1) We now have to face the fact that what is called "social science" and "evidence based" is scientifically shown to be equivalent to "bullshit vending". Except less elegant, of course.
Which brings us back to ancestral heuristics, #Lindy, #SkinintheGame
Which brings us back to ancestral heuristics, #Lindy, #SkinintheGame
https://twitter.com/HarryDCrane/status/1034130885711417345
2) 2 ½ years ago I proved MATHEMATICALLY p-values need to be lowered by at least 1 order of magnitude.
Now @BrianNosek et al, 75 authors mysteriously realize that p-values need to be lowered by.... 1 order of magnitude. List includes of course a bunch of Harvards I insulted.
Now @BrianNosek et al, 75 authors mysteriously realize that p-values need to be lowered by.... 1 order of magnitude. List includes of course a bunch of Harvards I insulted.
3) The REPLICATION CRISIS a VINDICATION for citizens SICK & TIRED of the Harvards w/a sense of entitlement treating practitioners as immature pple lost without their "science" when it is STATISTICALLY shown Harvards are no better than Ms Bré, fortune teller.
4) People don't realize the more severe problem is not absence of replication, rather the reduction of EFFECT for those that replicated.
If you lower effect by ½, the odds of being observed in REAL LIFE decreases nonlinearly. From 2 sigmas to 1 sigma, down by 86%.
If you lower effect by ½, the odds of being observed in REAL LIFE decreases nonlinearly. From 2 sigmas to 1 sigma, down by 86%.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
