thegarnet.net

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
===== Science: the good, the bad, the ugly =====

Science builds upon science: Publications as building blocks

Most scientists are honest and hard-working

Scientists are biased

Scientific studies often not reproducible

Errors in statistics and models

Sloppy research

Science misconduct


Articles:
  * Why most published research finding are false - PLOS One, 2005
  * Estimating the reproducibility of psychological science - Science, 2015


===== Publish or Perish? =====
Why do scientists make errors, intentionally via alsification...  

Publish or Perish cartoon “Welcome to the team. Remember if you follow the university motto, you’ll do fine....

===== Definition of Research Misconduct =====
  * Plagiarism - stealing text or ideas
  * Falsification - manipulating or omitting data
  * Fabrication - making up data

Self plagiarism - text recycling - does not fall under definition

Research misconduct does not equal honest errors or differences of opinion

Office of Research Integrity

==== Research Misconduct: Plagiarism ====

Cartoon: I need you to do a presentation on the topic of “plagiarism”. You dont have any time for it, so just steal something off the internet

=== Plagiarism in a review paper ===

Used text from at least 10 source papers. Author has 343 publications (113 reviews) - paper was retracted


=== Another example of plagiarism ===

At least 43 source documents were used that were NOT listed in the Referenes. Last author of this paper was the Senior Associate Editor at this journal.  Paper was retracted.

=== Plagiarism in Science Paper ===
  * $$: iThenticate, Turnitin, WriteCheck
  * Free: http://smallseotools.com/plagiarism-checker/
  * Google / Scholar: 5-10 words between quotes

Personal project:
  * 80 review/research papers found, reported to ournals
  * 35 retracted, 11 corrected, 34 addressed


Exceptions for definitions

Talking about complete paragraphs just copy pasted


==== Research falsification #1: Andrew Wakefield ====

Claimed 12 in a row kids with measles vaccination had immediate autistic issues.

Paper retracted

Falsified the results of colonoscopy to say colitis

This is the source of the whole anti-vaccine movement

===== Research falsification #2: Haruko Obokata =====
Claimed treating any cell with acid, can turn them back into stem cells.

Paper in nature

Immediately people started talking about this

Retracted within months

Found photos flipped and rotated

She was allowed to work under video, she was never able to replicate her research,

Her mentor hanged himself in a stairway of the institute, eventhough he was never suspected of involvement of the misconduct.

  * Very serious busines


===== Research Fabrication: Diederik Stapel =====

  * 137 scientific papers
  * 58 retracted so far
  * reported by his own students
  * odd data; most appears to have been fabricated

“Coping with Chaos: How disordered contexts promote stereotyping and discrimination”

Most of data fabricated.

Site: RetractionWatch - he is #4

===== Inappropriate image duplication =====

“The prevalence of Inappropriate Image Duplication in Biomedical Research Publications”
  * Elizabeth M. Bik, Arturo Casadevall , Ferric C. Fang.


Found:
  * 20,621 papers from 1995-2014. (4 years to go through)
  * 40 journals from 14 publishers
  * ~800 papers with problematic figures (4%)
  * three main categories of problems
  * limited to photographs - did not include tables, or numbers

==== Biology research images should be unique ====
  * microscoping photos - each unique
  * chromosome pictures
  * every band is unique 

==== Type 1: Simple image duplication ====
  * Messenger RNA experiments

“These look very similar to me”. Stops short of an accusation

==== Type 2: repositioned image duplication ====

  * Overlapping photos - shifted
  * shifted panels
  * mirrored panels

==== Type 3: Image Alteration ====
Even worse, image photoshopped

All examples found, the papers got retracted.

===== Spot the overlapping panels =====

Example of an overlap - several in the slide

30 panels

A dozen overlaps

===== Is all science corrupt?  Most papers are OK! =====

  * 4% of biomedical papers have inappropriate image duplication
  * of those, 10% will be retracted
  * 4/10,000 papers have been retracted (RetractionWatch)

Weird excuses
  * forgot we had duplication in a mockup
  * blame the students

Articles:
  * Misconduct accounts for the majority of retracted scientific publications
  * Analysis and correction of Inappropriate image duplication: the Molecular and Cellular Biology Experience

===== Predatory Publishers =====

Predatory publishers 
  * run under “open access model” - you pay to publish, it is fee for others
  * send spam emails to researchers
  * offer fast peer review, open access
  * publish many journals, many topics
  * low quality papers, not indexed
  * fake impact factors / editors
  * authors are:
    * inexperienced researchers
    * lone authors with wild ideas
    * shady companies

Contents
  * empty
  * poor quality


Examples:
  * IOSR Jourals - international organization of scientific research
  * Science domain international
  * eSciPub
  * OMICSGroup
  * JP Juniper Publishers
  * MedCrave

Hard for lay person to know which ones are red flagged. But an example is time between submission and published:
  * submitted: Feb 23, 2018
  * published: Feb 23, 2018


You can not republish after publishing in a predator. You can only publish once.


===== Predatory Conferences =====
  * WASET conferences - 200 conference happens in the same room
  * people submit fake papers - and people reports
  * everyone gets an award

  * multiple conference at Sam day/hotel
  * nice locations
 
===== SciGen to expose fake conferences / journals =====

Decoupling IPv4 from 802.11B in Reinforcement Learning
  * Elisabeth M Bik, Pudge Purrmeister and Oifa McFloof

===== Fake Peer Reviews and Affiliations =====

South Korean plant compound research faked email addressed so he could review his own studies


Peer reviewed scientific journals with people with impressive affiliations, but they never worked at their institutions, or the institutions didnt exist

Examples of fakery:
  * Physician working under the influence of alcohol: an analysis of past disciplinary proceedings and their outcomes - Forensic Science International
  * Mental and Physical orkload, Salivary Stress Biomarkers and Taste Perception: Mars Desert Research Station Expedition
  * A mathematical model for DNA
    * males and females with different polarized
thegarnet.net

User Tools

Site Tools

Page Tools
