While resistance to change is an intrinsic facet of human nature,
it never ceases to amaze me when it is exhibited by individuals
whose lofty purpose in life is precisely to be imaginative—able
to think for themselves, challenge false beliefs, and come up
with inventive, original, groundbreaking concepts or solutions.
I am referring to engineers, scientists, and other researchers,
who can be so rigorous and so creative in their technical work
yet so disorganized and so conformist in their communication.
Those same people who make fierce referees, willing and able
to criticize the work of their peers without mercy in an effort
to defend what they regard as the truth, will follow like sheep
the bad reporting practices of their company or of their field
in an incomprehensible attempt to “fit in the group” at any cost.
As an instructor of scientific communication, I am admittedly
not one to say “this is how people in your field typically write,
so this is how you should write, too.” Being rather immune
to peer pressure (for better or for worse), I find myself saying
more often “this is how people in your field typically write…
but it sucks. Let's discuss a better approach to writing papers.”
This approach has served me well with my (rational) audiences,
who normally see the point and agree with me—some of them
more readily than others, true, but all of them to a fair degree.
The barrier is thus not the participants… it is their supervisors,
usually group leaders, who did not take part in the discussion,
who do not appreciate to be challenged by their PhD students
or postdocs, and who feel the need to assert their authority.
Needless to say, the supervisors of my workshop participants
have a right to their own opinion and so to disagree with me.
The problem is, they just have no opinion of their own, really:
they follow a dogma—and one that is all the harder to dispute
because the authority that laid it down is ill-defined, not to say
nonexistent. Thus, whenever they see a more junior researcher
try something new, they exclaim, “that's not the way it is done!”
Through their recommendations and their own writing practice,
they imprint dubious habits on these budding scientific minds
eager for advice at the launch of their career. In the process,
they further propagate the myths of scientific communication,
such as the taboo against the use of the first person—a taboo
often decried in “author's guidelines” yet regarded by authors
as a sine qua non for their papers to be accepted by journals.
Such myths are viruses: they are deleterious habits endemic
to the scientific world that spread by replication inside papers.
One particularly strong instance of resistance to change I face
is the choice of graph type in the life sciences. The repertoire
of authors in life-sciences papers indeed seems to be limited
to vertical bars in all their variations (single, multiple, divided),
usually with one-sided error bars to boot as shown on the left.
Even when they do not constitute visual lies (as when they fail
to run from zero, are incorrectly spaced horizontally or, worse,
run along a logarithmic scale), these familiar telegraphs poles
use a lot of ink for what are few data, and fail to reveal trends.
In many cases, they could usefully be replaced by line graphs
or simply dot plots (revealing all the individual data points).
Recently, a PhD student in one of our training programs tried
such alternative graph types for his own experimental data,
only to be told by his group leader that this was “not the way”.
His attempt at a rational discussion around graphical displays
was soon cut short by an argument of authority on her part:
“change everything to the way you presented your data before”
she said, “or show me a highly cited publication which uses
your way of presenting data sets.” Is this the spirit of research?
Is a PhD thesis not all about making an “original contribution”?
I am a firm believer myself in peer review for reporting science,
but if peer review means that authors no longer dare to submit
what might be a (somewhat) original graphical representation
to their peers for review, there is something definitely wrong.
After all, what was the supervisor of our training participant
so afraid of? What is the worst that could happen? A comment
from one of the referees, perhaps? If that be the case, so what?
Just change the graph if you must, then resubmit your paper—
or better, write to the editor and defend your choice of graph.