Open notebook science

Akshat recently blogged about science blogging as a new frontier in communicating science to the public (including both scientists and non-scientists). A related recent development, that I’ll talk about more in this blog post, has been open notebook science (Wikipedia entry, see also this Wired article).

Open notebook science is a form of scientific practice where the scientist or researcher notes down each step of the discovery process and experiments, including the failed experiments or the insights/steps that did not lead to anything substantive. This helps tackle problems such as: the file-drawer effect (where researchers decide not to send certain results for publication because the results aren’t interesting, contradict established results, or just support established results without adding anything new), publication bias (where publishers are likely to accept research that produces certain kinds of conclusions as opposed to other research), and the absence of specific details and raw data in published papers that makes it hard for others to attempt to reproduce the results.

The UsefulChem project in the Bradley lab at Drexel University, headed by Jean-Claude Bradley, is one example. They have a lab notebook hosted by Wikispaces, which gives details of their experiments. In addition, the same research group coordinates an Open Notebook Solubility Challenge, whose goal is to get accurate and verifiable information on solubilities, backed with details of experiments that were done to compute these solubilities. The corresponding lab notebook, again hosted by Wikispaces, is here. UsefulChem also have a blog, that they use for more general observations. The ONS wiki also has links to similar open notebook science efforts by others. In addition to using a wiki, they also make their solubility and other data available through Google Spreadsheets (that can be downloaded, viewed online, or queried using a web-based API) and in archival formats, and periodically release print versions of the solubility data as books via Lulu.

In February 2010, Science Commons and Microsoft Research teamed up to hold a conference on open data sharing in the sciences — a summary blog post with links can be found here. The first session included presentations by Jean-Claude Bradley of the UsefulChem project mentioned above, as well as one by Cameron Neylon (blog), also an open notebook scientist. Video of the first session is available here. Bradley’s presentation gives some examples of how open sharing of data and experiments helped resolve controversies, such as the NaH oxidation.

Open notebook science proponents have proposed the Panton Principles for sharing raw data of experiments. These principles include the release of all raw data under a public domain-type license that allows unrestricted use of any kind. The two licenses they suggest are the PDDL (Public Domain and Distribution License) and CCZero (Creative Commons Zero Waiver).

There are very few examples of open notebook science principles being applied outside experimental fields. In mathematics, there have been a bunch of experiments in collaborative solution of specific mathematics problems, called the Polymath projects. These have had mixed success, and offer promise for the future, but are very different from the “open notebook science” approach of carrying out all one’s thinking in the open. One researcher maintains a lab notebook in differential geometry (aptly titled Deferential Geometry) but the trend doesn’t appear to have caught on. Also, a bunch of researchers, mostly in higher category theory, started a n-Category Lab which serves as a hybrid between a lab notebook and a reference resource for their topic.