A recently published Science article reported that efforts to reproduce findings from 100 studies published in three leading psychology journals concluded that more than 60% of the findings did not hold up when retested (Open Science Collation, 2015). While the Reproducibility Project found no evidence of fraud, it concluded that the evidence for most published findings was not nearly as strong as originally claimed. Some say the field of psychology is in crisis, while others say the failure to replicate is a normal part of how science works.
The editorial team of The Gerontologist sees these findings as a call to strengthen our journal. Like other top-tier journals, we are guilty of favoring novel studies over replication studies. We agree, however, with the Science article’s conclusion that “Innovation points out paths that are possible; replication points out paths that are likely; progress relies on both” (p. 943). To that end, we will strive for a better balance. We encourage scholars to submit high quality replication studies to our journal. The replication studies we seek are well-powered, based in theory, and have implications for bettering the lives of older people. The original studies they replicate are well-cited and important to the field. The introductions of these studies make clear the need for replication; their discussions explain why and how findings from the replication help build science. We believe that replications that reproduce results can bolster the knowledge base while replications that do not support the status quo can promote innovation.
Earlier this year, The Gerontologist called for literature reviews and meta-analyses, which we publish online. These reviews critically examine issues and draw conclusions based on many studies. Like replication studies, these reviews strengthen our confidence in scientific results by synthesizing and crystalizing complex findings, fitting them within theories of human behavior, identifying unanswered questions, and pointing toward applications that can improve the experience of aging.
The Reproducibility Project raises important issues for journals. The Gerontologist, like most journals, tends to favor studies that support hypotheses over studies that fail to do so. We suspect that important findings languish in file drawers because studies that do not support hypotheses are systematically under-published. Here too, we will endeavor to present a more unbiased perspective. Publishing such findings is tricky, because studies fail to support hypotheses for many reasons. Some studies do not support the hypothesis because they have poor designs or are inadequately powered. These studies are not of interest to The Gerontologist. The studies we seek are well-powered, based in theory, and have important applied implications. Their introductions explain the current state of the art; their discussions grapple with why hypotheses were not supported and explain how future research should build on these findings.
The Reproducibility Project found that the best predictor of the ability to reproduce a finding was the strength of the original effect. The most robust findings were easily detectable. The Gerontologist has long insisted on statistical rigor and well-powered studies and will continue to do so.
The Reproducibility Project examined studies published in psychology journals. But concerns about the state of science are not limited to psychology. As a multidisciplinary journal, The Gerontologist embraces the challenge presented by these findings and seeks to grow stronger from them.