Cookies   I display ads to cover the expenses. See the privacy policy for more information. You can keep or reject the ads.

Video thumbnail
It's hard to understand how science works if you don't
understand how scientists act upon the incentives that
scientific institutions provide them. And that has been a
real challenge for research in this area because up until
the very recent past we tended to study science at the
level of aggregates, such as sometimes entire countries
or maybe universities. And fortunately advances in
IT [information technologies] and the Internet, Google
LinkedIn and so on, have now allowed us to put together
massive amounts of data on individual scientists, being
able to keep track of where they are across time and space
for long periods of time. And being able to measure
their productivity in a very precise way.
One thing that people have been wondering for a long
time is what are the costs and benefits of scientific
collaboration. It's a very interesting question, but it's a
very difficult one to ask because people do not choose
their collaborators at random. So, in this research what me
and my co-authors have done is to focus instead on the
question of: What's happening to individual scientists
when one of their prominent collaborators is in some sense
taken out of their network because he or she dies
suddenly and unexpectedly? And what we find is that the
productivity of those collaborators declines quite sharply,
but also never recovers, so it's sort of a permanent loss
in output. So that's sort of the first finding but what's
even more interesting is understanding the types of
collaborators that are more or less affected by this event.
And what we find is that, in some sense, is the entire
field around the extinct superstar that atrophies following
the death of this individual.
I don't have a clue. But I know how we should try to
find out. And there is one word: experiment. The project
of my life is to convince the scientific community that the
way to figure out how things work is to actually subject
our hypothesis to systematic, randomized, controlled
experiments. And that's something that scientists have
been typically very reluctant to do. They think that the
scientific method applies everywhere except to themselves.
And so what really is needed here is a change of mindset
where instead of just listening to a very accomplished
and a very well scientist to figure out what to do, we
actually use their ideas as potential hypothesis' and then
we subject them to very rigorous tests — just like we would
for a scientific experiment.
Mostly I like to spend some time with my daughters.
They are three and five and they bring me and my wife
a lot of joy, sometimes some heartache and some grey
hair, but mostly it's quite wonderful. The other think I like to
do is bike; in particular I've now become a bike commuter.
We have this wonderful new building at Sloan with showers
in the basement, so one thing I've really enjoyed doing
since the beginning of this school year is to actually bike
from my house in Newton to MIT every morning and back.