Electoral geography is easy and powerful, because images can provide clear and simple messages, but they often distort reality. Maps of Red & Blue US or Remain & Leave UK exaggerate relatively modest differences. #thread
1. We have all seen "the" electoral map of 2016 US presidential elections, which shows a largely red US (i.e. Republican). 
2. It's even more pronounced when you look by county, which is the level most gerrymandered in the US.
3. But these maps don't take into account: (1) population density; and (2) margins of victory. Whereas there is massive differentiation in terms of (1), the differences in (2) are often not very big.
4. Consequently, most of US (or UK) is neither red, nor blue, but various shades of purple.
5. And when population size is taken into account, the predominantly red US becomes predominantly blue & dark purple US.
6. Or, in more complex maps...
7. Leaving aside that all these maps mostly refer to a minority of Americans. Non-voters "won" 41 US states! 😲
8. In short, electoral geography is highly simplistic. In the few district systems with first-past-the-post (ie US), they do accurately reflect the electoral outcome in terms of seats, but distort it in terms of votes.
9. In vast majority of countries, however, electoral geography offers an almost useless insight into electoral behavior, leading us to overly simplistic, and sometimes completely wrong, theorizing with serious real world consequences.
10. It essentializes geographical units and subscribes meaning to them, even though most geographical units are quite diverse in terms of eco/pol/soc characteristics, as reflected in diverse electorates.
11. This is not to say that electoral geography is useless, or new/growing "center-periphery" cleavage is not of importance in UK and US (or, for that matter, much of Central and Eastern Europe).
12. Rather, it means that geography is an empirically and theoretically very weak proxy. It is internally quite heterogeneous and we don't really know what it means.
13. At best, electoral geography should be used as a first, quick and simple, tool to hypothesize possible explanations, which are then explored in more detail at both a more accurate meso level and at the individual level.
14. Using simplistic red-blue maps without clear explanations is not just inaccurate, it is irresponsible, as it feeds into polarization and into populist narratives of representing "the people" as a homogeneous bloc.
15. Context matters, but most people's relevant geographical contexts do not perfectly reflect administrative geographical boundaries. #TheEnd
