oh gosh I am glad they included scatterplots but gosh https://twitter.com/NicoleBarbaro/status/1134103469533618176 …
-
This Tweet is unavailable.Show this thread
-
-
-
Replying to @KizunautFI
there are at least three problems here how many can u find
2 replies 0 retweets 5 likes -
Replying to @eigenrobot
I should know, I used statistics in my thesis but all I can spot is basicaly the scatter plot is so all over the place you can't really draw a trend from there. Also shouldn't you have the N and statistical significance shown somewhere on the graphs??
1 reply 0 retweets 3 likes -
Replying to @KizunautFI
There is a lot going on here All of those things are problems "There's clearly no meaningful pattern in the scatterplot" is the most important point, if you can't do an ocular regression (thats an econometrics joke ha ha) probably don't trust your linear regression overmuch
1 reply 0 retweets 14 likes -
Replying to @eigenrobot @KizunautFI
The beta there is . . . maybe it's something unusual, but typically it refers to a regression slope coefficient The fact that it has no subscript probably indicates that the author ran a bunch of univariate regressions to compare regressors Do . . . do not do this
1 reply 0 retweets 13 likes
Her response feature seems to be ordinal (probably "how much do you like chokers? Not at all, hardly at all, . . . , I love them more than my own mum!") which is okay but a linear least squares regression with an ordinal LHS is not going to return an interpretable coefficient
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.