[EM] Re: Automated districting

Toplak Jurij jure.toplak at uni-mb.si
Fri Jan 9 00:49:01 PST 2004


Do you know if it requires contiguity (physical proximity), or does it
pick random cities from around the state to form a given district?

Yes, it requires contiguity. Because it uses political units, there is also
no problem with preservation of communities of interest, since communities
usually live in same political units. The only decision that has to be made
is on which level of the political units should be used (counties,
municipalities, or something smaller, like school districts or
neighbourhoods). If you use larger units, districts will be more compact,
but the population deviances will be larger. If you use smaller units,
population will be more equal, but the district shapes might be wierd and
communities of interest might not be preserved.

Also, do you happen to have URLs for any of those articles?

Yes, Micah Altman's article is on the net - in a form of a chapter of his
PhD dissertation (as far as I saw, it is completely the same as published
article).
http://www.hmdc.harvard.edu/micah_altman/disab.shtml
Click on the Chapter 5 (it is in PDF): Is Automation the Answer?
He argues that automated districting is too complex to be handled by
computers.

Best,

Jurij Toplak




More information about the Election-Methods mailing list