mercredi, janvier 04, 2012

SOS! A statistical emergency!

I usually take vacation during unusual periods of the year. My last vacation was during October, to compete in the national cycling championships. As a result, I have no more leave days left for this year, therefore, no Christmas vacation for me. The upside is that the workplace will be much quieter and I’ll have the chance to focus on some work I’ve been neglecting for lack of time.
At work, we have a certain arrangement during long vacations. For every profession, we should have at least one person present at work, to ensure permanence for all studies, so if something urgent comes up, the statistician should have to deal with it.
I was wondering the other day, as sipping coffee with MMMMMMMMM at Starbucks Cappucines in Paris, what a statistical emergency could be.
Is it an urgent sample size calculation for an urgent recruitment of patients in an urgent project, studying concepts and aspects of public health that no one cares about, except the eventual authors, to be published in a high impact factor medical journal?
Or is it a p-value that gone missing.
The point is that, in research, things move in a very slow pace and changes don’t really occur on a daily basis but rather monthly.
That was actually some sort of a shock for me when I started working in that area, especially after I spent about 4 months in 2006 working in the humanitarian field in an emergency context, where I had the chance to test my qualities as “baroudeurs”.One of the differences between working in emergency and working in research is that, for the latter, the impact on society is not directly tangible and is more for the longer term. Therefore, one might feel his/her work is not necessarily useful nor rewarding because it’s important, whenever we achieve a certain job, to be able to see the end product, which is not always the case in research.