In Dan Ariely's Ted Talk, we are introduced to a phenomenon where, despite potential consequences or incentives, people lie. This isn't anything particularly novel-- everyone knows they're surrounded by liars, because everyone lies themselves. Despite this universal knowledge, the majority of people exist in a blissful world where science and scientific publishing is exempt from the “Fudge Factor”, and it's difficult to understand why.
Let’s examine for a second the tiers involved in scientific publishing—we have the scientist and collaborators who design an experiment based on observed and researched data, perform the experiments to the best of their ability, reproduce the experiments, and compile their data for publication. We have the publishing companies that receive and distribute the manuscript to editors and peer reviewers. The peer reviewers that diligently read the manuscript for credibility and relevance, and ultimately return the manuscripts back to their editors for revisions. The last player in the publishing of a manuscript is the reader—the scientist, media, or interested party that consumes the published manuscript.
With all that in mind, I see two glaring reasons why scientific publishing lends itself to deception. One—the demand for new manuscripts, new ideas, and even new “miracles” is so high. Take the findings by JAMA regarding the nomenclature of press releases on new oncotherapies. When every new study is hailed as a potentially lifesaving new breakthrough in the field, it is no wonder that reproducibility in the field of cancer research and others is so dismal. The second problem I see, is the aforementioned structure of scientific publishing itself. As it exists today, the system is made so that faulty methods and non-novel findings do not slip through the cracks into publication. The problem is, with so many people involved in publishing a manuscript, all it takes is for one tier of the publishing process to succumb to the "Fudge Factor"-- lying just a little bit, for the integrity of the process to fail.
I don't see any easy solution to the flaws that exist in scientific publishing. The one thing that I do believe contributes to the integrity of the scientist, the publishers, and the public, that was not mentioned in any of the recent articles on unreliable research is skepticism. A scientist should be skeptical of their research instead of trying to shape their data to fit their hypotheses. Publishing companies should accept openness and discussion in the post review process. Perhaps most importantly when viewing skepticism, the public should maintain a healthy amount of skepticism and scrutiny when reading new publications and press releases.
It is easy to get swept into the glamour of a promising new result, whether on the bench or over breakfast while watching the news one morning. The job of responsible science lies on all parties to maintain skepticism while producing and consuming science, and hopefully, this scrutiny will lead to a more honest system of scientific publishing.