‘Bad science’ and how new ideas in scientific publishing can combat it
In a position paper published in The Leadership Quarterly earlier this year, Professor John Antonakis described factors that he believes are stifling the production of useful scientific research. In response to this, Elsevier colleagues have highlighted developments in scientific publishing that may combat these trends.
Professor Antonakis identified issues such as an undue focus on statistical significance, rather than effect size, and on novelty at the expense of replication, extension or ‘negative’ studies. He also highlighted the practice of hypothesising after results are known (so-called HARKing), a lack of scientific rigour in theoretical or empirical work, and the penchant for quantity over quality when it comes to producing publications.
The team at Elsevier point to a number of potential solutions. These include journal policies that aim to improve transparency standards with regards to data, methods and reporting, and the move by some publishers to actively encourage the publication of replication studies. Another idea being piloted is the introduction of new article types, such as Registered Reports and Results-Masked Review articles, which are initially reviewed without the results and discussion, in a two-step peer review process. Here, papers may be accepted for publication in principle based on the experimental methods and proposed analyses alone, thus moving the focus towards the research question rather the outcome, and limiting potential publication bias. The authors also mention technologies that can be utilised to distinguish quality research, such as Mendeley Suggest.
The authors conclude by recommending that individuals make the best use of available technologies so that research can be judged appropriately on transparency and replicability.
Summary by Alice Wareham, PhD from Aspire Scientific
Leave a Reply