The Seats/Votes Relationship and the Identification of Partisan Gerrymandering in State Legislatures
In: American politics quarterly, Band 15, Heft 4, S. 484-498
ISSN: 1532-673X
This study explores the relationship between seats and votes and the value of this relationship in evaluating districting plans of the lower houses of 15 state legislatures. Partisan bias, the difference between the seats won by the Democratic and Republican parties when they are projected to have the same percentage of the vote, is calculated for each state for the 1982 and 1984 elections. For partisan bias to be useful in evaluating districting plans, it must be stable enough from one election to the next to predict the share of the seats a party will get at the next election given x percentage of the vote. Predictions for 1984 are made for each state by using projection analysis; unfortunately, many of the predictions are not accurate. This calls into question the use of the seats/votes relationship to evaluate districting plans.