NSF OIG: Stories from the Case Files
National Science Foundation
Office of the Inspector General
Dr. Jim Kroll
Director, Research Integrity and Administrative Investigations Unit
Who is NSF OIG?
Inspector
General (IG)
Deputy IG
Office of Audits
Office of Legal,
Legislative, and
External Affairs
Admin. Investigations
(Investigative Scientists)
Financial Audits
CPA Contract Audits
Grant and External Audits
Performance Audits
Office of
Investigations
Expertise in all
areas of research,
grant, and contract
administration
Civil/Criminal
Investigations
(Special Agents)
Legal and Outreach
(Investigative Attorneys)
Investigations Specialists
and Analysts
Who is NSF OIG?
–
–
–
–
Independent office reporting to the Congress and NSB.
Promote economy, efficiency, and effectiveness.
Prevent and detect fraud, waste, and abuse.
Accomplishes mission through:
• Audits
• Investigations
– Criminal and Civil (e.g., false claims, false statements,
embezzlement).
– Administrative (e.g., regulatory and policy violations).
Where does research misconduct fit in?
OIG is delegated the responsibility for investigating
research misconduct allegations involving NSF programs.
–
Unique among the IG Community
NSF/OIG and ORI
–
–
–
–
Responsible for intake/assessment of allegations
Refer matters to university for investigation
Make recommendations regarding administrative actions
Work together on matters of joint jurisdiction
But there are some subtle differences
ORI
NSF
 Negotiates Voluntary Exclusions
Refers exclusion requests to
(VE)
NSF/OGC
 Oversees grantee investigations
Ability to independently investigate
 Not a law enforcement agency
LE agency with subpoena authority
Search warrant capability (criminal)
 Division of Education/Integrity
Limited outreach by investigative staff
 Publishes all findings/VEs with
All closeouts online but names
names
redacted except in debarment cases
Why do we care about RM?
• Fairness
– An NSF proposal is a request to obtain Federal funds from the
taxpayer.
– NSF program officers/reviewers assess the proposers’ ability to
carry out the proposed work and the proposers’ understanding of
the current state of the field.
– False representation of data and plagiarism misleads
reviewers assessment of PI’s capabilities/knowledge
• Economy, efficiency, and effectiveness
– Do not want to pay for research already completed.
– Do not want subsequent work to be based on misrepresented
work.
How does this compare to fraud?
RM Relationship to Fraud
Pearls Before Swine – Stephen Pastis
Fraud:
a misrepresentation of a material fact to
induce another to act to their detriment.
If NSF awards funds based on a proposal containing research
misconduct – the case is analyzed under the criminal and civil fraud
statutes and common law fraud doctrine.
NSF funding may be temporarily suspended during the process.
What is “appropriate credit”?
• Depends in part on the “relevant research community.”
• Basic idea: tell the reader what material you copied or
paraphrased and give the reader a map back to where
you got it.
(Q) C R
Quotation
Citation
Reference
(Q) C R
Quotation
• General rule – it should be clear and obvious what is copied and
what is your original work.
• Distinguish material copied verbatim from your original work with
quotation marks, block indentation, or other community standard.
•
If paraphrasing, make sure it is sufficiently paraphrased and not
substantially similar text.
– i.e., do not just rearrange the clauses or change every nth word to
a synonym, put it in your own words.
What if there is no other way to say it?
technically constrained text
or
common phrase
Hypothetical: Embedded Citations
This information not only furthers our understanding of cobalamin (B12) systems
[1], but also produces a large database of structural and chemical information
applicable to other areas of chemistry, such as the development of molecular
mechanics force fields for organocobalt systems [2]. For example, the Co–C
stretching frequencies of cobaloxime models with R = CH3 are very similar to
this frequency in the methyl B12 coenzyme [3]. Of particular interest here is the
fact that multiple factors influence NMR chemical shifts in such models and in
B12 compounds themselves [4]. Specifically, both Co anisotropy and equatorial
ligand anisotropy can affect shifts [4]. Model compounds have proved to be
useful in deconvoluting the various contributions to NMR shifts [5] and [6].
Moore, et al., furthers our understanding of cobalamin (B12) systems [1], and
also produces a large database of structural and chemical information
applicable to other areas of chemistry, such as the development of molecular
mechanics force fields for organocobalt systems [2]. The Co–C stretching
frequencies of cobaloxime models with R = CH3 are very similar to this
frequency in the methyl B12 coenzyme [3]. Of particular interest here is the fact
that multiple factors influence NMR chemical shifts in such models and in B12
compounds themselves. Specifically, both Co anisotropy and equatorial ligand
anisotropy can affect shifts [7].
[7] Moore, et al. 2009
Source
Subject
= embedded
object
Community Standards
•
•
•
Institutional Policies
Professional Societies
Journal Policies
Is there a different standard for faculty versus students?
Are there different standards for proposals and peer-reviewed journals?
NSF Grant Proposal Guide: The full proposal . . .
•
“should present the merits of the proposed project clearly and should be
prepared with the care and thoroughness of a paper submitted for
publication.”
•
“NSF expects strict adherence to the rules of proper scholarship
and attribution. The responsibility for proper attribution and citation
rests with authors of a proposal; all parts of the proposal should be prepared
with equal care for this concern. Authors other than the PI (or any coPI) should be named and acknowledged. Serious failure to adhere to
such standards can result in findings of research misconduct.”
