The power of numbers

With a January Ard Fheis called possibly called the last barrier to 7th March election has may have been removed plus being off gave me the time to do some number crunching. Voters are generally creatures of habit so past elections can give a reliable guide to the likely outcome of the upcoming election. So what does the 2005 Westminster and Local Government elections predict for March 2007?It is good news for the DUP and Sinn Fein with both the Westminster and Local government models predicting all present seats are safe and gaining at least two seats but the possibility of more. For the UUP it is a mixed bag, Westminster has them dropping to 19 but the Local Government results have them holding at 24. For the SDLP and Alliance both indicators predict a loss of 1 seat.

The full results by the Westminster 2005 result:
Alliance 5 seats
DUP 38 seats
SDLP 17 seats
SF 27 seats
UUP 19 seats
Others 2 seats

The full results by Local Government 2005 result:
Alliance 5 seats
DUP 35 seats
SDLP 17 seats
SF 26 seats
UUP 24 seats
Others 1 seat

Area profiles to follow.

IMPORTANT NOTES: Seats were assigned to parties strictly on the basis of who was closest to a quota. No assignment was made on the basis of possible/likely transfers nor has any projection of existing trends been extrapolated (ie DUP vote grew X% between 2003 and 2005 but if that continues their likely vote be). Westminster results tend to show a party’s maximum potential vote in a constituency while Local Government shows the core vote. When the two agree it seems the highly likely result and were they disagree it highlights were a party should target a constituency for extra effort. In individual constituencies the perceived geographcal base and number of candidates a party chooses and the vote management system adopted can have an effect on the final result. Examples of how this can impact are – the choice of two Ballymena UUP candidates in North contributed to their loss of a seat, vote mismanagement in West and East Belfast cost SF and the DUP a seat respectively. It is possible to confound both indicators as David Ford did in South Antrim in 2003.

The same methodology using the 2001 Westminster results would have predicted the following results in 2003:
Alliance 3 seats (underestimate by 3 seats)
DUP 29 seats (underestimate by 1 seat)
SDLP 20 seats (overestimate by 2 seats)
SF 23 seats (underestimate by 1 seat)
UUP 32 seats (overestimate by 5 seats)
Others 1 seat (underestimate by 2 seats)
The UUP skew was because of the DUP standing aside in a number of constituencies.

The same methodology using the 2001 Local Government results would have predicted the following results in 2003:
Alliance 6 seats (exact match)
DUP 29 seats (underestimate by 1 seat)
SDLP 19 seats (overestimate by 1 seat)
SF 23 seats (underestimate by 1 seat)
UUP 28 seats (overestimate by 1 seat)
Others 3 seats (exact match)

Of the two models the 2001 Local Government results proved the more accurate predictor.