Another aspect of open notebooks that has troubled me of late is the prospect of getting scooped. When initiating this exploration, I knew that this was the primary concern of most scientists when the idea of an open notebook is proposed, and I allayed my own fears by convincing myself that: (1) open notebooks themselves provide proof of priority of a sort; (2) anyone seeking to mine my postings for scoopable material would likely remain a step behind me, since they’d only get access to my work only up to my latest posting; and (3) being scooped may be an acceptable cost-of-business if the benefits of open notebooks become more prevalent. Of course, (1) is only really relevant for future historians of science, since within the annals of science itself peer-reviewed publication is (and should?) be the standard for priority. (2) is also somewhat dubious in that it necessitates intentionally delayed posting of material, which somewhat violates the spirit of openness. So lately I am left with (3), which again necessitates elevating an long-term/abstract gain against immediate concerns of cost.

On the other hand, I recently experienced a great example of the efficiency of openness while attending the 50th annual meeting of the Psychonomics Society last week in Boston. The poster sessions were a great opportunity to interact with other scientists and share ideas about each other’s work. I’ve hopefully drummed up a couple collaborations out of the experience. I did, however, feel a pang of regret immediately after pointing out a flaw in one poster and suggesting to the author an easy solution. I think it was the fact that the solution was easy, so easy I could have done it myself, that made the interaction feel like a missed opportunity. On the other hand, while it would have been easy for me to implement the solution, it would certainly have been even easier for the poster’s author to implement it, since he obviously had most of the experiment already coded. Hopefully by fostering an atmosphere where scientists value the long-term efficiency of openness over immediate potential gains through closed competition, we’ll speed the progress of science.

In a similarly vein, hopefully down the road I don’t look back on sentiments like this as exemplifying the naïveté of youth!

I’ve also uploaded code and resulting graphics from an attempt at replicating Miller (1988)‘s simulations demonstrating the dangers of using the median in the context of skewed distributions (ex. human response time data). The code demonstrates Monte Carlo simulation in R using the plyr package and the use of ggplot2 for graphics.

Subscribe to the blog RSS feed and stay tuned!