[EM] "Ideological districts" in some PR methods

[Kristofer Munsterhjelm]

Some opponents of multimember districts say that electing multiple 
winners weakens the connection between the representatives and their 
constituents, because the representative doesn't represent the district 
in total any longer.

However, some PR methods (like Monroe) return, as side information, an 
assignment of voters to candidates, producing, in effect, an 
"ideological district" per winner.

Could this provide a link between voters and their representatives? 
While the given district no longer has a single representative, any 
voter would know (assuming it's doable without breaking the anonymity of 
the vote) who they helped elect, and thus what candidate the method 
chose to represent them.

Or would this just reinforce polarization? If each representative is 
"chosen" by a group of voters, then that representative may not see any 
need to appeal to the opinions of the voters outside the group.

-km