š» John 
Zach Weinersmith
Morten@ZachWeinersmith Haha, yeah *social* science. Surely no other science would have this problem.
Leo RĆ© Jorge@drgroftehauge Yeah, came here to say the same... All science is a social endeavour, so I guess only in the sense every science is social can you say this is a social sciences problem. Can start with P < 0.05 there, and all of statistics falls into this trope.
theMimic@drgroftehauge @ZachWeinersmith
Haha, what do you mean. I NEVER uncountered such a thing in algorithmic, physic or electronics š but just in case it happen some days (it will), I keep the comic available. It is a great one page summary to anyone doubting the metric issue.
Isaac Erbele, MD@ZachWeinersmith Big yikes.
Not a Spring OnionFor the Body-Mass-Index (BMI), it took only 190 years to get to step 5:
Janelle Shane@wakame @ZachWeinersmith I was just thinking of the BMI on this one!
Fits the bill all the way to the ad hoc Victorian beginnings (Lambert Adolphe Jacques Quetelet was trying to find the average human, and based it on French and Scottish men)
@janellecshane @ZachWeinersmith
And it makes sense as a simple rule-of-thumb when you want to check your male soldier-apprentices of roughly the same age for who should be looked over more closely by a doctor.
But otherwise, BMI is more of a test for who approximates a quadratic mattress best.
Propriety
R. Brent Stansfield@ZachWeinersmith Brilliant (p<.05)
Supportive Tree-based lifeform@ZachWeinersmith
this needs to be taped to cabinet in most science classes.
this needs to be taped to cabinet in most science classes.
Lizard@ZachWeinersmith This is how buggy libraries sit at the core of thousands of applications,because itās impossible to rewrite the code anymore.
Justin Tracey@ZachWeinersmith Sans the Victorian part, this is also how empirical CS works in my experience (the SZZ algorithm in particular comes to mind -- it's the most common way to link software bug fixes to bug causes, despite having repeatedly been shown to be in the category of "better than guessing, but not good enough to draw conclusions from").
Tom Forsyth@ZachWeinersmith Life goals for every statistician! Although inevitably Stigler's Law of Eponymy means it'll be named after someone else.
Paul Chernoff@ZachWeinersmith Its like how sample code is used in production and we end up with routers lacking security.
Luke@ZachWeinersmith see also: Net Promoter ScoreĀ® (though to be fair, its creator did hype it up as a legit metric before later recanting on his deathbed)
(okay he's still alive but might as well be dead for all anyone listens to him now that his monster's out in the world)
soaproot@luke @ZachWeinersmith I once killed off my tiny startup's urge to use Net Promoter Score. I mean, I didn't really know what it was or even that I was killing off anything. But the founder was like "oh I guess we should build this thing which asks people questions after they use the site" and I was like "I don't know if I'm the target audience but speaking for myself I'm kind of burned out on those surveys" and it never came up again.
Virginia S. O'Possum
Jonathan Horowitz@ZachWeinersmith I kind of like null hypothesis testing in some ways, but IIRC this is a near perfect reconstruction of the history of p-values.
Steve Davidian@ZachWeinersmith
I'd add "race" and "IQ" to the list of invalid concepts.
Great comic.
Cian O'ConnorGDP. Of course the damage from that one extends far beyond social sciences.
ryaninteractive
Mike Babyak@ZachWeinersmith @healthstatsdude hahaha. Precisely the story of categorizing continuous variables. Done initially because calculations were done by hand but everyone knew it was less than ideal. There were even papers showing how to try to correct for the lost information. Within a generation it became a norm despite being no longer necessary as computation improved.
1/
1/
@ZachWeinersmith @healthstatsdude I used to tell the joke about making the holiday roast. Mom cuts the end of the roast off before it goes in the pan. Kids ask why she does it. Mom says her mom did it that way. Kids ask grandma why she did it. Grandma says thatās how her mom did it. Kids ask great grandma. She says, when she was first married they only had a small pan, so she had to cut the end of the roast off so it fit. 2/2