You are on page 1of 23

Free Powerpoint Templates

The Ethics of Data


Tyler Park and Matt Zachreson

Free Powerpoint Templates


Page 1
Fraud-Why do it?
Free Powerpoint Templates

Page 2
Fraud-Why do it?
From Dr. Goodstein
Free Powerpoint Templates

ØCareer Pressure
Science is still a very strongly career-driven discipline.
Scientists depend on a good reputation to receive
ongoing support and funding; and a good reputation
relies largely on the publication of high-profile scientific
papers. Hence, there is a strong imperative to "publish
or perish". Clearly, this may motivate desperate (or
fame-hungry) scientists to fabricate results.
ØLaziness
Even on the rare occasions when scientists do falsify
data, they almost never do so with the active intent to
introduce false information into the body of scientific
knowledge. Rather, they intend to introduce a fact that
they believe is true, without going to the trouble and
difficulty of actually performing the experiments
required.
Ø
Page 3
Fraud-Why do it? (cont.)
Free Powerpoint Templates

ØEasiness of fabrication
In many scientific fields, results are often difficult to
reproduce accurately, being obscured by noise, artifacts
and other extraneous data. That means that even if a
scientist does falsify data, they can expect to get away
with it - or at least claim innocence if their results
conflict with others in the same field. There are no
"scientific police" which are trained to fight scientific
crimes, all investigations are made by experts in
science but amateurs in dealing with criminals. It is
relatively easy to cheat.
Ø

Page 4
Fraud-Why do it? (cont.)
Free Powerpoint Templates

ØFinances
There is the additional incentive of money. If one has a
promising proposal in an area in which federal or other
grant money or funding is available, especially in a new
technology in which there is no existing standard to
compare it with, the submission of preliminary data
cannot be confirmed until further research is done.
ØIdeology
While perhaps the least common incentive, it is still
there. The classic example would be anti-abortionists
claiming sonograms show the silent scream of an
aborted fetus demonstrates the fetus is alive with
feeling, while pro-abortionists would submit
demographic studies showing that women who
considered abortion but later decided against it are
doomed to life of dependency on welfare, lower
socioeconomic status, relationship abuse, child abuse,
drug abuse, etc. Page 5
Misuse of Statistics
Free Powerpoint Templates

Pulled from Wikipedia

Examples:

ØDiscarding Unfavorable Data


ØLoaded Questions
ØOver-generalization
ØBiased samples
ØMisreporting error
ØCorrelation/Causation
ØProof of the null hypothesis
ØData dredging
ØData manipulation

Page 6
Huh?
Free Powerpoint Templates

ESPN’s SportScience has made numerous blunders


statistically…

"In talking about how important speed in the half pipe,


Sport Science claims that for every 1 mph of speed a skier
loses, he/she will lose 3 feet in height" (Allain)

... How do you figure that?

Page 7
Correlation is not causation
Free Powerpoint Templates

There are two clocks


which keep perfect time.
When “a” points to the
hour, “b” strikes. Did
“a” cause “b” to strike?

Quito, Equador

Page 8
Correlation is not causation
Free Powerpoint Templates

“In the last episode, Sport Science wanted to predict the results of the
upcoming Super Bowl game between the Colts and the Saints. To do
this, they looked at some stuff from the past 10 years….
As you can see, they
looked at 4 things for
the two quarterbacks
playing in the game:
height, weight, age
difference, and state
they were born in.
From this, they
concluded that Peyton
Manning "has the
edge". I am
paraphrasing what
they actually said, but
this is basically what it
Page 9
was” (Allain).
How do I avoid post hoc?
Free Powerpoint Templates

“To avoid falling for the post hocfallacy and thus wind up believing
many things that are not so, you need to put any statement of
relationship through a sharp inspection” (Huff, 89). Is it by chance?
Are there more than one variable? Which is the cause and which is the
effect? Can one exist without the other?

Take a look at Coulomb’s Law—there is a correlation (although inverse)


between the electrostatic force and the distance, squared. How does
this hold up?

Page 10
Overgeneralization
Free Powerpoint Templates

Sub group a (usually a minority or even minuscule minority) of


Group A, has trait b but it is said that Group A has trait b.

A
a

Page 11
Where do we see overgeneralization?
Free Powerpoint Templates

“Our revised interpretation of students’ statements was based on


considerations that may be illustrated by the following example.
Consider the situation of a ball moving over a horizontal surface (see
Fig. 5). Friction on the ball is assumed to be absent (and whether the
ball rolls or slides is ignored). When asked to describe the movement,
to draw and name the forces that are present, and to explain the
answer, around 70% of the students at the start of the course give an
answer like: ‘There is a forward force that keeps the movement going.
It is not opposed, so the speed is constant’” (Dekkers).

* Also see subgroup problem in quantum computing.


Page 12
…or is it overgeneralization?
Free Powerpoint Templates

•Newton’s laws: "...when physicists began to


apply Newtonian principles to cosmological
conditions not sampled by Newton, discrepancies
began to appear." (And may I add likewise on the
atomic/subatomic level.) "Newton's laws failed
when physicists attempted to generalize the
Universe, and a new set of laws had to be
developed for ‘universal’ conditions” (Hammond).

