Are Psychological Scientists Overvaluing Significance? 

Promoting nonsignificance • New solutions for an age-old problem

Publish or perish. This phrase is often cited by scientists as their motivation to submit a continuous flow of papers to academic journals. But how valuable is research if the findings are not significant? And what effect will those papers have on the reputations of the scientists who author them? 

Though studies with nonsignificant findings are often welcomed in psychological journals, scientists tend to believe they are discouraged and are hesitant to submit them. Over time, a bias toward significance has powerful impacts on the literature—it can inflate misleading associations between variables, undermine meta-analyses, and contribute to replication failures. For applied sciences like clinical psychology, those biases can move downstream, impacting treatment recommendations. 

“Cumulative science requires that all results, including null findings, are published,” said Louis Schiekiera, a research associate at Humboldt Universität zu Berlin. “Otherwise, the literature becomes biased, replication problems emerge, and resources are wasted.” 

Schiekiera and his coauthors examined the publication biases displayed by psychological scientists in their recent paper in Advances in Methods and Practices in Psychological Science. 

In an online experiment with clinical psychology scientists, researchers asked participants to quickly evaluate abstracts for their likelihood of being submitted, read, or cited. They followed up shortly after this initial assessment to see how participants felt about their original responses.  

Participants rated abstracts with nonsignificant results as less likely to be submitted, read, or cited, and the researchers did not change their responses after further consideration. Although nonsignificant results were penalized in their assessments, participants did not show the same bias for abstracts that included inconsistent hypotheses (Schiekiera et al., 2025). 

“We had expected hypothesis-inconsistency to matter more, given past debates about confirmation bias and its relationship to publication bias,” Schiekiera said.  

Previous research has shown that scientists submit fewer nonsignificant results because they believe articles will not be accepted by journals but, in reality, journal editors are just as likely to accept them as those with significant results (Lee et al., 2006; Okike et al., 2008; Olson et al., 2002; Timmer et al., 2002). Scientists also worry about their reputations when they are cited for nonsignificant findings (Schiekiera et al., 2025).  

“Academic careers, tenure decisions, and funding are strongly tied to publishing in high-impact journals, which often tend to reward novelty and positive results over replication or null findings,” Schiekiera said. “This creates a perception that negative or nonsignificant results ‘don’t count’ for career advancement.”  

Promoting nonsignificance  

Still, some scientists have found continued success from the practice. Matt Williams, associate professor at Massey University, is no stranger to submitting nonsignificant results to journals, yet he rarely encounters resistance when publishing nonsignificant findings.  

Williams is a coauthor on a recently published article titled, “Do Stress, Depression, and Anxiety Lead to Beliefs in Conspiracy Theories?” in Clinical Psychological Science. The study examined the relationship between conspiracy theories and mental health in nearly 1,000 participants from Australia, New Zealand, and the United Kingdom.  

Only one of their 15 hypotheses was statistically significant (and that one was just on the cusp of significance), implying little evidence that psychological distress leads to a belief in conspiracy theories or vice versa (Fox et al., 2025).  

“This doesn’t completely rule out the theory that psychological distress can increase vulnerability to conspiracy theories, but it does suggest we should be quite skeptical of that idea,” Williams said. 

Williams had a positive experience with his submission, saying he found the peer reviews to be insightful and constructive, focusing on the methods and theory of the paper. 

“I have reported a lot of nonsignificant findings, and I’ve never really felt that it’s harmed my career,” he said. Williams added that, in contrast, he has been encouraged to submit Registered Reports and pursue other ways of reducing the bias toward publishing only significant work.  

Though authors must take the first step to report nonsignificant findings, Williams believes editors and reviewers hold the most power in normalizing their publication. 

Related: AMPPS Accepts First Manuscript From Innovative Peer Review System 

“That’s partly why I do a lot of editing and peer reviewing,” he said. “When I’m editing and reviewing, I try to judge studies based on their rigor and whether the research questions seem useful to answer.” 

APS Fellow Jennifer Tackett is the editor-in-chief of Clinical Psychological Science, the APS journal that published Williams’ study. Tackett agrees that it is crucial for researchers to contribute to the peer-review system by serving as reviewers and editors as they are able.  

“We need more than authors to be on board if we hope to see long-lasting systemic change,” she said. 

New solutions for an age-old problem 

This bias toward significance is not a new phenomenon. It was described as the “file drawer problem” nearly half a century ago (Rosenthal, 1979). But in today’s publishing landscape, cultural shifts and emerging practices serve to provide potential solutions for an age-old problem.  

“Ideally, criteria for publication should not be any different for significant and nonsignificant results,” Tackett said. “This is why models like Registered Reports are so critical. Papers should be evaluated based on whether they ask important, scientifically grounded questions and use robust and rigorous methods and analyses.”  

With Registered Reports, authors submit their manuscript before they collect data. The primary review process for the manuscript happens at this preliminary stage and, if the manuscript is in line with publishing requirements, authors receive in-principle acceptance. This means that the entire study has already been accepted for publication regardless of the results.  

However, despite this tool’s power for ensuring that research is considered regardless of its statistical significance, Tackett says clinical psychologists have been slow to adopt the approach.  

“I would love to see more folks in the field try to produce a Registered Report. It’s an eye-opening experience, and I think people would quickly realize the value of that approach,” she said. “But it has been very difficult to understand why there remains so much resistance to that model.” 

Schiekiera also sees Registered Reports as a potential solution, alongside other shifts in the culture of psychological science, such as creating dedicated venues for nonsignificant results and improving the ways studies are indexed to make it easier to discover nonsignificant findings.  

He insists that the onus to make these shifts is not centralized to any one party.  

“Realistically, responsibility is distributed,” Schiekiera said. “Authors through their submission and citation choices, reviewers through the signals they reward, and journals through both explicit and implicit policies—all contribute to shaping the current incentive structure.” 

Feedback on this article? Email [email protected] or login to comment.