Scientists behaving badly – Martinson et al. - 2005 - Article

What is the issue?

We need to look beyond falsification, fabrication, and plagiarism, to a wider range of questionable research practices to protect scientific integrity. Misbehaviour in research damages the reputation of science and undermines public support of it. People have always focused on headline-grabbing cases of scientific misconduct, but the authors believe that researchers can not afford to ignore a wider range of questionable behaviour threatening the integrity of science anymore.

The Study

Surveying thousands of early- and mid-career scientists, based in the US and funded by the NIH, to report their own behaviours. Evidence shows a range of questionable practices. It suggests that ‘regular’ misbehaviours pose greater threats to science than those caused by high-profile misconduct cases like fraud.

Background

The US office of Science and Technology policy defined research misconduct as “fabrication, falsification, or plagiarism (FFP) in proposing, performing, reviewing, or reporting research”.

To understand misbehaviour requires attending to the negative aspects of the research environment. Modern scientists face a combination of pressures) intense competition, difficulty and unreasonable regulatory, social, and managerial demands) that creates many possibilities for the compromise scientific integrity, extending beyond FFP.

Getting Data

• Assuring anonymity, responses never linked to respondents’ identities.
• mid-career scientists: 1,768 usable surveys à 52% response rate.
• Early-career scientists: 1,479 usable surveys à 43% response rate.
• This approach leaves room for potential non-response bias; misbehaving scientists may be less likely to respond to the survey – fear of discovery or potential sanction.
  • Combined with potential underreporting of misbehaviours, this suggests that these estimates of misbehaviour are conservative.
• The specific behaviours chosen to be examined arose from six focus-group discussions held with 51 scientists from several research universities. They discussed which misbehaviours concerned them most.

Admitting to Misconduct

Survey respondents were asked to report whether they had engaged in a specified behaviour during the past three years. Summary of results:

• For six behaviours, reported frequencies under 2%, including falsification and plagiarism (consistent with previous estimates).
• Frequencies for remaining behaviours were 5% or above; most exceed 10%.
• Overall, 33% of respondents said they had engaged in at least one of the top ten behaviours in the last three years (38% among mid-career scientists and 28% in early-career group).

Possible explanations for sub-group differences:

• Opportunities to misbehave, and perceptions of the likelihood/consequences of being caught, may change during a scientist’s career.
• Groups received their education, training, and work experience in eras that had different behavioural standards.
• Underreporting of misbehaviours by those in relatively early-career positions.

Addressing the Problem

Results show US scientists to engage in various behaviours beyond FFP that can damage scientific integrity. Therefore, attempts to foster integrity that focus only on FFP miss a lot.

Early approaches to misconduct focused on ‘bad apples’. Consequently, analyses of misbehaviour focused on discussions of individual traits and local contexts as the most likely determinants. The 1992 academy reports helped shift attention from individuals with ‘bad traits’ towards general scientific integrity and the ‘responsible conduct of research’.

In the last decade, professional associations have focused on responsible conduct in research to promote research. But these efforts still prioritize the immediate contexts of scientists’ work, and are typically confined to ‘fixing’ the behaviour of individuals.

A consideration of the wider research environment is still missing from current analyses. The authors’ main concerns of this paper lie in the scientists’ perceptions of how the resource distribution processes work (peer-reviewing, features of funding and publishing, markets for research positions, graduate students, journal pages, and grants).

Now it’s time for the scientific community to consider which aspects of this environment are most important to research integrity, which are most amenable to change, and which changes are likely to be most effective in ensuring scientific integrity.