A 2018 anonymous survey found that many responding biomedical researchers in Belgium said they had committed or witnessed at least one of 22 listed forms of research misconduct or questionable practice in the previous three years. It did not establish the share of all researchers who committed proven misconduct: its percentages describe usable survey respondents’ self-reports and observations.
What the survey found
In its February 2, 2018 report, Chemistry World said 71% of university respondents and 61% of industry respondents admitted at least one listed act during the prior three years. On witnessing conduct, 93% of university respondents and 84% of industry respondents said they had observed misconduct among colleagues.
As an Amazon Associate I earn from qualifying purchases.
Those figures measure two different things: respondents’ admissions about their own behavior and their reports of what they saw others do. Observations were not adjudicated findings. As study author Simon Godecharle put it, “This survey, based on anonymous self-report, shows that research misconduct occurs to a substantial degree among biomedical researchers from both industry and universities.” Chemistry World’s report presents the headline comparisons.
Recommended Free Tools
Who was surveyed, and how
The researchers invited 1,766 university researchers who had received biomedical research grants or scholarships from Belgian national academic funders between 2010 and 2014. They also invited an estimated 255 researchers at large biomedical companies or spin-offs in Belgium. An anonymous online questionnaire asked whether participants had committed or observed listed behaviors in the preceding three years.
#1 Best Overall
| Group | Invited | Usable responses | Response rate |
|---|---|---|---|
| University researchers | 1,766 | 617 | 43% |
| Industry researchers | Estimated 255 | 100 | 48% |
The response counts and rates come from the study record; the percentages reported in the article refer to respondents, not everyone invited. The publication is Simon Godecharle, Steffen Fieuws, Ben Nemery and Kris Dierickx, “Scientists Still Behaving Badly? A Survey Within Industry and Universities,” Science and Engineering Ethics 24(6), 1697–1717 (2018). The journal record and abstract describe the study.
What counted as misconduct in this survey?
The questionnaire covered 22 kinds of conduct, extending beyond fabrication, falsification and plagiarism. Examples reported by Chemistry World include cutting corners, ignoring safety or regulatory requirements, and changing study results in response to pressure from funders. The figures therefore should not be read as rates of confirmed fraud or plagiarism alone.
Godecharle told Retraction Watch that the team chose a broader scope because research misconduct and questionable practices in everyday work go beyond the traditional categories of fabrication, falsification and plagiarism. The survey adapted a prior US questionnaire in light of European guidance and the team’s earlier qualitative work. The interview with Godecharle discusses that approach.
Does industry report less misconduct than academia?
In this survey, industry respondents less often admitted at least one listed behavior and less often said they had observed misconduct than university respondents. But the study does not show that the underlying rate of misconduct was lower in industry. Differences in willingness to disclose, recall, or interpret behavior could affect the comparison, and the available findings do not establish why the groups differed.
Rank #3
For individual behaviors, the report described few significant differences between the groups. Plagiarism was identified as an exception: it was admitted and observed more often by industry respondents than by university respondents. The headline percentages cannot be used to infer item-level rates.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the percentages can—and cannot—tell you
- They describe respondents. The 71% and 61% figures are proportions of usable university and industry respondents who acknowledged at least one item, not proportions of all academics or industry scientists.
- They combine a wide range of conduct. One respondent acknowledging any item counts toward the headline measure, whether the behavior was a serious form of misconduct or a questionable practice on the broader list.
- They rely on self-report. Anonymity may make disclosure easier, but answers still depend on what participants recalled and chose to report. The witnessing figures are reports by respondents, not formal investigations or verified cases.
- They cover a limited population. The sample concerned biomedical researchers in Belgium linked to specified academic funders and large biomedical companies or spin-offs. It cannot establish rates for other countries, fields, career groups, or smaller organizations.
- Nonresponse matters. The response rates were 43% among university invitees and 48% among industry invitees; the behavior of people who did not respond is unknown.
The most defensible conclusion is that substantial numbers of respondents in both settings disclosed or witnessed conduct on a broad survey list. The study provides evidence about those reports, not a census of misconduct or a causal explanation for differences between workplaces.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
