Don't you mean "for the children," Larry?

The +/- 3.5% on partisan registration to define a competitive district appears to be taken directly from my work for the Arizona commission.  This is how I defined a competitive district for the commission.  I regressed a number of predictor variables (partisan registration, previous elections, and incumbency) on congressional and state legislative election results to determine the range where the 95% confidence interval of the regression prediction covered 50%, i.e., we could not know with a reasonable degree of statistical certainty which of the two party's candidates would win the election.  By happy chance, the state legislative and congressional ranges for this definition of a competitive districts were both +/-3.5%, or a 7% spread.

This analysis was only done for Arizona elections in the 1990s.  I have two picky concerns.  One, that California and Arizona elections are not entirely comparable.  And two, that the spread of what defines a competitive district may change with the degree of partisan polarization in the electorate.  +/- 7% on partisan registration is probably not a bad competitiveness measure for current elections, but it might be a good idea for the children if some flexibility were inserted into the range of what defines a competitive district. 

----- Original Message -----
From: Larry Levine <larrylevine@earthlink.net>
Date: Friday, January 14, 2005 5:05 pm
Subject: Re: CA Governor's Redistricting Plan

Picky, Picky, Picky. Don't bother me with details. Ve have to 
change the
system for the people.
Larry Levine

----- Original Message ----- 
From: <karin@cain.berkeley.edu>
To: "Rick Hasen" <Rick.Hasen@lls.edu>
Cc: "election-law" <election-law@majordomo.lls.edu>
Sent: Friday, January 14, 2005 1:10 PM
Subject: Re: CA Governor's Redistricting Plan

ACA3 reads:

Data regarding party affiliation or voting history of electors 

could not

be used. To the extent possible, district boundaries would be 

required to

be drawn to ensure a prescribed level of competitiveness between 

the two

largest political parties.

can someone explain to me how one would assess competitiveness if
political data can not be used?

thank you
.
karin mac donald
statewide database
igs/ucb