Teaching Research Ethics or Learning in Practice ? Preventing Fraud in Science

DIES NATALIS LECTURE ISSTHE HAGUE, 9 OCTOBER 2014

KEES SCHUYT, PHD, LL.MSOCIOLOGY PROFESSOR EMERITUS, UNIVERSITY OF AMSTERDAM;CHAIR NATIONAL OFFICE OF RESEARCH INTEGRITY (2006-2015)

Teaching Research Ethics or Learning in Practice?

Preventing Fraud in Science

Two phenomena, five topics

Scientific integrity (what it is and isn’t)

Data-management (good and bad practices)

Five topics:

1. What do we want to prevent?

2. Good and bad practices

3. Why does it happen? - Tentative explanations

4. What is to be done? - Rules or principles

5. Educating, learning, mentoring

1. What do we want to prevent ?

History of fraud in science (Baltimore-case (1986-1996) as turning point; US Office of Research Integrity, 1994

Broad and Wade (1983); Van Kolfschooten (1996, 2012); Grant (2008)

Levelt - report on the Stapel-case (2011/2012)

What can we learn from incidents (outliers)? (teamwork; the system is not watertight: good datamanagement)

Scientific integrity

Integrity is a self-chosen commitment to professional values (B. Williams 1973)

Resnik: “striving to follow the highest standards of evidence and reasoning in the quest to obtain knowledge and to avoid ignorance” (The Ethics of Science,1998)

Integrity is context bound, eg. fabulation in novels and fabulation in science; leading values in science (Merton 1949)

Codes of Conduct: NL 2005/2012; ESF 2010

Violations

Violations of the game rules of science:

FFP : fabrication or fabulation

falsification plagiarism

Difference between F and P?

2. Good and bad practices

Questionable research practices (trimming, cooking, pimping, sloppiness, uncareful data management, not archiving)

Drawing the line (raw data, co-authorship, impolite behaviour)

Trimming and cooking (Babbage 1830)

Trimming: “consists of clipping of little bits here and there from those observations which differ most in excess of the mean, and in sticking them on to those which are too small”

Cooking: “to give ordinary observations the appearance and character of those of the highest degree of accurance. One of its numerous processes is to make multitudes of observations, and out of these to select only those which agree, or very nearly agree”

Metaphorically:

“if a hundred observations are made, the cook must be very unlucky if he cannot pick

out fifteen or twenty which will do for serving up” (Charles Babbage, Reflections on the

decline of science in England and some of its causes, 1830; 1989 edited by Hyman)

Four main distinctions:

honest vs dishonest, fraudulent

good vs bad practices

controversies vs dishonest research

game rules vs goal rules

Data-management

The scientific research cycle: 3 strong controlling points: grants, peer review, scientific community2 weak points: primary process and data-archivingWide variations between disciplines: is everything okay? Bad to good practices: single vs teamworkScale of research: international data-gathering; protocols

Variations in data and in data-gathering

Experimental design data (lab)Stemcells, MRI-scan dataMathematical data, logical analysis Survey-data (pen and pencil)Public data (time series, economic data,

populations figures, official statistics)Historical data (archives)Anthropological field observationSimulations

3. Why does it happen?

Three main explanations: o Publication pressure: from who to whom?o Sloppy science o Pressure from contract research

Alternative tentative explanatory scheme:misplaced ambition, loose mentoring, ignoring early signals, poor peer review, no institutional response

Contract research

What is the problem? Köbben 1995: scientific independence; pressure from above (yes, minister); conflicts of interests

Research biases? Biomedical research; Roozendaal

Patents, secrecy, firm’s data not publicRemedies: “good fences make good

neighbours” (R.Frost), applied to contractsResearch codes, guidance committees, High

Prestigious Research Group (hprg)Conclusion: be a hprg: integrity high, high

skills, independent

4. What is to be done?

Learn from best practices across disciplinesPeer pressure before peer review; data-

manager and/or statistical counseling; open discussions to keep alert (not too often!)

Scientific pledge or oath taking!? Lowering publication pressure? (causality!)Educating ethics in science; integrated in

data-management courses

5. Educating, learning, mentoring

The sixpack:

a learning rules, discussing ethicsb training research skills (eg. advanced statistics, philosophy of science)c good mentoring (becoming a good scientist)d oath-taking (!?)e online learning, the dilemma gamef reading Being a scientist

Select your own best combination

Gift to all PhD students:

Thank you very much indeed for your attention