No, what’s racist or sexist is to suggest that a genetic difference in trait distributions, observable only at the population level, has ANY application to observations about individuals or the laws that govern them. #EcologicalFallacyhttps://twitter.com/EPoe187/status/1092350710681866240 …
If I know that giraffes are taller than gophers on average, and someone tells me animal X is a giraffe, can't I use Bayesian reasoning to predict that they're likely to be taller than a gopher? If not, why not?
-
-
The difference in population trait distributions between human sexes and races is miniscule, and the difference in population trait distributions between giraffes and gophers is enormous. But then again, you already knew that. Enjoy your fellow travelers here, dude.
-
What d value concerning mean differences is required before Bayesian reasoning is acceptable to you? Give me a number.
- 4 more replies
New conversation -
-
-
I asked a similar question (seems like bad reasoning to ignore base rates) and was told I had forwarded the bullshittiest non sequitur in history, which made me proud for having at least accomplished something profound in my life
.Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
When inter-group variance exceeds intra-group, then yeah, make default assumptions about individual cases based on group membership.
-
*while leaving room to be pleasantly surprised by outliers.
End of conversation
New conversation -
-
-
I would guess it is easier with physical attributes than things relating to complex behavior. For the latter, one deals with distributions where variance dominates differences in averages, and systematic errors on the latter are poorly controlled even when they are statistically.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Why Bayesisan? Wouldn't you use Bayesian instead to predict that an animal taller than a gopher is likely to be a giraffe?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The answer is 'no', because Bayesian reasoning is just for them hi-faluttin' acadeemics.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Only if the distribution is Bayesian and the measurement is as accurate and precise as measurement of length.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Geoff, the difference is that choosing based on Bayesian inference when the difference is race, gender,.. on cognitive traits subject to high neuroplasticity, you may actually be increasing the bias that caused the population difference. It is a reinforcement algorithm.
-
The problem is that we know many of the genes for height and understand the limits of environmental influence fairly well. But we have been very wrong about neuroplasticity, have bee oppressive because of it, and still pretend too much about genes.
- 1 more reply
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.