Survey finds a quarter of researchers see an integrity problem

A new survey finds problem particularly bad in biomedical sciences, where there are major issues with planning, conducting and reporting of research.

A quarter of researchers see a lack of integrity in their scientific field.

Over a quarter of Australian scientists and research managers believe there is a lack of research integrity in their field that compromises the ability to arrive at reliable and reproducible scientific findings.

The survey of nearly 1000 research staff, carried out by the Australian Academy of Science and Springer Nature, found the problem is particularly bad in the biomedical sciences where 43 per cent believe there is a problem with lack of integrity in the planning, conducting and reporting of research.

Researchers in the physical sciences saw the least problems, with only 11 per cent seeing integrity failures. Overall, 26 per cent believed lack of integrity was a problem.

One survey respondent, speaking of their research institute, replied with these personal observations: “Many studies that are done at the institute lack transparency and appropriate controls. Researchers at senior positions are not good models for research integrity (prevalent of ghost/gift authorships and not enough quality control of junior members).”

Asked what issues were important in assessing research integrity over 90 per cent of respondents agreed with “acknowledging the work of others”, “reporting research transparently”, detailing research methods and procedures” and “declaring conflicts of interest”.

Other aspects of research integrity seen as important by over two-thirds of respondents included “research project design”, “sharing negative results”, “making protocols openly available” and “statistical methods”.

Over two-thirds (68 per cent) of survey respondents said that their institution provided training on research integrity, and 50 per cent said such training was mandatory. However far more, 73 per cent, said they believed such training should be mandatory.

Most (61 per cent) agreed that the research integrity training offered by their institution was effective, while 14 per cent disagreed. Of those who believed it was ineffective 53 per cent said it lacked depth and detail, and 18 per cent said there was no accountability for non-compliance.

On survey participant said their group of higher degree by research students asked for training seminars on research integrity and were told there was no money for it.

“Given the entire ‘research integrity’ portion of the induction is under 15 minutes – once in a PhD (or masters) – and does not cover the university ethics policy, data storage or availability, and does not touch on evolving problems like the nature of authorship, republishing of data, self-plagiarism, contributing to fair data principles, I would generally like more comprehensive and more regular training,” they said.

The Australian Academy of Science’s secretary for science policy Ian Chubb said it was vital that research was conducted ethically and with integrity, so that there was confidence in its outcomes.

“Greater consistency in the provision of research integrity training along with a greater uptake among those working in the sector, particularly among early career researchers, is one way to meet that objective,” said Professor Chubb, who is a former chief scientist of Australia.

The survey also revealed that over a quarter of respondents were not satisfied with the quality of mentorship of early career researchers received in matters of research integrity at their institution. Twenty six per cent disagreed with the statement: “The quality of mentorship in relation to research integrity by senior researchers at my institution is high”.

Researchers from 35 universities responded to the survey between December 2020 and November 2021.