Wednesday, March 01, 2017

Reproducibility!

I guess it is old news but it only caught my eye today at the BBC website. Apparently Nature did a survey last summer (2016) were scientist were asked if they had "failed to reproduce another scientist's experiments"? What really surprised me was that only 70% on average said yes to that question. Obviously, scientists from all disciplines participated in the survey and chemistry did better in reproducibility than biology and medicine (when judged by themselves) but it looks pretty bleak overall. In synthetic organic chemistry, I would estimate that >95% of chemists have been unable to reproduce a published synthesis (as in get any of the desired molecule and not necessarily just the same high yield). Personally, I have had methods across the impact factor landscape fail in my hands, from Synlett to Nature Methods. On more than one occasion everything has just decomposed and gone black. Unless it is published in Organic Syntheses it's a bit of a lottery. Anyway, both articles are quite interesting and worth reading (although they are somewhat depressing). D! 

3 comments:

Anonymous said...

And chemists are champions in not reproducing their own research!
As chemistry in general is more reproducible, I see two reasons: a) chemists are more honest; b) some fields (ahem, biology) do not even try to reproduce the most exciting results

Unknown said...

I think the real reason why chemistry is more reproducible is that someone is actualy trying to reproduce someones work.

Anonymous said...

Ha! Well.... I have discovered that many organic chemists lie about their yields. I am trying to finish up a PhD in Pharmaceutical Chemistry and fell behind struggling With reproducibility in my own lab! I would love to chat with someone for advice. My one first author paper contradicts yields in my PI's previous work. I feel like I am going to destroy my career if I'm honest.