i'm really frustrated that the NYT as an organization has been actively conspiring to protect child rapists - trump, musk, summers, etc. - for over a decade

but, hey, you go ahead and be frustrated that people are frustrated about that

by Amanda Marcotte

whoa, slow down there. this is dizzyingly complex

i can see a reality in which some people can work their way into the top 10% - though admittedly, the chances of that are vanishing with each passing year - but there is no real way for someone to honestly work their way into the top 1%. you're either born into it or purely lucky/exploitative

i suppose that depends on what one considers wealthy. from this forbes article, to be in the top 10% requires "only" a networth of over $1 million, while being in the top 1% is for those whose networth is over $11 and a half million.

www.forbes.com/sites/jackke...

the mechanism for this reality is surprisingly simple, too:

when you are wealthy, you pay people to solve all your problems for you. you never have to learn anything or even to maintain any previously-learned things

wealth is an obstacle to learning

i suspect this because in the text of their actual article, they alternate between cited standards of acceptable inter-rater reliability scores. also, because of just how circuitous i had to be when trying to reproduce their reported output

i never could figure out what columns TOR21, TOR22, and TORAVG are supposed to actually represent, but my suspicion is that they are left in columns from earlier attempts to try and see if they had sufficient inter-rater reliability (which, they do not) to justify their "content meta-analysis"

oh, you have no idea. it took me quite a while to try and piece together how what was described in their published article corresponds to the data file they shared, because it is so poorly described. which itself might be a strategic incoherence to help hide their ulterior motives

unlikely. the numbers in columns FSR1&2, RBR1&2, RER1&2, NRR1&2, and NWN1&2 simply represent a binary 1 or 0 to indicate whether qualitative coders coded presence (1) or absence (0) of codes abbreviated FSR, RBR, RER, NRR, and NWN. TOR1&2 seems to be simple summation of the preceeding 5 columns

their only provided supplemental material is a .xlsx file, which is what i downloaded and took a screenshot of

wouldn't something like that result in consistent issues across the spreadsheet, rather than the inconsistency seen in the screenshot?

e.g. some instances of correctly averaging 0 and 1 to get 0.5 in the same column as averaging 1 and 1 to get 0.5 (and averaging 1 and 0 to get 1 as in line 21)

but i am petty, and don't want woodley at al to skate by without anyone mentioning publicly that they have very likely engaged in scientific fraud in an attempt to defend their racist pseudoscience from legitimate critique

now, bird et al's critique is thorough and wonderfully unravels the strategic incoherence of woodley et al's bullshit (in the frankfurtian sense), so throwing on an accusation of data tampering probably doesn't contribute much more, and risks the real message getting lost in sauce

i do think the authors were exceptionally kind in their critique of woodley et al's content analysis

if you dig into the details of woodley et al's supplemental data, there's evidence suggestive of data tampering

for example, they average 1 and 1 together to get 0.5 multiple times, seen here