Crazy Statistics: Margarine Causes Divorces By Brian Simpson

     In academia there is something called a replication crisis in pseudo-scientific disciplines like psychology. Part of the problem must be due to the fact that these soft academics don’t have much of a mathematical grounding and just feed their garbage into the computer program R, or if they know a bit more, MATLAB, and out pops the latest piece of bs research. Much of this waste is at the level of statistical correlations, and some truly stunning statistical tricks like bootstrapping, where constant feeding and cycling of the same data somehow is supposed to generate valid results. Anyway, here are some sites that show a whole range of spurious correlations, like margarine eating and divorce rates, and that US spending on science, technology and space, correlates with suicide by hanging, strangulation and suffocation. I imagine that doing psychology also correlates with a range mental illness, but that may not be spurious.

     Real science, as in physics and chemistry, does not solely rely upon statistical bs correlations, but seeks out explanatory hypotheses, that have qualities such a simplicity, coherence, and suggest experimental testing to enable spurious correlations to be replaced by causation. The psychology, let alone sociology I have seen, does not reach this basic level of scientificity, and I think it needs defunding.



No comments made yet. Be the first to submit a comment
Already Registered? Login Here
Wednesday, 24 July 2024

Captcha Image