Elites, Interest Groups, and Average Citizens

[aphar]

Gilens, M., & Page, B. (2014). Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens Perspectives on Politics, 12(3), 564-581
doi:10.1017/S1537592714001595 PDF

An interesting study with a pessimistic conclusion:
...policymaking is dominated by powerful business organizations and a small number of affluent Americans...

prompted morons to claim
American democracy is a sham... basically similar to Russia or most other dubious 'electoral' 'democratic' countries.

The paper is relatively short - 18 pages, please read it!
They start with honestly admitting that their work is tentative and preliminary.
They consider 1779 policy issues for which they have national surveys with income breakdowns for respondents and build a model predicting legislative response to the issues from 4 attributes:

1. opinion of the general population


2. opinion of the "affluent" respondents (top decile, household income above $146k in 2012 dollars)


3. organized interest groups
   1. mass-based
   2. business-oriented

What they discover is:

1. Individually each variable has a high statistically significant effect on the legislative outcome


2. Correlation of the first and second variables is 78%


3. Other correlations are negligible


4. All but the first variable have a significant effect on outcome


5. R² is 0.074 when all interest groups are combined into a single variable (i.e., 3 variables are used for modeling) and 0.07 (smaller!) when all 4 variables are used.

My conclusions:

1. The extremely low predictive power of the models casts a huge shadow of doubt on any possible practical conclusion


2. The high correlation of popular and affluent opinion means that
   
   
   1. the methodology is unsound: instead of these two variables, one should use the sum and difference (i.e., orthogonolize them)
   
   
   2. the conclusion made by morons that "the poor don't matter" is unfounded


3. the research is very interesting and should be continued:
   
   
   1. I would try adding a variable which indicates whether the outcome is policy change or continuation (the authors mention strong status quo bias themselves)
   
   
   2. I would love to see something similar done for other countries
   
   
   3. I think the definition of "affluent" is too inclusive (2% is more reasonable than 10%), but I am afraid it is impossible to get the survey data for the truly affluent


4. The best counterargument to the alleged irrelevance of "common man" to the American polity is the election of Trump: he was opposed by everyone: all business groups, all elites - political, intellectual, scholarly, but he still won. Admittedly, this is just one example, and, alas, it is a high price to pay (4 years of incompetent erratic egomaniac in the white house), but it does debunk the claim that the US is an "oligarchy".

PS. Comment here or on fb.