7 December 2019 | New Scientist | 23C
LAIMS of “fake news”
within the UK Houses
of Parliament are nothing
new. This time, however, the
charge has been laid not at
the door of politicians, but of
scientists. And it was scientists
themselves making the claims.
They came at a meeting I attended
last week where the British
Neuroscience Association (BNA)
launched a fightback against bad
science with its “Credibility in
Neuroscience” campaign.
The problem isn’t just that some
findings turn out to be wrong. It is,
after all, the point of science to be
constantly questioning, testing
and refining hypotheses. The BNA
campaign claims that the entire
structural edifice of academia now
encourages mistakes to be made.
This starts with well-meaning
efforts by managers and funders
to judge researchers’ productivity.
That is done by gauging how many
papers they write and the prestige
of the journals that publish them,
as quantified by their “impact
factor” – basically, an average of
how often the papers they publish
are cited by other papers.
Researchers’ publication
records increasingly govern
every aspect of their career
success, including pay rises,
future jobs and funding for new
projects. In this “publish or perish”
culture, it is in their interests to
produce a blizzard of papers that
are groundbreaking and flashy,
so as to get published in high-
impact journals. With an eye to
JOSmaintaining their impact factor,
IE^ F
OR
D
Comment
Clare Wilson is a biomedical
reporter for New Scientist
@ClareWilsonMedViews
The columnist
Graham Lawton on
the environmental
impact of pets p24Letters
Transparency on
political micro-
targeting p26Aperture
A stunning view
of Bali’s subak
rice terraces p28Culture
The best science-
themed gifts to
buy this year p30Culture columnist
Silicon Valley lets tech
giants off easy, says
Chelsea Whyte p32journals are incentivised to
publish such papers, rather than
ones that, for example, describe
attempts to replicate others’ work.
The resulting system is the
antithesis of how good science
should be done, namely by tackling
questions in a thoughtful and
systematic way and by testing and
retesting any unexpected result in
different labs and circumstances.
Prime evidence of how bad thingshave become is the “replication
crisis” in psychology, where
doubts have been raised over
classic findings such as priming,
the idea that behaviour can be
changed by subtle, unconscious
cues. In psychiatry, a review
published this year called into
question two decades of work on
a link between depression and a
gene affecting the brain chemical
serotonin. “It wasn’t just thatpeople said it mattered and it
didn’t, it’s that we built whole
castles in the air on it mattering,”
said psychologist Dorothy Bishop
of the University of Oxford at the
BNA event.
The campaign is a laudable
attempt to change this underlying
culture. As Bishop said, “We have
to do things in a way that ensures
discoveries are robust.” The aim is
to do that by lobbying universities
and funders, as well as by training
scientists in best practice, such as
data sharing and registering
studies before publication,
meaning mistakes are more
likely to be noticed.
Happily, it isn’t a lone
initiative. The recently launched
UK Reproducibility Network is
encouraging higher research
standards more broadly. Growing
numbers of bodies are signing the
San Francisco Declaration on
Research Assessment, made at a
cell biology conference in 2012, to
say they won’t use journal impact
factors in decisions on funding
and job appointments.
If the movement succeeds,
that would mean fewer interesting
stories for journalists like me to
write about, but the ones we do
cover would be more likely to be
true. “If everyone’s trying to do
groundbreaking research, you
just end up with a lot of holes in
the ground,” said Bishop. “You
don’t get anything built.” ❚Science’s fake news problem
A productivity-driven research culture encourages false results – but
there are welcome signs of a fightback, says Clare Wilson