The Inquiry/Investigation Process
Inquiry – Confidential; establishes substance; 120 days; may close institution ever knowing;
Potential QRP letter; data fabrication usually referred
Investigation - Substantive matters referred unless institution conflicted; 180 days to
complete; we use institution report as basis for our investigation;
OIG investigation independent – additional 180 days; may come back to you
to address unanswered questions
Adjudication - Institution should act only to protects its interests; OIG makes
recommendations to protect federal interests; NSF adjudicates, not OIG
Appeal - Director is final appeal
Closeout – We will inform you of our case close. All case closeout documents are
available online
http://www.nsf.gov/oig/closeouts.jsp
The Elements of an RM Finding
• Act must meet the definition
• Must be with a culpable intent (reckless,
knowing or purposeful, not careless)
• Preponderance of the evidence
• Act must be a significant departure from
accepted practice of the relevant community
The Research Misconduct Finding
For an NSF finding of RM the preponderance of the evidence
must support:
The act (e.g., plagiarism) committed by the subject; and
The subject’s intent in doing the act was at least reckless.
Careless
Reckless
Reasonable Person Standard
No Finding
Knowing
Intentional
(purposeful)
Individual Standard
Finding of Research Misconduct
Where would you put copy-and-paste plagiarism
on the intent continuum?
Plagiarism Detection Methods
• Review process
– NSF program officers and reviewers frequently notice text copied
from their own works appearing in proposals.
• General complaints/allegations.
• Software
–
–
–
–
Only finds textual similarities not figures, images, or ideas.
Does not determine plagiarism.
Proactive reviews by OIG.
Many programs out there with different features.
BOTTOM LINE:
There is no substitute for a manual analysis.
Community Standards
•
•
•
Institutional Policies
Professional Societies
Journal Policies
Is there a different standard for faculty versus students?
Are there different standards for proposals and peer-reviewed journals?
NSF Grant Proposal Guide: The full proposal . . .
•
“should present the merits of the proposed project clearly and should be
prepared with the care and thoroughness of a paper submitted for
publication.”
•
“NSF expects strict adherence to the rules of proper scholarship
and attribution. The responsibility for proper attribution and citation
rests with authors of a proposal; all parts of the proposal should be prepared
with equal care for this concern. Authors other than the PI (or any coPI) should be named and acknowledged. Serious failure to adhere to
such standards can result in findings of research misconduct.”
Case Statistics
Number of RM Allegations
140
120
100
80
P
FF
60
FFP
40
20
0
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
Case Statistics
Data Fabrication/Falsification Allegations
30
Data Fabrication/Falsification Allegations PI/co-PI vs Student/Post Doc
25
20
FF Allegations
15
FF Post Doc/Student
FF PI/co-PI
10
5
0
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
Case Statistics
90
80
70
60
50
Cases Completed
Research Misconduct Finding
40
Debarments
Cases Referred
30
20
10
0
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
Case Statistics - Trends
• Allegations peaked in 2008-2010
• Upward slope on number of substantive allegations
• Upward slope on the number of RM findings by NSF
• Recent increase in number of substantive data fabrication allegations in research
by students and post-docs
• Disconcerting number of PIs who believe that copying text is ok if you include a
citation to the source or if it is just “background” material
• Disconcerting number of PIs who rely on post-doc or grad students to write their
proposals without making sure they understand the rules/expectations
• Upward trend in violations of peer review confidentiality
• Conflicts between faculty and small business time and effort
Why the Increase?
• We are better at what we do
– More experienced staff, better tools, internet
• Culture clash
– Explanation, not an excuse
• Technology is a game changer
– Makes it easier to cheat
• High profile cases raises awareness regarding RM
• RCR training raises awareness regarding RM
• Government agency interactions with the research
communities raise awareness of where complaints can be
sent
Avoiding the dilemma
• When in doubt, make a clear distinction between what is
your original work and what is someone else’s.
• Know your communities’ practices (hint: look at the
standards for the journals in which you publish).
• When you work cross-discipline, know the other
discipline’s standards also.
• Adhere to the standards.
• Check with work of co-authors; especially of post-docs or
grad students are involved
A few of our favorite excuses
Can you explain why these do not work?
•
•
•
•
•
•
•
•
•
It’s only background material.
I did not put the text taken from a specific reference in quotes since it
usually makes reading a proposal difficult.
The reviewers are smart enough to know what is mine and what is not.
I used the same words, but I meant something different.
I copied the original sources that the review paper used so it’s cited.
I was told that having between 70-80 citations in a proposal was enough.
Anymore and I would look like I wasn’t proposing to do something new.
It’s not plagiarism if you change every seventh word.
My graduate student / post doc / lab manager / etc. wrote that part.
I was distracted by bird vocalizations outside my thatched roof hut, grabbed
my digital camera … , and when I returned to my computer where I thought
I had saved my changes to the material, it had crashed with the wrong draft
saved.
Case Files
•
•
•
•
•
•
SBIR gone bad
SUNY shenanigans
US – Egypt program
A three-fer in PA
Can I borrow your awarded proposal?
Disney Syndrome – It’s a small small world
Contact Information
www.nsf.gov/oig
Hotline:1-800-428-2189
E-mail:oig@nsf.gov
Fax:(703) 292-9158
Mail:
4201 Wilson Boulevard
Suite II-705
Arlington, VA 22230
ATTN: OIG HOTLINE
Jim Kroll
jkroll @ nsf.gov
703-292-5012
?