the irreproducibility crisis in the first place because every professional incentive encourages scientists to publish exciting new results, regardless of whether they’re reproducible. Most scientists won’t want to reform—especially the ones foisting politicized agendas on the public rather than conducting disengaged research.
Only a federal requirement to preregister research will work—a requirement linked both to federal grant money and to qualifying research as eligible to inform federal policy. The federal government is the most important single funder of scientific research in the world. It has funded, rather, a gusher of irreproducible science: an abuse of taxpayer dollars. Scientists don’t preregister hypotheses because they are disincentivized from doing so. The federal government needs to change that landscape of incentives. It amounts to a fiduciary duty.
This new research also tells us that a federal requirement to preregister is doable. The means for voluntary, verifiable, and trustworthy preregistration already exists. If scientists can preregister effectively, it’s a short leap for federal regulators to require preregistration effectively. Metascience regulation ain’t rocket science.
Preregistration isn’t a silver bullet. Every system can be gamed, and requiring preregistration will encourage scientists to figure out ways to get around the requirement. But preregistration surely will help make science more trustworthy, an attribute in short supply lately, and the federal government can and should use its power to make preregistration reform effective.
Ruan J/peopleimages.com — Adobe Stock — Asset ID#: 602489238