When looking to publish data, reports should be written in a way that is sensitive to the participant group involved. Language should be unbiased, not allowing gender or racial stereotypes without qualification. Labels should be carefully applied, such as in linking different ethnic groups together. The authorship of studies is also an ethical issue, and should not merely be attributed as personal favours or in the hope of eliciting authorship on other studies in return. The Vancouver Protocol was developed by a group of medical journal editors in Vancouver, British Columbia in 1978 to set out uniform requirements for manuscripts submitted for journal publication, including setting out guidelines for who should be named as an author on research. According to the protocol, authorship should be based only on substantial contributions to:
- 1. Conception and design OR analysis and interpretation of data
- 2. AND drafting the article or revizing it critically for important intellectual content
- 3. AND on final approval of the version to be published
To be named as authors, those involved must satisfy all three conditions. Notably, involvement in acquisition of funding or collection of data on its own does not justify authorship, nor does general supervision of the research group undertaking the research. People who do not fulfil the three criteria but who have contributed to the project should instead be ‘acknowledged’, which is often through a separate section on a research paper.
Another ethical issue pertains to which data are reported and how. Problems that can arise within science include the falsifying of data, suppressing negative or insignificant results, and even inventing data to support specific conclusions. At best, these practices lead to wasted money and ineffectual treatments. At worst, they can lead to participants being harmed. Given the large amounts of data that are collected in research studies, it is common for not all of them to be reported. However, certain principles can guide whether reporting is ethical or not. Suggestions for these include:
- ? Demographic data should be provided with enough detail that it is clear who is represented by the study results. If participants in a study are all musically trained, this is an important point for inclusion in a study about a music intervention as it could imply that the intervention only has effects on people with musical backgrounds. However, other pieces of demographic data that may have been collected, such as what city people live in, may not be as important if this is not part of the study hypotheses and if the defining features of their city living (such as socioeconomic status) are already being reported
- ? Non-results should be acknowledged. If there are some findings alongside other null findings, it is encouraged that both types are reported. If the entire study finds no significant results, it can be a challenge to find a journal that will publish it. However, there are now a wide range of science and arts journals that publish results from well-designed studies, regardless of the outcome. And it is also possible to publish study reports on project websites so that people can be aware of null findings
- ? Limitations should be outlined. All studies have limitations, whether a weakness in the study design, a limited number of participants, or questions that could not be answered yet are still relevant to the topic. These should be honestly stated so that readers can judge for themselves how much store to place on findings and can develop follow-up research
The topic of publication in research was made famous by doctor Ben Goldacre, who for more than a decade wrote a column in the Times entitled ‘Bad Science’, as well as producing best-selling books Bad Science (2009) and Bad Pharma (2013).(7,8)
Finally, when publishing results, the opportunity for repercussions based on these results should be considered: comprehensive limitations sections can help to reduce sweeping generalizations; press releases that accompany results should be carefully worded to ensure they do not mispresent results to a nonscientific reader; and if there are contentious issues brought to light by data, these should be discussed within the ‘discussion’ section of a research paper, in particular with any caveats to interpreting the data carefully outlined.