Monday 5 September 2011

Never be rude to a scientist, no matter what you do...

Never trust a scientist, and that includes me.


I've become increasingly pissed off by the industry I work in, an industry where people who lie and cheat do well, get jobs, build careers while some who are honest and have integrity get anally raped by their employers and told that they should 'probably go and do something else for a career'. People far better than me are in danger of being thrown out of science because they don't 'make money'. 


This is why, and this is how it works...


Scientists need money to do experiments (as chemicals, equipment, cells, animals, lab coats and cocaine all cost money). This money can come directly from Universities, from charities or from the Government (channelled through organisations known as 'Research Councils'). You can also get money from  Europe but only if you have some European mates, some friendly companies and lots of time.


In order to get this money you have to write a grant application, and to increase the chances of your application being successful you have to show that you have pretty much already done all the work and just need to confirm a few things and finish off a few loose ends. Do you see where the system falls down there? A good idea won't get funded unless you've pretty much already done it, and in order to do it you need to be funded...


The best way to write a successful funding application is to have already published your 'preliminary' work in a scientific journal. This publication will have been checked by independent scientists (peer reviewed) to make sure it is of the required quality. Of course they don't see raw data, you just send them graphs you make on the computer. Graphs that you can make say whatever you want. So, not a good system really... 


So this gives scientists a quandary. In order to be successful you can either


a) be honest (and lucky) and see if you can scrimp around and get enough data to get some papers and preliminary data which you can use to 'sex up' your application.


or  


b) Make a load of data up that looks impressive and pretend that you're a really good scientist (after all, not many people are going to ever check to see if you've done the work are they?). This way your application looks amazing, and you're more likely to get the money.


The latest estimates are that about 70% of published science has at least one fabricated or embellished piece of data in it. So basically put, most science is bad.


Now imagine (it's not hard to) that in this economic climate, which has hit Universities as hard as anyone, your employer is putting pressure on you to get money in (and to get high impact papers). If you don't, you're out. End of your job, and possibly career. No money for the mortgage, fuel or food. Would you not be slightly, just slightly tempted to think 'If I make up some data, and get it published, I have a better chance of getting a good paper out and then some money in'. I would be. So far I haven't come under that pressure but i don't know how I would react if I was. 


The thing is I've seen people who have been in this position, and far worse I've seen people doing 'bad science' even when they're not under pressure. This has ranged from doing one experiment, liking the result and saying that you repeated it 5 times to check the result (which would make it statistically significant, an essential part of any scientific experiment) to completely and utterly fabricating a result to give you a graph that supports the hypothesis you want. I have sat in meetings where a junior scientist has been told to 'say we did it 5 times' when he had done his experiment once (or possibly made it up)  and I have seen people get papers from blood samples that do not exist (from which they haven't even tested the stats correctly for multiple comparisons #geektalk).


And who, in all of this, has the better career? So far I have been honest, I haven't made up data, I haven't cheated. And my career is going OK, but not spectacularly. I'm not under immediate threat but things will get interesting at the end of my contract. But people I used to work with are doing better than me, getting more papers, better papers when I know they are cheating. In an interview situation, they would blow me out of the water.


It fucking stinks. It hurts. Never trust a scientist. 



6 comments:

  1. I've just done a graph which shows the number of crocodillapigs I've cloned against the amount of cancer that has been cured over the last decade, and according to that graph there is a direct and significant correlation. I've even had it peer reviewed by my mum, so it's legit.

    Just waiting for the grant money to be delivered by forklift now *looks hopefully out of shed window*

    ReplyDelete
  2. Dude, ahhh but what was your n number (quickly, whatever it was times it by 5 or something)

    ReplyDelete
  3. I am glad to see that the hypocritical experts that teach morals and then ignore them, which exists in the worlds of IT and psychotherapy, are alive and well in other professions.

    ReplyDelete
  4. It lives, breathes and thrives...

    ReplyDelete