Good and Bad
Scientific Practice

Good Scientific Practice …

... lesser a set of rules than an attitude to be aquired

Terminology

scientific misconduct \(\longleftrightarrow\) good scientific practice

• scientific misconduct: a serious offence, defined by law
• good practice: doing honest and trustworthy work
  (less rigorously defined)
• bad practice is between these poles
  

• doesn't bring you to jail
• but may affect your reputation among
  your peers, colleages, and students

Scientific misconduct

• Definition

  falsification, fabrication, plagiarism and other serious violations of good scientific practice commited wilfully .. on planning, performance or reporting of research results.

• Examples
  • theft of ideas
  • undisclosed excluding undesired data or manipulating images
  • undisclosed unusual and misleading use of statistical methods
  • utilizing (co-)authorship of a person without his consent
  • (co-)authorship of a person without essential contribution

Protection systems against research misconduct

examples of high rank investigation comissions

• Denmark: Committees on Scientific Dishonesty
• USA: Office of research Integrity
• Germany: Ombudsman für die Wissenschaft

• contact persons for scientific integrity/counseling system
• commission at our university
• certainly similar also at your university

the less these commissions have to do, the better
but, there mere existence shows: they have their work …

Useful Documents

• On being a scientist
  Committee on Science Engineering and Public Policy (2009)
• Singapore Statement on Research Integrity
  The 2nd World Conference on Research Integrity (2010)
• Safeguarding Good Scientific Practice
  Deutsche Forschungsgemeinschaft (2010 + 2013)

How to survive

Documentation is your ensurance

• lab records, data treatment and statistics
• computer codes, output, source code for image generation

• Everything needed for reproduction (may depend …)

• primary data: a minimum of 10 years (recommends DFG)

• ask for local regulations!

That's pretty obvious, what's the problem?

• durability of media, data managment
• consistency of data (is it complete? refers to same moment in time, …)
• understandability (what does it mean? is it really related to published results or just "stored away", …)

Publication

• as a scientist you have to publish your findings

• characteristics of a good paper:
  1) reporting sound, original and interesting research
  2) necessary documentation and clear presentation
  3) fair credit to previous work
  4) proper credit to contributors (title page or acknowledgement)
  

• 4) and 3) deserve attention to avoid misconduct

Authorship

An example with many authors

Rights and obligations for coauthors (see: American Physical Society)

• all collaborators share some degree of responsibility
  • either responsibility for the entire paper as accurate and verifiable report of research

    e.g., integrity of data, main analysis, manuscript writing, conference presentation, scientific leadership

  • or responsibility only for specific, limited contribution (quality ensurance process!)

    e.g., analysis of a limited part of the data, a special, non-decisive way of analysis

• every coauthor should review the manuscript and has obligation to prompt correction of errors or even retractions of parts

Order of author names

• is not a matter of taste but of
  • alphabet (e.g., mathematics, theoretical computer science)
  • amount/kind of contribution (e.g., live sciences)
  • tradition (all communities)

• non-author contributions should be acknowledged

• non-contribution authors are to be avoided

Quotations and Permissions

• proper referencing is essential:
  • gives credit to related work, puts current into context
  • supports verification and guides readers
  • helps to avoid duplication of effort

• clearly identify the source of:
  • important findings, ideas, arguments
  • mathematical derivations, algorithms, computer programs
  • photos, diagrams, tables, verbatim text

• ask the copyright holder for permission to use figures!

All very understandable, what's the problem?

• to identify the true source of an idea, photo, …
• to find an appropriate trade-off between completeness (survey paper) and representative sample (research paper)
• to estimate a collaborators contribution
• to agree on a joint publication policy

Note on the refereeing process

• peer-reviewed publications are a key ingredient of a scientific career
• as a researcher you typically live in both worlds
  • your papers get reviewed and
  • you act as a reviewer on other author's papers

How to (not) run into misconduct

Categories of misconduct reported in press

• fabricated or fictive data
• plagiarism
• manipulating statistics
• questionable scientific basis for drawn conclusions

• people get carried away or
• are under pressure

Concluding Remarks

Self protection against being involved (not 100% …)

• observe your scientific environment:
  How would you decide?

• train your senses for scientific (mis-)conduct

• read about research integrity (see references)

• discuss with others

References

• Peter Sigmund: Good and Bad Scientific Practice. Lecture at the University of Southern Denmark (2011)
• The dilemma game Professionalism and integrity in research. Card game developed by the Erasmus University Rotterdam

Technical acknowledgement

This presentation is based on reveal.js and was prepared using the org-mode of emacs and its extension org-reveal.