However in addition to the NI assembly parties, we also included the PUP because of their recent local government election performance in the constituency. We excluded the NI Conservatives because they have yet to demonstrate significant support either in NI public elections or via the polls. These sorts of judgements have to be made otherwise the practical process of running the poll can become unworkable, and is sort of similar to the current discussions affecting the election TV debates. It's probable of course that not all the parties included in our poll will end up running in this May's election.Most of the difficult 'who to include', 'who to leave out' decisions, is caused by the fragmentation of Unionism. It's surprising to note that there are 11 active political parties in Northern Ireland with representation either at local government, or at the NI Assembly. Only two of these 11 are Nationalist/Republican i.e. the SDLP and Sinn Fein. Defining the Alliance party as neutral, we then find that the remaining eight parties are Pro-Union. Thus we see how fragmented the pro-union political camp is. But before the Unionists rush out and form election pacts with each other they should note that you can't assume that the support for one type of unionist party will automatically transfer to another.
You may have also noticed our previous commentary article about polls and the role of polls. People understandably conclude that polls exist to predict elections but that isn't their main function. Political polls are meant to track ups and downs in opinion, and the ebb and flow of views on particular issues, in between elections, and during the build-up to elections. Yes, polls the day before an election, asking the basic 'Who are you going to vote for' question, should turn out to be accurate, and reflect the election result to a reasonable degree. Such was the case in the recent Scottish referendum with the last polls accurately predicting the final result.
In fact you could put the reverse point - If Opinion Polls could accurately predict public elections every time, then there would be no need to run the elections, would there? Incidentally it should be noted that running public elections (like the upcoming Westminster election this May) is an extremely expensive business, mainly because it is very people intensive and requires a lot of staff to run. If polling science could reach the stage, were it could always accurately and reliably represent opinion to within a very small error then this would remove the requirement for public elections. This would save the enormous amount of money spent on public elections, which could then be re-invested elsewhere e.g. health and education! However, maybe we still need the formal elections to keep people involved, and provide authentication to the democratic process.