6 ms·

I wish the table of underlying data was available so that a logistic regression could be run to see whether variables like campaign contributions do indeed pred

I wish the table of underlying data was available so that a logistic regression could be run to see whether variables like campaign contributions do indeed predict support for SOPA after age and number of years served have been partialled out.

tst 15y ago
The data is pretty easily accessible as JSON - however, here's the data I used: http://pastebin.com/nRYv40U8 http://pastebin.com/nRYv40U8 Firstly, a boxplot with the quotient of entertainment contributions to entertainment & internet contributions. http://i.imgur.com/FWQWy.png http://i.imgur.com/FWQWy.png You can see quite easily that there's a difference which is also significant (95%, t = -4.73). I've also done a logistic regression correcting with age, party (is_democrat), seniority and quota of contributions (quota_ent). ------------------------------------------------------------------------------ support | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0258551 .0358136 0.72 0.470 -.0443382 .0960485 is_democrat | -1.252883 .6243361 -2.01 0.045 -2.476559 -.0292067 seniority | -.0262688 .0381962 -0.69 0.492 -.101132 .0485943 quota_ent | 5.839435 1.447732 4.03 0.000 3.001933 8.676938 _cons | -1.968467 2.01512 -0.98 0.329 -5.918029 1.981096 ------------------------------------------------------------------------------ The AUC is 0.8089 which is quite okay. Furthermore, it would be interesting to test whether location is a significant factor. Edit: @adamtaylor: Here's a scatter plot with each contribution, transformed with log(1 + x) for readability: http://i.imgur.com/MRciL.png http://i.imgur.com/MRciL.png
- _delirium 15y ago
  Hmm, that scatterplot suggests a less straightforward relationship; the for/against dots look like they plausibly came from slightly different distributions, but not very much different ones.

tst 15y ago

The data is pretty easily accessible as JSON - however, here's the data I used: http://pastebin.com/nRYv40U8 http://pastebin.com/nRYv40U8 Firstly, a boxplot with the quotient of entertainment contributions to entertainment & internet contributions. http://i.imgur.com/FWQWy.png http://i.imgur.com/FWQWy.png You can see quite easily that there's a difference which is also significant (95%, t = -4.73). I've also done a logistic regression correcting with age, party (is_democrat), seniority and quota of contributions (quota_ent). ------------------------------------------------------------------------------ support | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0258551 .0358136 0.72 0.470 -.0443382 .0960485 is_democrat | -1.252883 .6243361 -2.01 0.045 -2.476559 -.0292067 seniority | -.0262688 .0381962 -0.69 0.492 -.101132 .0485943 quota_ent | 5.839435 1.447732 4.03 0.000 3.001933 8.676938 _cons | -1.968467 2.01512 -0.98 0.329 -5.918029 1.981096 ------------------------------------------------------------------------------ The AUC is 0.8089 which is quite okay. Furthermore, it would be interesting to test whether location is a significant factor. Edit: @adamtaylor: Here's a scatter plot with each contribution, transformed with log(1 + x) for readability: http://i.imgur.com/MRciL.png http://i.imgur.com/MRciL.png

_delirium 15y ago
Hmm, that scatterplot suggests a less straightforward relationship; the for/against dots look like they plausibly came from slightly different distributions, but not very much different ones.