Remembering that PolitiFact actually reminds me to be more worried about this. My impression is PolitiFact started off reasonably neutral and then veered into being a partisan mouthpiece.
(I am maybe less worried about a version that has more specific goal of “decide between candidates you plausibly like”, but, it’s the sort of the thing that would have a natural tendency to turn into an org, and then get audience-captured)
Random aside: I did recently find out about a thing called “Center for Effective Lawmaking” that seems to rate legislators based on how well they accomplish the policies they (or their party?) set out to do. I haven’t looked into it at all but it seemed like another angle from PolitiFact of “past example of someone trying to do the thing”.
I did recently find out about a thing called “Center for Effective Lawmaking” that seems to rate legislators based on how well they accomplish the policies they (or their party?) set out to do.
That sounds very useful, but it also sounds like an intractably difficult thing to figure out, even putting aside the issue of motivation that you brought up.
Suppose I’m a politician with a single issue that I care about. Is sponsoring ten bills on that issue a plus even if none of them get passed? But what if I’m in a party facing a 49-51 minority, and no bill I sponsor is going to get passed regardless of what I do—is there no way to evaluate me, if non-passed legislation doesn’t count? What about my ability to shift the Overton Window, or talk my party members into supporting my position in the future? I would imagine that the targeted function looks like...
...where policy_alignment is some approximation of the U.S. government’s alignment with my stated positions, and policy_alignment' is the same, but in a counterfactual world where my opponent won the general election, such that my political impact across the entire collective future of America is computed. But that’s a very difficult function to build a proxy for, because, even putting the credit assignment problem aside, who can say whether a line item giving my issue of choice $50 million in federal funding is going to have more long-term influence than a high-profile showdown that wins me no friends on Capitol Hill but forces the next president to adopt it as part of his platform, let alone backing down quietly on the line-item in exchange for a favor from a senior congressman that I can use later on?
Remembering that PolitiFact actually reminds me to be more worried about this. My impression is PolitiFact started off reasonably neutral and then veered into being a partisan mouthpiece.
(I am maybe less worried about a version that has more specific goal of “decide between candidates you plausibly like”, but, it’s the sort of the thing that would have a natural tendency to turn into an org, and then get audience-captured)
Random aside: I did recently find out about a thing called “Center for Effective Lawmaking” that seems to rate legislators based on how well they accomplish the policies they (or their party?) set out to do. I haven’t looked into it at all but it seemed like another angle from PolitiFact of “past example of someone trying to do the thing”.
https://thelawmakers.org/find-representatives
That sounds very useful, but it also sounds like an intractably difficult thing to figure out, even putting aside the issue of motivation that you brought up.
Suppose I’m a politician with a single issue that I care about. Is sponsoring ten bills on that issue a plus even if none of them get passed? But what if I’m in a party facing a 49-51 minority, and no bill I sponsor is going to get passed regardless of what I do—is there no way to evaluate me, if non-passed legislation doesn’t count? What about my ability to shift the Overton Window, or talk my party members into supporting my position in the future? I would imagine that the targeted function looks like...
∫∞t=nowpolicy_alignmentt−∫∞t=nowpolicy_alignment′t
...where
policy_alignmentis some approximation of the U.S. government’s alignment with my stated positions, andpolicy_alignment'is the same, but in a counterfactual world where my opponent won the general election, such that my political impact across the entire collective future of America is computed. But that’s a very difficult function to build a proxy for, because, even putting the credit assignment problem aside, who can say whether a line item giving my issue of choice $50 million in federal funding is going to have more long-term influence than a high-profile showdown that wins me no friends on Capitol Hill but forces the next president to adopt it as part of his platform, let alone backing down quietly on the line-item in exchange for a favor from a senior congressman that I can use later on?