I issued a brief survey of EAs on the subject of American politics.
I only received 10 responses from the forum, so I also shared it in a Twitter group chat of EAs (6 responses), then filled it out myself. (I probably could have gotten more responses by calling it a politics survey in the headline, but I did not want to disproportionately attract people who are super concerned about politics, as I think that generally produces a less healthy set of responses.)

With a sample size of only 17 this survey still has very little power but I said I would share the results, so here it is. https://docs.google.com/forms/d/1s2o1_D-CpxeihRJ8UoRj0EMufw6HHTwliqnsL9tghQY/viewanalytics 

Some highlights and takeaways.

First, across a wide variety of significant American political disputes, EAs do not exhibit polarized disagreement. That is, either we broadly agree, or our disagreement is between people who have one point of view and people who are undecided/moderate. We do not reflect typical ideological polarization. The two exceptions are "Should we cut the defense budget?" where 47% emphatically say yes and 29% emphatically say no, and "What should be the federal minimum wage?" where people are pretty widely distributed from $0 to $15.

Second, across a wide variety of significant American political disputes, my own conclusions after doing research and writing the policy platform on my website are broadly consistent or at least adjacent to the views of most EAs. The one exception here is "Should we cut the defense budget?" were I say no but half of other EAs say yes. (I intend to study and revisit my views on this issue more carefully in the near future.)

Third, few EAs say that there are practical tradeoffs between their own values and utilitarianism on practical policymaking. There are a few issues where someone has remarked that one policy may be welfare-maximizing but immoral for other reasons, but these are rare exceptions.

I think the above three points are encouraging and undermine the worry that political disputes are inherently epistemically and socially intractable for EAs.

However, there is deep, wide disagreement on fundamental questions, like how much priority to assign to x-risk as opposed to short-term issues.

I have responded to some specific comments here and will update various things in my website and model to cite and be more consistent with various people's judgments.