What do you think about using p<0.05 hypothesis tests?
>>7756693
pretty gay imo
nevertheless pretty easy
my sample includes 200 19 year old psychology majors from the art department
>using meme-values
>2016
>>7756693
Something is wrong with her leg
>>7756693
5Sigma or noting
>>7756732
>5 sigma
>not using a system based on non-axiomatic mathematics developed by ayyliums to find the objective truth of the universe
T O P K E K
O
P
K
E
K
>>7756732
5sigma is just as arbitrary as 0.05 (~2sigma)
Is this a frequentist vs. bayesian thing?
I side with bayesians, so fuck the hypothesis thingy.
>>7756693
I like it because it's the easiest way to justify if some difference is significant or not.
A lot of the theoretical assumptions behind t or z-tests are rarely met on actual data. And even when they are the conclusion given by low p-values are often not especially interesting ("Wow the effect isn't 0! Isn't that great?").
Also it's problematic since there's like a million research projects going on using these tests, and those with non-significant results never get published, so we really get a bad grasp on the likeliness that any significant result is purely coincidental. If it is doable journals should ask researchers to replicate the result before publishing it.
>>7756875
>fuck the hypothesis thingy
>thingy
the most-faggot word ever written
>>7756693
dat waist makes me wet
>>7756910
yup, science is all about faggot words
go back to... your trailer? mothers basement?
whatever.
also: english is not the native language of all people on earth ... faggot.
>>7756693
Practically, the line has to be drawn somewhere. If not 0.05, then where?
Man, she's skinny.
>>7756693
Alpha=.01 masterrace.
.05 fags can't into post-hoc analysis, and just do it to seem cool.