22 August 2020 | New Scientist | 39politicians. But if people keep publishing
these hyped results which end up not
actually delivering on the promise, trust
in science will decline.
We see it already in nutrition research.
I think most people don’t take that seriously
any more. I would hate to get into a situation
where other sciences are in that same boat
and research becomes a laughing stock.Do you worry about giving ammunition
to anti-science voices?
I do. I worry particularly about already
politicised areas like climate change, vaccines
and evolution. The last thing we want is to
give climate deniers and creationists yet
more reasons to be critical of science.
I’m critical of the way science is done, but
I’m not critical of the idea that science is the
best way we have of discovering how the
world works, of making new things that
improve the world for everyone. That is the
whole reason that we must be sceptical and
critical of it. I would be disappointed if people
took this book as saying, “Well, we can just
chuck away science.” That’s absolutely the
opposite of what I’m trying to advocate.How can we fix it?
I think the broad answer is scientists being
more open and transparent about what theyactually did in their research. But I don’t
think that’s the only answer. Universities
need to stop hiring people just on the basis
of how many papers and citations they
have and start rewarding a different kind
of science: people who contribute data to
the world, people who are part of large
replication projects, people who are creating
new tools to help other scientists do research
more efficiently. Being a good scientific
citizen, essentially.
Journals need to change the way
they publish articles. They can incentivise
people to preregister their work, so they’ve
published their data analysis plan before
they actually get to the analysis and can’t
do stuff like p-hacking.Ultimately, I feel like this is an optimistic book,
even though it is essentially a litany of human
failure and weakness.
The amazing thing is that even with all these
problems, we can still know stuff and make
progress. But imagine how much more we
could know if we managed to build a better
system than the one we have. We can do
much better. We can change the way the
system works.
But I also think we need to change
what we think science is, and have
different expectations of it. We should
expect scientists to be much more open,
but also more boring.
The perverse incentive to make
research look exciting is at the root
of a lot of the problems. Science is not
an endless march of exciting, flashy
findings. There are transformative
discoveries, and we should try and
encourage those. But in general,
science is incremental and small scale
and requires a new kind of intellectual
humility. I think we will be much better
off when we realise that. ❚PL
AINPICTURE
/KRISTA^ KELTANENGraham Lawton is a staff
writer at New Scientistus this is true” rather than revealing how
the process really works.
Why should people who aren’t scientists
care about this?
So much of what we do relies on science. So if
the whole system is off, then that has major
consequences. It couldn’t be more clearly
illustrated than with covid-19. All of the
things I’ve talked about have come up in the
pandemic: accusations of fraud, retractions
from top journals, screwed-up statistical
analyses, lack of data sharing, negligence.
And it has put rocket boosters under hype.
Another lesson we’ve learned is the
dangers of pushing science out too
quickly. We’ve seen a huge upsurge in
research about covid-19, but a lot of it is
probably completely useless. It’s kind of
unbelievable. There couldn’t be a better
illustration of the need for change.
Also, a huge amount of taxpayers’ money
is wasted on useless science. It actually makes
it harder to discover the truth in future
because it muddies the picture.
What is at stake? If science just shrugs and
carries on as normal, what will happen?
At the moment, people have a high degree of
trust in science and I think that’s justified to
some degree, certainly compared with
Even the world’s
leading journals
occasionally publish
poor science