Monday, December 29, 2008

Predicting Outcomes in Special Elections is complicated by several factors.

Each Special Election is a stand alone, occasional event (not random, but not regularly scheduled) and is not part of a pattern of ongoing events like the primary and general elections held every two years.


(1)They happen so infrequently and irregularly that the sample size may be too small and unpredictable to make accurate predictions on a regular basis.

Every ten years following the census & reapportionment, the entire 21 member state senate is on the ballot. The other election cycles stagger state senate elections with 11 on the ballot in one cycle and 10 on the ballot the next cycle.

This means that from 1994 to 2008 there were potentially 94 Delaware state senate elections. Here is how it worked:
1994 & 1996 21 seats up
1998 & 2000 21 seats up
2002 21 seats up
2004& 2006 21 seats up
2008 10 seats up
Total 94 seats

All 42 state representative seats are up every 2 years which means between 1994 and 2008 there was the potential of having 336 contests, if none of these had been unopposed.

Over that same period there were 3 state senate Special Elections and 6 State Representative Special Elections. There have been four Special Elections since April 2007, but there were none between Jan 2001 and April 2007.
It’s a much smaller sample and scheduling (which is usually based on the death of a legislator or a change in a legislator’s ability to serve) is not regular.


(2)By definition they occur at a time other than when people are not used to voting. The recent 6th Representative District Special Election was held 6 weeks after the longest election cycle in recent history. Some voters may have tired of politics before Dec 20th. It was held
the last shopping day before Christmas, the day before Hannukah and the last Saturday before Kwanzaa. These combined to distract some percentage of voters away from the polls is my guess, although I don’t have polling data to support that assumption.



(3) It’s the only race on the ballot, so there are no coattails. In November 2008 voters had the marquee races for President, US Senator and Governor. In December there was only the 6th State Representative seat on the ballot.
One of the statistics I found interesting on the County Election Department website was voter drop-off which represents voters taking part in the voting process,but not voting in a specific race. In 2006 when the 6th State Representative District was contested in a three-way race, voter dropoff was 2.06% .152 voters entered the voting booth and did not cast a ballot in the 6th State Rep race.

I am betting these 152 voters did not come out on 12-20 which is possibly meaningful in a race decided by 73 votes. If 2% would not make it to the bottom of the ballot when they are already in the booth why would they come out for an election when one race at the bottom of the ballot is the only race on the ballot?

We don’t know the party affiliation of these dropoff voters,but in a district with 2700 more Democrats than Republicans where the Democratic ticket ran well in November there is a good chance a majority of these dropoff voters are Democrats.

On Nov 16th the Philadelphia Eagles played to a 13-13 tie with the Bengals,a team that ended the season 4-11-1 .Yesterday the Eagles beat the Cowboys 44-6 to clinch a NFL playoff berth. The Eagles are 9-6-1 and the Cowboys ended with 9-7 season. Earlier in the season the Cowboys beat the Eagles 41-37. The Eagles lost a close game against a good team, tied an awful team and then later routed the good team that had beaten them. It’s a “any given Sunday” situation.
I think that is what is at work in some of these Special Elections. The team that looks to have the advantage on paper wins more often, but every once in awhile the underdog wins (or ties in football). It’s all about who shows up on game day (election day).




Here is the link detailing voter dropoff in 2006

http://electionsncc.delaware.gov/2006g/residual.pdf

No comments: