I was expecting the libertarian quadrant to be a lot more populated.
Quote Tweet
Sexual fetishes on the political compass, men and women. Total sample size was over 19,000!
Show this thread
116
88
2,006
Conversation
there's basically two things to pay attention to - how strong is the correlation, and how significant it is?
The strength is like... "are tall people heavier?" There's a strong correlation between height and weight.
1
6
generally in this type of loose research with tons of variables that impact stuff, <0.2 is a small correlation, and stuff 0.4-0.5 and higher is really big. A correlation of 1 means it's basically the same thing - it would mean height and weight are so correlated theyre the same #
1
1
4
if I ask 10 people their height and weight, and it's not correlated at all (we can run a test to check the correlation number), we have a few questions - did we just randomly draw a 'bad' 10 people to check? If we gather more people, will we see a correlation?
1
6
this is what significant means - its basically a test (theres diff kinds of tests) to ask the question "how likely is it that this correlation is showing up by total chance, and how likely is it it measures something real?" The bigger your sample size, the less it's chance.
in this survey I did, the sample size was huge - tons of people responded. But the correlations I found were really weak (like, authright people are only very teenyslightly more likely to be into raceplay). But despite being weak, they were *significant* - unlikely to be chance.
5