A correspondent sends me the lovely FiveThirtyEight site, on how to p-hack your way to scientific glory.
Should I be proud or ashamed that it only took me 30 seconds to get a 1% p value? I fault it though for much too modest an effort, compared to many papers I have read. Include judges, mayors, state legislators. Measure performance with levels, growth rates, unemployment, inequality, demographic breakdowns, house prices, health measures, investment, exports and more. Control for far more than recessions — exchange rates, or all the other outcome measures. Let each variable take its turn on right and left hand sides. Instruments…
Have fun
