I wanted to write the subject of this post for long time. The death of science the way I grew up seeing it since I was a kid and the rise of the era of information processing. What made me sit and write this time is a brilliantly written article by Chris Anderson, editor- in- chief of Wired magazine and author of the popular blog “thelongtail.com”
Anderson noticed in this article that appeared in July 2008 issue of Wired that the approach to science in its old form which is “hypothesize, model, test” is becoming “obsolete”. We live in the age of “Petabytes”. We collect data. We find correlations and….Well, that’s it really. We are in a less need for models to experiment on. Data is paramount and “correlation is enough”. That reminds me of few facts. Generally I find reading a new study or article a bit painful. nearly all the studies in scientific and medical journals are full of complex statistical analysis that tries to find appropriate “correlations”. The only study I really enjoyed reading over the past 12 months is one published by Nature in their August 2 issue 2007 and was about using DBS for improving the functionality of a patient with a minimal consciousness state following a severe TBI. There was not much statistics. It was a great example of “hypothesize, model, experiment” form. A real joy to read.
A good example of what I am talking about is this study that was published last month in Neurosurgery (well, actually it is their May issue that is released in July!) about the correlation between subarachnoid haemorrhage and seasonal changes in Dusseldorf area in Germany. Authors found after analysing a sizable set of data that SAH incidence peaks in springtime. No satisfying physiological explanation was give