Conversation

Replying to
but as a basic summary: all the correlations were pretty weak, so don't go around being like 'the authright is definitely into raceplay' the closer a correlation is to the center, the more 'noisy' it is and the more likely it is to 'move' after gathering more data.
24
995
Replying to
When you say you combined the correlations via subtraction, do you mean you removed the "base" correlation between auth/lib and left/right, so the trendline is an effect above this? Since that base rate is the only reasonable explanation I can think of for the trendline...
2
Replying to
so your sources are... a reddit post and twitter polls, on asking people what they themselves correlate in their mind? That has to be one of the worst methodology ever. Instead make a form about what people think for themselves and their own alignment on the compass...
1
Replying to
Points are not how I would expect to see that data presented. Rather long lines with error bars is what I'd expect, in each axis. Isnt male/female a third axis - a short one? Isnt hetero-homo or preference for partner an axis?
Replying to
It's mostly on a diagonal line because the test looks at two variables which correlate strongly. Auth left and lib right barely exist because the test mainly graphs how pro/anti hierarchy you are. The compass test itself is bad not ur data.
Show more replies