Many studies over the past few years have identified problems with reproducibility in scientific research, even to the point of calling it a “reproducibility crisis.” Reproducibility is a key tenet of scientific discovery. The ability for researchers to replicate the findings of a particular study gives it credibility that not only improves scientific literature, but also speeds scientific discovery. As more reproducibility problems across all scientific disciplines come to light, there is a tendency for the scientific community to throw up its hands and declare scientific research as unreliable.
Although reproducibility, or lack thereof, is a concern, it’s important to discover the underlying reasons researchers cannot reproduce the findings. Is it because of bad science, faulty conclusions, or poor methodologies for either the original or subsequent studies? Often, the inability to replicate a study can be as important as the original study. When these findings are published, they provide an opportunity for delving deeper — and that’s good for both the scientific community and the general public.
The importance of reproducibility
Reproducibility allows the production of trustworthy science by reducing both fraudulent research as well as honest mistakes. Reproducibility not only makes writing and publishing papers easier and more efficient, it also helps peer reviewers better evaluate the research.
If studies can be replicated and the same results are reproduced, it lends credibility to the research. When results are not reproduced, the earlier research may be perceived as questionable. Essentially, it’s a way for science to self-correct.
The underlying challenges and reasons for reproducibility problems
There are many reasons and factors involved when a study fails to reproduce earlier experimental results. For example, there may be problems with the original or subsequent replication studies, such as a too small sample size or unforeseen circumstances like temperature changes, timing errors, or power outages. Conflicts of interest or manipulation of data — whether conscious or subconscious — to produce a desired P-value (probability value) can hinder reproducibility, as can a lack of data sharing between researchers conducting replication studies.
In some cases, it’s not that the research is not reproducible, it’s that there is a lack of replication studies. Many researchers prefer to conduct unique or novel studies with significant new findings. They also may have the perception that academic journals are not interested in publishing replication studies.
Overcoming challenges and using reproducibility problems to further science
Whether a reproducibility crisis exists or not, both researchers and publishers can work together to facilitate reproducibility and protect the integrity and validity of scientific research. Researchers not only need to be meticulous in their methodologies and data analysis, they should assist those trying to reproduce the results by being transparent about their methods and any uncertainty inherent in the results. They also need to limit bias by being careful about funding sources.
It’s important to understand that a failure to reproduce experimental results is not necessarily a bad thing. It is a way for science to self-correct. It also can lead to the discovery of new variables and interactions, furthering scientific discovery.
Replication studies are important scientific endeavors and should be published whether they agree with the original study or not. Researchers should register an intent to publish, ensuring results will be shared, regardless of the outcome.
As publishers, we need to be open to publishing replication studies and provide a means for making results available for further studies. Access to failed results is as important as corroborative results for the purposes of advancing science. You can find other ways to support reproducibility in ‘’A manifesto for reproducible science.’’
Reproducibility is a cornerstone of good science. Corroborative results provide validation for earlier research, but when replication studies fail to produce the same results, it is an indication to take another look. Both the ability and inability to reproduce results are important for the advancement of scientific research. To that end, publishers can contribute to the transparency and efficiency of scientific discovery by providing an outlet and incentive for researchers to make their research available, whether it supports the hypothesis behind it or not.
Contact your Sheridan representative for a consultation or visit our contact page to learn how Sheridan experts can help streamline and simplify your publishing processes.