Tag Archives: dark data

Dark Science: on freeing the negative results

Hi all. Sorry for the long silence. My dissertation is defended – I passed. I’m qualified as a PhD!

So this whole presentation and dissertation thing reminded me of some work I did a few years ago. I got it published after a struggle, but one reviewer specifically recommended against publication because it was a ‘negative result’. I showed that a peptide implicated in Alzheimer’s disease does not affect synaptic vesicles. It seemed like al ikely hypothesis at the time.

I think a “Journal of Negative Results” would be a good idea. There’s been some rumbling about this on the net. Wired magazine did a great piece on it called “Freeing the Dark Data”.

Well, in any case, there is one, now, it turns out. The Journal of Negative Results in BioMedicine publishes studies that show that show things that are useful in the sense that they say “don’t bother trying this hypothesis. We already tested it.”

That’s a simplification, of course, but from a researcher’s standpoint, things in that spirit have tended to go unpublished. Articles like this make it sound like scientists are hiding data that doesn’t suit their fancy. Really, if they have some hypothesis (“I’ll bet drug A will work a lot better if we include drug B”) and then they test it and find out it’s totally false, it feels like failure. And it’s hard to publish. Journals don’t want low-impact articles like “Drug B does not change the activity of Drug A.” They want titles like “Drug B enhances effects of Drug A by 1000%”

Plus, it seems like scientists should know what they are doing. Why were we wrong when we predicted that Drug B would enhance Drug A? Were we being foolish? Nobody wants to look foolish.

So maybe with this kind of journal, this will start to change. This is good for science in the long run. If somebody, later, reads that the fact that Drug A and B don’t affect each other, it might have huge implications that are not obvious now. Why repeat the experiment? If the data is out there, that’s to everyone’s benefit. It even seems like Google (“don’t be evil”) is getting on board.