Neural correlates of interspecies perspective taking in the post-mortem Atlantic Salmon: an argument for multiple comparisons correction (Humor column)
David Hand's book "Dark Data" ("Dark data. Why what you don't know matters") came across this whitepaper It's about doing brain experiments on dead salmon. This experiment won the Ig Nobel Prize in Neuroscience and was staged as follows: Dead salmon underwent an MRI brain scan. At the same time, they showed a series of photos depicting people in different situations. And also asked the subject to determine what emotions are used by the depicted people Fixed the definition of emotions by response, which was defined as the activity of brain cells visible on voxels (pixels) It turned out that several voxels were active. The authors discovered that
Either we stumbled upon a startling discovery at the intersection of ichthyology and postmortem cognitive function, or something is wrong with our statistical approach. Can we conclude from this data that salmon expressed their opinion when asked? Of course not. By controlling the subject's cognitive abilities, in this case we have completely eliminated this possibility.
David Hand, author of a book on dark data and president of the Statistical Society and a member of the British Academy, recalls the anecdote.
Experimenter A tells experimenter B that he has a great problem reproducing the results obtained by B. "Not surprising," he replies, "because I too could not get their first results." 100 (b) when conducting an experiment.”
And in order to ensure that our experiments at Tinkoff were protected from this, we are building an a/b platform inside the company that I am already talking about. told when posting a job vacancy of a staff engineer in this platform. There are also a number of books that are useful for those who like statistics and experiments, which I have talked about before.
- How to Lie with Statistics (How to Lie with Statistics) - on the fingers explains how to lie with the help of statistics, and from here it becomes clear the motivation for creating a system of summing up the results of experiments
- Understanding Statistics and Experimental Design. How to Not Lie With Statistics (Statistics and Experiment Planning for the Uninitiated) This book talks about the design of experiments and the mathematics behind them.
- Confidential a/b testing (Trustworthy Online Controlled Experiments) This book also provides insights into how to build a platform for company-wide experimentation.
#Math #PopularScience #SelfDevelopment #Humor #Statistics