3 ms·

And like most statistics, they can be read to support almost any conclusion you want. A deeper dive on that study shows that it's not really that simple, or tha

And like most statistics, they can be read to support almost any conclusion you want. A deeper dive on that study shows that it's not really that simple, or that distressing: https://www.vox.com/2016/5/9/11502464/gilens-page-oligarchy-study https://www.vox.com/2016/5/9/11502464/gilens-page-oligarchy-... Important quote: America is an imperfect democracy, in other words — but it's hardly an oligarchy. When the rich and middle class disagree, each wins about half the time.

hedora 4y ago
The people that wrote that article don't address claims of the original study. They cite academic work that divides the country into four groups: the middle class, the rich, business interests and non-business interests. (This is not a covering set, which is already a red flag.) They show the rich and middle class agree on bill outcomes the vast majority of the time, with a 10% spread when they don't (e.g., 45% - 55%, from the article). As an example, they point out the rich are more supportive of public campaign financing than the middle class. This is a second red flag: It is unlikely that people that use their money to influence legislative outcomes are more supportive of banning the practice than the general population. The original study claimed that (1) lobbyists and oligarchs are basically writing legislation and deciding who gets elected, and (2) there is no correlation between popular opinion and actions of congress. You can't rebut that by claiming that the 90-99% usually agree with the 50-90%, and congress doesn't really seem to favor one of those groups' opinions over the other. It is as though I said that, once confounding factors are accounted for, there's no correlation between listening to rock music and catching cancer, and the rebuttal comes back as "Well, Eagles fans and Jay-Z fans both get lots of cancer, but half the outlier cancers are more common amongst Hotel California fans, so electric guitars must be carcinogens!" To rebut the original claim, the academic study would need to either reproduce the original study (showing statistical irreproducibility) or dive into causality. They could look into the actions / opinions of lobbyists and people like the Koch brothers, Soros and the ULine family. The could trace how many line items those groups are sneaking into bills, and whether the popularity of the line item matters at all to its success. (Hint: Most voters don't form an opinion on section 1.35.2.456.63356-b that was introduced at 2am the night before the 10,000 page military appropriations was passed by voice vote.) Anyway, the damage control piece was an interesting read. Thanks for the link.
- calgarymicro 4y ago
  >don't address claims of the original study They address several things, like the low R-squared value (0.074) or the fact that interest groups win against both the rich and the middle class at roughly the same rate if you use their own data sets. >They cite academic work that divides the country into four groups: the middle class, the rich, business interests and non-business interests. (This is not a covering set, which is already a red flag.) These groups are literally from the Gilens study, not the rebuttal. If you view this as discrediting the rebuttals then you should view it as discrediting the original paper as well. >they point out the rich are more supportive of public campaign financing than the middle class. This is a second red flag: It is unlikely that people that use their money to influence legislative outcomes are more supportive of banning the practice than the general population. This policy preference comes from the surveys used by the original Gilens survey, so if you view this as discrediting the rebuttals then you should view it as discrediting the original paper as well. >The original study claimed that (1) lobbyists and oligarchs are basically writing legislation and deciding who gets elected, and (2) there is no correlation between popular opinion and actions of congress. You can't rebut that by claiming that the 90-99% usually agree with the 50-90%, and congress doesn't really seem to favor one of those groups' opinions over the other. The original paper explicitly claimed that >ordinary citizens get what they want from government only when they happen to agree with elites or interest groups that are really calling the shots So it makes plenty of sense to exclude those instances where the two groups agreed if you're trying to determine how influential their policy preferences actually are.

hedora 4y ago

The people that wrote that article don't address claims of the original study. They cite academic work that divides the country into four groups: the middle class, the rich, business interests and non-business interests. (This is not a covering set, which is already a red flag.) They show the rich and middle class agree on bill outcomes the vast majority of the time, with a 10% spread when they don't (e.g., 45% - 55%, from the article). As an example, they point out the rich are more supportive of public campaign financing than the middle class. This is a second red flag: It is unlikely that people that use their money to influence legislative outcomes are more supportive of banning the practice than the general population. The original study claimed that (1) lobbyists and oligarchs are basically writing legislation and deciding who gets elected, and (2) there is no correlation between popular opinion and actions of congress. You can't rebut that by claiming that the 90-99% usually agree with the 50-90%, and congress doesn't really seem to favor one of those groups' opinions over the other. It is as though I said that, once confounding factors are accounted for, there's no correlation between listening to rock music and catching cancer, and the rebuttal comes back as "Well, Eagles fans and Jay-Z fans both get lots of cancer, but half the outlier cancers are more common amongst Hotel California fans, so electric guitars must be carcinogens!" To rebut the original claim, the academic study would need to either reproduce the original study (showing statistical irreproducibility) or dive into causality. They could look into the actions / opinions of lobbyists and people like the Koch brothers, Soros and the ULine family. The could trace how many line items those groups are sneaking into bills, and whether the popularity of the line item matters at all to its success. (Hint: Most voters don't form an opinion on section 1.35.2.456.63356-b that was introduced at 2am the night before the 10,000 page military appropriations was passed by voice vote.) Anyway, the damage control piece was an interesting read. Thanks for the link.

calgarymicro 4y ago
>don't address claims of the original study They address several things, like the low R-squared value (0.074) or the fact that interest groups win against both the rich and the middle class at roughly the same rate if you use their own data sets. >They cite academic work that divides the country into four groups: the middle class, the rich, business interests and non-business interests. (This is not a covering set, which is already a red flag.) These groups are literally from the Gilens study, not the rebuttal. If you view this as discrediting the rebuttals then you should view it as discrediting the original paper as well. >they point out the rich are more supportive of public campaign financing than the middle class. This is a second red flag: It is unlikely that people that use their money to influence legislative outcomes are more supportive of banning the practice than the general population. This policy preference comes from the surveys used by the original Gilens survey, so if you view this as discrediting the rebuttals then you should view it as discrediting the original paper as well. >The original study claimed that (1) lobbyists and oligarchs are basically writing legislation and deciding who gets elected, and (2) there is no correlation between popular opinion and actions of congress. You can't rebut that by claiming that the 90-99% usually agree with the 50-90%, and congress doesn't really seem to favor one of those groups' opinions over the other. The original paper explicitly claimed that >ordinary citizens get what they want from government only when they happen to agree with elites or interest groups that are really calling the shots So it makes plenty of sense to exclude those instances where the two groups agreed if you're trying to determine how influential their policy preferences actually are.