scientific fraud and misconduct · 2019. 11. 19. · scientific fraud and misconduct ... aws s....
Post on 19-Dec-2020
5 Views
Preview:
TRANSCRIPT
SCIENTIFIC FRAUD AND
MISCONDUCT
https://www.pewresearch.org/science/2019/08/02/trust-and-mistrust-in-americans-views-of-scientific-experts/
Ernst Haeckel Jan Hendrik Schön
On track for Nobel prizeOrganic conductors
Bell Labs
WikipediaList of scientific misconduct incidents
Malcolm Pearce
a small case series with no controls, linked three common conditions, and relied on parental recall and beliefs
MMR linked to autism
Andrew Wakefield
ontogeny recapitulates phylogeny
Brian and Allan Kaspar were the chief scientific officer and vice president at AveXis (gene therapy division of Novartis)
May 24 2019, FDA approve Zolgensma, a gene therapy product used to treat children (<2 yrs) with spinal muscular atrophy (SMA).
June 28 2019, FDA informed by AveXis Inc. - data manipulation issue that impacts the accuracy of data from animal testing submitted in the biologics license application (BLA)
Novartis purchased AveXis for $8.7 billion in 2018. Brian Kaspar made over $380 million.
Within Business Community
https://retractionwatch.com/ Tracking retractions as a window into the scientific process
https://www.youtube.com/watch?v=PGBrfyOCCII Ivan Oransky - most concerning retractions
Tracking scale of scientific/publication fraud
A quantifiable characteristic may be RETRACTIONS
The Fraud SpectrumDefinitions….
FRAUD TRIANGLE
Donald Cressey…………
Motivation/Pressure
OpportunityRationalise
Must be the data, not my theory that’s wrong
“NON-MALEVOLENT” FRAUD
Bernard Fisher
Fisher-Poisson : breast cancer
1989
Martin Fleischmann of the University of Southampton and Stanley Pons of the University of Utah
Pons-Fleischmann : cold fusion
Fisher: giant of refining treatmentsPoisson: study clinician
Poisson falsified data to admit patients (1990)
3 yr investigation ORI - Fisher innocent of any scientific misconduct
“DELIBERATE FRAUD”
1. Pressure to publish (career progression, >250000 new papers pa)
2. Mental illness
3. Population deviants
4. Messianic
Alleged Examples……………………………..
David MazieresMIT Computer Science
2014
1. Pressure to publish: ‘Predatory journals’
https://www.youtube.com/watch?v=XomE5TquwhMEugenie Scott, Executive Director of National Centre for Science Education - Sceptic
2. Mental illness
Seminal ‘Nature’ paper
• Not reproducible• Significant consequences• Revealed?
Veronica James
Molecular structural changes in human fetal tissue during the early stages of embryogenesis Veronica J. James, Jack F. McConnell, Yoshiyuki Amemiya. Biochimica et Biophysica Acta 1379 1998 282–288
Synchrotron fibre diffraction identifies and locates foetal collagenous breast tissue associated with breast carcinomaV. James. J. Synchrotron Rad. (2002). 9, 71-76
Still being explored !!
3. Population deviant
>60 single author papersHigh impact journalsSignificant consequencesRevealed?
Aws S. Salim
Yoshihiro Sato (died 2017) bone researcher at a hospital in southern Japan, fabricated data for dozens of clinical trials published in international journals.
4. Messianic
Did the Japanese scientific community ever question how he managed to publish more than 200 papers, many of them ambitious studies that would have taken most researchers years to complete?
https://www.sciencemag.org/news/2018/08/researcher-center-epic-fraud-remains-enigma-those-who-exposed-him
Studies (2005-2010): 374 patients in 4 months280 patients in 2 months500 patients in 2 months
Demonstrated 70%-80% reduction in hip fractures
Governance & Policing
Doesn’t Peer Review help?
write MS -> submit MS -> review (>1) MS -> revise MS -> publish
Peters et al (Behav & Brain Sci. 1982 5 187-225)
Testing Peer Review………………………….
Douglas Peters – psychologist (test for author/institution bias)
12 published articles. All authors in prestigious departments.
Fictitious names (authors & institutes) substituted.
Submit to 12 high impact journals with high rejection rates (>80%) (same as original journals)
Met
ho
d
• 3 out of 38 editors & reviewers detected the resubmissions
• 9 out of 12 MS’s evaluated further
• 8 out of 9 rejected (serious methodological flaws)
Re
sult
s
Developed from Congressional OversightRun by CongressmanPowers of investigation, sequestion & subpoena
Some controversy
Committee on Publication Ethics (COPE)
Governance & Policing
ALLEA, the European Federation of Academies of Sciences
and Humanities,
2017
LANGUAGE
It has long been known that….
I cannot be bothered to look up the original reference
Three of the specimens were chosen for detailed study
The other results didn’t make sense and were ignored
Typical results are shown…
The best results are shown
Correct within an order of magnitude
Wrong
The correlation is excellent
The correlation is fair
The correlation is fair
The correlation is imaginary
While it was not possible to provide definite answers….
Nothing really worked, but I got a publication out of it
top related