In the past weeks I've been visiting quite the number of jobs fairs, networking events, trainings on how to hunt for jobs and the like. I certainly learned a lot, albeit mostly things that are obvious after a moment of contemplation or come to you by common sense. Yet a simple reminder and a bit of practice are surely beneficial. High in demand are all these abstract skills we all have been told one too many times to include on your CV (complex and analytical thinking ...) and programming skills (although many employers seem to be happy to train on the job). What somehow fell of my radar and which came up more than once, is a basic knowledge of statistics. This leaves us in a bit of a pickle because most available material seems to fall into one of the two following categories: 1. Statistic for statisticians: Something we all have very actively decided against studying. All the proofs and constructions and endless excursions into theoretical concepts that haven't yet and maybe never will make their way into practice. Scientific exploration that is fun if and only if you have a thing for statistics. 2. Statistics for undergrads who don't even know if they need it, lectured by junior-professors or assistants who rather spend their time on research than on preparations: What you get is a incoherent bunch of powerpoint slides that seem to live more through examples than concise explanations and miss out on the most important part, namely a self-contained overview over available methods. To the rescue comes Arnaud Delorme, a computational neuroscientist from the University San Diego. His paper on statistical methods does basically everything you want on 23 pages: A quick recap of the necessary definition, a clean-cut overview over the most common methods and a short concise explanation for each of them with just the right amount of theoretical background and a short demonstrating example. Sure, towards the end it can be a bit tiresome to go through method after method and you won't want to read it front to back in one go. Of course I also can't guarantee for its exhaustiveness, but it seems like we have a winner when it comes to ratio of usefulness over required time-investment.
4 Replies to “Statistics, and a neuroscientist to the rescue”