- We had a treat today - a talk by Ed Yong, science journalist and purveyor-of-wonder *extraordinaire*, about scientific misconduct and fraud. I would have loved to add more of my notes to this, but then it would be unreasonably long. Just follow Ed on Twitter for more (of course you already do).
I will live tweet @edyong209’s talk on scientific misconduct today, thus blowing my cover in not attending PhD upgrade talks. Sorry guys :/- Oops... But how could I miss this? Some pictures (N.B. speaker brought sweets):
- Ed Yong’s talk about public engagement with sciences at U of Bristol lockerz.com/s/250482586
- The idea of this talk was to initiate undergraduates into the true world of science. To open the doors of perception. To teach them to be cynical and sceptical. To...well, to break them.
Ed spoke in the manner that we know him well: fascinating, funny, backed up by evidence, and a bit sweary. - Scientists aren't only on a pure quest for truth. They also want to publish papers, to find jobs, to get funding. Scientists are human: and therefore they are "flawed, egotistical, power-hungry and biased".
- .@edyong209 "I want to talk about science, warts and all...mostly the warts". (I'll use a #scienceishuman hashtag from now on)
- He talked not only of wilful fraud and misconduct, but also (presumably) unintentional bad practice: dodgy statistics and methodology, such as collecting data until you reach statistical significance and then stopping.
He talked of how difficult it is to reveal these problems. That replication studies languish unknown, unpublished, providing fodder only for private conversations between scientists. - .@edyong209: Studies that can't reproduce well-cited classic results are not published: they are "water-cooler gossip" #scienceishuman
- "We delude ourselves that because the scientific process is self-correcting that all scientists are so." -- @edyong209 #scienceishuman
- The problem is that we have an environment in which bad practices can survive, or even thrive. To me, the following was one of the most important points:
- "Incentives for the individual (eye catching good) are bad for the field (eye catching probably means it's wrong)" #scienceishuman
- Most journals have a policy of not publishing replications. And you can be the subject of attacks by the original authors. #scienceishuman
- The misty-eyed view of science being a self-correcting process is not entirely realistic...
- It's too simple to see science as a self-correction long-term process, for 6 reasons. 1. It's hard to replicate studies. #scienceishuman
- 2. It's hard to publish replications. 3. It's hard to interpret replications. 4. Some studies don't get replicated. #scienceishuman
- 5. Some replications aren't exact replications. 6. Fraud is usually caught by whistle-blowing or accident (bumbling authors) #scienceishuman
- And self-correcting is slow, expensive, time-wasting, career-wasting and raises hopes unnecessarily in the case of medicine. #scienceishuman
- I'm not sure at what point SG was mentioned (I must have been concentrating on tweeting)...
- Ed gave one reason many people get into science blogging, and later into science journalism:
- Like me, @edyong209 motivated to start blogging because annoyed at misrepresentation of science by popular press. #scienceishuman
flimsin
Tamsin Edwards
Climate scientist with a new blog: http://allmodelsarewrong.com
Share & notify
Notify
140
