“My manager says he ‘gets better regressions most of the time when he does this’” 
[ht: @wytham88]https://twitter.com/randal_olson/status/1088134618723737600 …
-
now I kind of want to know what this regression does - it's almost testing how unequal the tails are of the two univartiate distributions, but that's not quite it...
I’m going to go out on a limb and say \hat{\beta} is positive. (Don’t @ me ... I included an intercept.) And its magnitude is just picking up something about relative dispersions of y and x?
-
-
Special case of y and x are both normal, I’m guessing expected value of \hat{\beta} depends only on \sigma_y and \sigma_x, incereasing in former and decreasing in latter. (“Proof left as an exercise to the reader.”)
-
I don't think beta_hat should be positive. Here's a simulation where it doesn't. https://gist.github.com/paulgp/86f67b10ed0e1e2063735999cfcef060 …pic.twitter.com/2b3phj7T6y
-
I believe it but I don’t understand it. In particular, I can’t visualize a scatterplot of (sort(x), sort(y)) [i.e., every observation is northeast or southwest of every other] for which the ols fit is downward sloping.
-
Then again,
@paulgp figured out the Tweet that would get me to finally install Rpic.twitter.com/ltg2jJiS17
-
It's pretty fun. I still use Stata for everything but R is great.
-
Be sure to use the tidyverse pacakages to do data cleaning, ggplot2 to handle plotting, felm to do panel regressions; stargazer for LaTeX output
-
If you’re looking for an intro that’s closer to showing how R is used optimally (as opposed to a beginner tutorial) I’d suggest looking at (and following along with) my friend
@drob’s livecasts:http://varianceexplained.org/r/tidy-tuesday-college-major/ …
End of conversation
New conversation -
-
-
LOL! I hereby propose a new robustness statistic that is the t-stat of regression coefficient obtained by sorting Ys and Xs independently before regressions. I call it the "Manager's t" :)
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.