Hacker News Reader

Comment by impendia | Hacker News Reader

As an academic in math (perhaps the lab sciences are different, I speak only for what I know), I find this absurd. Perhaps mathematicians are nicer than other academics, but within math I have never heard firsthand of any story of this kind.

The advice given is breathtakingly cynical. You take risks by sharing your accomplishment, yes, but if you don't describe your accomplishment to others then you are depriving yourself of oxygen. Communication is the lifeblood of science, and in math at least, others are 1000x more likely to help you than to screw you over.

I would recommend, to anyone pursuing a Ph.D. in math at least, that they categorically ignore this advice. Hell, I'm confused as to why anyone prepared to believe this would want to get a Ph.D. at all.

View on HN

I have a PhD in Applied Math from one of the Math departments in the country, and now do biomedical research. Believe me, the cynicism is well placed.

I wanted to get a PhD because I believed as you do. The fact is, though biomedical research is a field where cheating can take you a long, long way. The experiments have gotten so big and expensive that scientific accountability is at an all-time low (for example, check out this commentary [1] on pre-clinical cancer research. And the money in biomedical research attracts power-hungry people. At the same time, the pressure to publish in order to obtain grant funding is enormous, and the incentives to cheat to get ahead by e.g. playing games with credit or retarding competitors with sly reviews is correspondingly intense.

[1] http://www.nature.com/nature/journal/v483/n7391/full/483531a...

  TLDR: drug company attempts to replicate results of high-impact cancer
  biology research papers, in a search for new drug targets.  Failed to
  replicate the claims in 47 out of 53 papers.  Authors refused to
  provide materials necessary for replication until company signed NDA
  stating they would not publish the results of the replication
  attempts.  One author tells the head of the replication effort "Well,
  we only got this result 1 time out of 5 repetitions, but we decided to
  report the 1 instead of the other 4, because it's the better story."
  (NDA story is from the subsequent Nature clarification, sentence 4 is 
  from a news report on the commentary.)

Er, s/one of the/one of the best/ :)

As someone who spent time with both biology and computer science PhDs, every bit of that article is true when applied to biology (much less so in computer science, and probably none of it applies to math). Actually it gets worse in biology than what the article describes. There are networks of "friends" established across the country which survive by sharing credit as much as possible. (Someone should write a paper on how such cliques get established -- it is almost like Darwinian evolution, except the other way around.) Members of those cliques often have impressive publication records, but when you start drilling down to what their individual contribution was to each paper, you start believing that you are watching some circus unfold in front of you. In an individual lab, members of such cliques can achieve significant power and influence (since friends know tricks to get paper closer to publication) even despite the fact that their individual ability to interpret experimental data is often very unimpressive to put it nicely.

It's an article about biology, where it's actually understating the behavior, as others have noticed. The subtext that no one has pointed out yet is how provincial biology is. Biologists fervently ignore other fields. I think it's partially because they're not as bright as physicists, chemists, or other scientists, partially because they are poorly trained, and partially because it would be so embarrassing to have their culture contrasted with a healthier one.

Yes, yes, and yes, however I believe one of the primary reasons why the academic culture in the field of biology became toxic is that it had become a common practice to have multiple authors in a publication. Initially the practice was quite benign because many experiments indeed require expensive instruments and knowledge of some quite specialized skills, however, this practice was kept unchecked (there is no real penalty of any sort how many authors you add to the paper, except that some authors themselves may object to having to share credit), which allowed for cliques to become established where you basically have a few "friends", and you try to add those friends to every paper you publish for some very minor help/advice they offer, and your friends in turn try to add you to every paper they publish for very minor help/advice from you. This has become an effective method to boost publication count and thus ability to obtain funding without doing any real hard work.

It's because biologists, like social scientists, are too busy massaging their data until it yields the intended result.

I've seen these sorts of things happen, and had one paper where the author list was a subject of much debate. Though it seems much more common in the life sciences.

I'm pretty sure the advice given was sarcasm.

Reply on news.ycombinator.com