We take up Fiorina's charge by using a new statistical method (King 1997) to produce more accurate estimates of the direction and magnitude of split-ticket voting in each state and congressional district in 1988.
We employ King's technique to produce more accurate estimates of split-ticket voting within districts and states.
His operationalization of split-ticket voting is the simple difference between the highest and lowest Democratic percentage of the two-party vote among races within each state.
For instance, a value of zero may mean that no split-ticket voting occurred (as Burnham and Rusk assume) or that every voter is casting a split ticket.
4) Below, we show that once the direction of ticket splitting is properly estimated, there is a conditional relationship between split-ticket voting and split outcomes.
When discussing our estimates of split-ticket voting, we will use a two-letter abbreviation to indicate the four possible voting patterns; the first letter is the party of the presidential candidate chosen, and the second letter is the party of the congressional candidate chosen.
If divided outcomes and split-ticket voting were unrelated, then one would expect no connection between the fraction of Bush splitters in a district and an RD election result.
Logit Equation Predicting Divided Outcomes (RD versus RR Districts) by Split-Ticket Voting Variable b Bush splitters 63.
For the 33 Senate races in 1988, split-ticket voting was about as common as in the House contests.
In this section we attempt to explain why split-ticket voting is more common in some electoral units than in others.
Thus, split-ticket voting is consciously used to produce divided government.
On the other hand, unintentional theories posit that split-ticket voting is a by-product of other forces, such as weak partisan attachments (Petrocik and Doherty 1996), ballot mechanisms (Beck 1997; Campbell et al.
They found that split-ticket voting was about 8% higher in states without a straight-party option in the 1950s.
This examination of the causes of split-ticket voting at the district and state level finds that most of the influences are outside the voter's control, which supports the unintentional theories of divided government.