Nate Cohn

Nate Cohn @Nate_Cohn

@im_sorry_wtf @Libnex2 @PoliticsWolf they're both 50% black — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf lol https://t.co/ePgwIdb0eT — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf but i just don't think the kinds of questions you'd raise about this map are any different than those i'd raise about yours? — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf and again, this map appears to have come together in 24 minutes, based on the timestamp on your tweet, so you certainly won't find me defending it as the perfect map! — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf alright here's a 2010 map that at least began in the same way as the prior map, until i found myself needing a new district. it's 9-5 trump in '16; i'd guess it's 7-7 in '12 (using 12/16 pvi, it's 8-6 with a mean-median gap is like a net-9 points) https://t.co/MGUiKVPj1n — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf ugh lol. fair enough. but in a certain sense, does it matter? is it your contention that you can't draw a more-or-less equivalent map without similar bias? — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf now i won't defend that it's *the* map--i spent like 30 min on it! but i guess i'd just ask you this: what is your basic case for why your map would prove there's no bias but this map wouldn't prove the opposite? — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf and just for fun i threw together a pretty boring map of michigan that's 8-5 trump in '16 (and nearly 9-4) with a 4 pt mean/median gap and 10 pt efficiency gap https://t.co/4QUoltya7l — PolitiTweet.org

Nate Cohn @Nate_Cohn

@ProfNickStephan @PoliticsWolf yeah, tx/ca should be/are a democratic geograhpy goldmine. surprised by NY (though it's admittedly not a state i've thought much about) — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf i don't have that joined up here (this data isn't being used for redistricting purposes at the moment) — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf as i said, alone it's not dispositive. but it's not meaningless, since there *is* spatial correlation. if they were totally random, then it would be meaningless. — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf that doesn't mean the bias can't be overcome--that your map does prove! but it does mean that lots of nonpartisan maps making totally defensible and slightly different choices will also be very biased! — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf that alone doesn't ensure map bias, since results need to be both spatially correlated and inefficient. but we all know that results are spatially correlated, as well, thanks to urban-rural gap etc. — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf that alone doesn't ensure map bias, since it depends on whether the results need to be both spatially correlated and inefficient. but we all know that results are spatially correlated, as well, thanks to urban-rural gap etc. — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf the democrats have an underlying spatial inefficiency problem. obama won precincts with 47.5% of the major party votes in michigan (population would be better, i know, but i've got this data up already coincidently!) v. 55% of the major party vote — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf and for the narrow purpose of this exercise, i do think the choice to constrain oneself to limiting county splits, as opposed to municipal splits--does have a benefit, as it does more to constrain the mapdrawer's discretion — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf it is a choice, but i'm not saying that *my* map of michigan is *the* map, and therefore proves geographic bias. what i am saying that the existence of these kinds of choices makes it hard for a map to prove or disprove whether there's geographic bias — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf and consequently, that makes it less of a test of geographic bias, imo — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf again, i'm not saying you *can't* make this choice or defend it. but it is a choice (and it's not the one i'd make fwiw). more importantly, it's a way of drawing CDs that makes them a little more subjective than always drawing CDs within a county, when possible — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf i do think it is successful at disputing the implication of chen-rodden, that there's like no possible way to draw a map without partisan bias. but i don't think it follows that it shows there's no bias — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf anyway, that's just one example. but i think it highlights why it's so difficult to evaluate geographic bias this way, especially if you want to take the COI > jursidictional lines kind of approach, which is inherently more subjective — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf but speaking personally, i think the likeliest nonpartisan map draws a CD that's entirely in Macomb county and votes for Trump — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf to take MI, since it's one of your two examples of a state where there's no evidence of bias based on a map (and i'll admit to not thinking much about NY), i'll note that you split Macomb to draw an extra blue CD. maybe that's justifiable. i'm not saying you *can't* do that — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf and as i said long ago, i don't think your maps are always drawn in a way that makes it easy for me to accept the conclusion that they're *the* or the *likeliest* nonpartisan maps, as opposed to just a nonpartisan map — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf and as i said long ago, i do think always think your maps are drawn in a way that makes it easy for me to accept the conclusion that they're *the* or the *likeliest* nonpartisan maps, as opposed to just a nonpartisan map — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf and i don't want to relitigate our old debates about the important of your map, which i always enjoy, but my old position holds: i think it shows that it's possible to draw fair maps, not that there's no bias — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf it's not only latino voters, but because of their age, eligibility rate, and low turnout that's the major predictor (including in new york) — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PoliticsWolf (doesn't miss that: https://t.co/lgvECoVA5r) — PolitiTweet.org

Nate Cohn @Nate_Cohn

@mlatner @ElectProject @davedaley3 notice that there's no claim in my thread about a net-effect on gerrymandering; there's only a claim on the net-effect of geography/spatial inefficiency. those are separate questions, and i happen to be taking the side of this one that is good for... antigerrymandering activists — PolitiTweet.org

Nate Cohn @Nate_Cohn

@mlatner @ElectProject @davedaley3 notice, however, that there's no claim in my thread about a net-attenuation in the effect of gerrymandering; there's only a claim on the net-effect of geography/spatial inefficiency. those are separate issues — PolitiTweet.org