Deceit and ignorance are the primary themes I am left with
from the Dan Ariely Ted Talk and assigned readings.
Fortunately, there appears to be a strong movement to mitigate their
effects in science.
The speaker discussed his research which showed a tendency
of people to show a lack of integrity unless reminded to do so, and that people
will practice more deception as their deceptive actions become further separated
from the reward. These results are not
very surprising to me, but do strongly suggest that some scientists engage in
deceptive practices (becoming a scientist does not automatically confer
immunity from engaging in a behavior that seems to plague humans). Moreover, as individuals whose livelihood
depends on funding, which is directly linked to publications, scientists have
incentive to practice deception. Thus, despite
the best intentions of peer review to eliminate deception, it should not be
surprising to find faulty experimental designs, hand-picked results, and
outright lies in scientific literature.
I was, however, struck by the alarmist attention given to
irreproducible scientific results in the assigned readings, especially the
Economist “Trouble at the lab” article. I
was stuck because I: 1) was not aware that there are so many examples of the
inability to reproduce results of previously published heretofore important science;
2) tend to be laissez faire, assuming that nearly all canons of scientific
discovery are vetted over time (and not single publications that gather dust) –
so question whether what I perceive as alarmist rhetoric is really warranted, and
3) am encouraged that efforts are being made by journals such as PLOSone and
Science to encourage replication of data.
Ignorance of proper application of statistical tools in designing and
interpreting experiments (on the part of authors and reviewers) is given as a
major contributing factor to publication of unreliable or unrepeatable data. Seeking to produce the best science I can, I am keener than ever to learn how
to properly apply statistical tools.
One last observation: From the Horvath blog, it was stated, “Many of my
colleagues worry that honesty and full disclosure will tarnish the reputation
of science.” Luckily, Horvath
disagrees with his colleagues, but who are these colleagues? It is hard to imagine that honesty and full
disclosure do any more damage than the stance of his colleagues. “Oh what a tangled web we weave unless we
publish and get a grant?” Is that the
mantra?
Horvath, J The Replication Myth: Shedding Light on One of Science’s Dirty Little Secrets Scientific American Guest Blog, Dec 4, 2013
No comments:
Post a Comment