•How about Einstein’s Theory of General Relativity or Special
Relativity?
"...the question arises as to whether Einstein's theory
incorporates this notion. That is, does the General Theory of
Relativity imply that one can generalize only so far as he has
sampled? To phrase the problem differently, after one builds
the theory, should he simply deduce the kind of results one
might obtain under certain limited, or immediately feasible,
conditions and report therefrom the adequacy or non-adequacy
of the theory?" (Hammond). Page 13
Be careful…
Free Powerpoint Templates

Page 14
Prevent overgeneralization:
Free Powerpoint Templates

"The rule is: generalize only those elements of the population


sampled" (Hammond).

What is your frame of reference?

Should we ignore this problem (the generalizability of the results) and


assume that generalization holds until proved otherwise?

Page 15
Outright Fraud: Case 1
Suppressing DataTemplates
Free Powerpoint

For years you’ve been debating a hypothesis with a colleague.


You’ve been doing experiment after experiment testing it, and often
your experiments prove your point. However, sometimes you get
results that are contrary to your hypothesis, but you can see how
they could be related to experimental error. Do you report your
conflicting results? If so how do you report them?

Page 16
Outright Fraud: Case 1
Suppressing DataTemplates
Free Powerpoint

Real Life Example: Louis Pasteur and Spontaneous Generation

When doing his famous broth experiments, Pasteur marked


all tests that grew life as “failed” (Geison 1996)

Page 17
Outright Fraud: Case 2
Tweaking/Inventing Data
Free Powerpoint Templates

You are right on the edge of a discovery that could change the world.
Your results could mean better, cheaper electronics, an end to the
energy crisis, or a super-effective treatment for a disease. However,
your results aren’t conclusive enough to get noticed, or at least won’t
be for several years, time that society could be benefitting from your
discoveries. (and you getting rich from the patents) What do you do?

Page 18
Outright Fraud: Case 2
Inventing Data
Free Powerpoint Templates

Real life Examples: Cold Fusion and Micro-transistors


Fleischmann and Pons really did see unexplained temperature spikes
in their experiment, but, under pressure from their university, they
came out with their results early and added detected neutrons where
they didn’t detect neutrons. End result: ridicule, and a current
aversion to continue their studies.

Schon claims that he really did see his dies behave as reported, but
manufactured data to make his evidence more convincing.

Page 19
Introducing: Erratum
Free Powerpoint Templates

“…since physics is a continually evolving field, a statement which


may seem quite reasonable at one time can be completely
overturned a decade later. Thus while correctness is not strictly
speaking time dependent, the standards by which physicists judge
correctness is time dependent” (Thomsen).

There is no doubt that fraud (intentional


error) is unethical—what about the
unintentional error?

Again, with Isaac Newton….

Page 20
The role of erratum
Free Powerpoint Templates

To correct errors, there are three


standard mechanisms:

1.The Erratum, in which the


authors describe one or more
errors in their paper and discuss
the implications, if any, for their
key results.
2.An independent author or authors
publish a new paper or comment
on the original paper in which
the errors are diagnosed and
implementations are discussed.
3.Allow word of mouth to
propagate the message that
there is an error in the paper.
(Thomsen)
Page 21
The role of erratum
Free Powerpoint Templates

“…by correcting their errors in print they likely saved other


physicists the time and expense of discovering the errors on their
own” (Thomsen).

See “The Effectiveness of the Erratum in Avoiding Error


Propagation in Physics” by M. Thomsen and D. Resnik to see the
results of their study.

Page 22
Bibliography
Free Powerpoint Templates

Allain, Rhett. "Sport Science: Pulling and Power." Dot Physics. Science Blogs, 01 Feb 2010. Web. 14
Oct 2010. <http://scienceblogs.com/dotphysics/2010/02/sport_science_pulling_and_powe.php>.

Dekkers, Peter J. J. M., and Gerard D. Thijs. "Making Productive Use of Students’ Initial Conceptions
in Developing the Concept of Force." Science Education 82.1 (1998): n. pag. Web. 18 Oct 2010.
<http://onlinelibrary.wiley.com/doi/10.1002/%28SICI%291098-237X
%28199801%2982:1%3C31::AID-SCE3%3E3.0.CO;2-1/pdf>.

Geison, Gerald L. “The Private Science of Louis Pasteur” Princeton University Press, 1996

Goodstein, David; “On Fact and Fraud: Cautionary Tales from the frontline of science.”, Princeton
University Press, 2010

Hammond, Kenneth R. "Relativity and Representativeness." Philosophy of Science 18.3 (1951): 208-
211. Web. 18 Oct 2010. <http://www.jstor.org/stable/185702>.

Huff, Darrel. How to Lie with Statistics. 1st ed. Vail-Ballou Press, 1954. 87-99. Print.

“Report of the investigation committee on the Possibility of Scientific Misconduct in the Work of
Heinrick Schon and Coauthors.”

Thomsen, M., and D. Resnik. "The Effectiveness of the Erratum in Avoiding Error Propagation in
Physics." Science and Engineering Ethics. 1.3 (1995): 231-240. Print.

Comics taken from various sources, including xkcd.com and phdcomics.com Page 23

You might also like