Message from @ping
Discord ID: 407324721986666497
there are simpler experiements in behavior science that have more predictable results, complex ones less so
does .3333333 repeating = 1? 😄 😄 😄
you want to present broad correlations
correlation isn't causation and all that :p
yep
but psychology is a really tall mountain
and not everyone that makes important decisions can climb it
@LotheronPrime no it's simply equal to .333333 repeating
to put it simply with a medical example
If a person is sick and gets Medicine A
Medicine A cures 60% of the people
40% are unaffected
then medicine isn't a science because theres no guarantee that "giving a person medicine A" will cure him
i mean since nobody is dying then if you are making life and death decisions then you want to use the medicine
hopefully someone can come around and figure out why 40% are unaffected and filter them out of those qualified to take it
@ping, sorry .33333 repeating x 3.. as in 1/3 *3 and all that,nevermind
leerooooy
i'm more about the actual application in society than proving things concretely
and that figure out why 40% are unaffected is irrelevant, in modern medicine these sorts of random % of X happening is so common they even place it on the medicine "May cause unintended side effects"
chemistry / physics need to be absolutely perfect and concrete in order to help, which is why it's so strict
yes but thats what makes them sciences
getting a correlation wrong by 10% in sociology is usually not that big of a deal
yes, because its not following the scientific method
its following statistics
you're using experiments to collect the data
experiments being methods of collecting your data
yep, good data and good experimentation lays credit to the correllation, no one will want to say its 100% truth, but can be confident nonetheless, thats good science
as in selection strategies and in the case of something fuzzy, a classification algorithm
@pukeblood i think we should define "good science" as something that has an actual use in society that offsets it's cost
well its hard to say if something is good if its, say, new
supporting hypothesis are needed to make a theorem in any science
nah, he means a reliable conclusion when he says "good science" he doesn't mean how much it matters to society
like traffic studies that reduce road deaths by 1% is still going to be good
oh yeah nah I dont mean societal benefits
but instead that its drawn from reliable experiments and data to draw an accurate conclusion
implementation of good science will probably, *usually*, result in good benefits
yeah but like
i can prove my brainfuck theories for decades and decades and waste countless hours of other people's time
so can anime
thats not a problem
if its scientifically proven, its still good
the issue is also my hypothesis can be so loose that it would always be a hit
then its a bad hypothesis 😛
youd get called on your shit immediately