I read an interesting article in this week’s edition of The Economist, which discusses reforms on the use of animals for scientific research (see the complete Economist article describing the details of the European Parliament vote here). While the article goes into great detail about the changes, I was particularly drawn to a paragraph that describes ways of reducing the amount of animals used for laboratory research. In a nutshell, the article suggests that scientists should share negative data in addition to positive data, with the idea that this will prevent other scientists from repeating failed experiments and ultimately use fewer animals. In theory this idea makes sense. In practice, however, this is quite a difficult task.

Let’s simplify things.

Proof of the theory: Here is my estimate of how many animals we can save by sharing negative data. In any given year, let’s say a scientist will generate at least 50% negative data (total experiments defined as e). Most scientists have at least one competitor carrying out similar experiments, and let’s estimate that roughly 10% of their experiments will be similar. So… (negative data)e x (similar experiments) = 0.5e x 0.1 = 0.05e. 5% of a scientist’s total experiments will be repeats.

It never leaves the lab: While it is clear that sharing negative data can help curb animal use, there aren’t many platforms for negative data sharing. Negative data rarely makes it beyond the realms of lab meeting, ending up in the same place as those mysterious missing socks. Only in a question/answer session may you have the chance of hearing about a failed experiment. But there is hope. Movements such as open notebook science are a great way for scientists to communicate both positive and negative data. A caveat to this, of course, is you have to be willing to freely share your primary data prior to publication. There are also sites like Nature Precedings, which allow scientists to make their negative data available to the public without showing all of their primary work (and it’s citable). As described here on the Nature Precedings FAQ page, “Submissions detailing negative results are welcome providing that they are likely to be of interest to other scientists.” Thus, scientists are able to share their negative data without risking getting scooped (on the positive data, that is), while limiting the number of duplicated experiments among competitors.

So, perhaps this is the right path to reducing numbers in animal research. Hopefully, as sites like Nature Proceedings and other data consortiums grow and their audience begins to read and cite negative data, they will aid in more directed and less wasteful science.