Predicting results.. what works and what doesn’t

Predicting election results is a mix of art and science.  We can all think of polls and pundits in the past who have got it wrong.  But there are some signs observers can look out for to make at least a partial judgement.

Yesterday (March 5th) I attended an excellent event at the BBC in Media City which looked at the North West (the region with the joint highest number of marginal constituencies) and attempted to make some predictions.

This got me thinking about what we should be looking for as indications of potential political success at election time.  So I’ll be writing a few pieces about the various measures and what they might mean.

Yesterday’s predictions, which only involved four seats in the North West region changing hands, used a number of factors including financial donations, polling data, local election results, incumbency factors and so on.  Audience members raised other potentially significant aspects such as levels of activism, key campaign issues and the personalities of the various contenders.

The aspect I am concentrating on today is local election results.

It is easy to look at a constituency, look at the local election results in the run up to a General Election, and make a party-based judgement on those figures.  In fact, in isolation, local election results are among the weakest of predictors of GE success.  Let’s take Edge Hill’s local constituency, West Lancashire.  In the run up to 2010 there had been Conservative local victories in Skelmersdale.  (If you don’t know the area, think in terms of Labour suddenly winning Surrey).

These were unexpected and many felt this pointed to a growth in Tory support that would turn out the incumbent Rosie Cooper.  In fact , although there was a small vote-share increase for the Conservatives, in line with the national trend, Ms Cooper is still the MP and had a 2010 majority of more than four thousand.

The thing is, local elections are simply different.  Firstly the turnout can be considerably lower than in a General.  Secondly, party allegiance can be weaker.  Finally the personality and activity level of a local candidate, particularly one running on an “us against them” ticket can be significant in a way that simply does not transfer.   This becomes apparent on those polling days when a General Election and Local Election take place on the same day in the same area.  Vote-splitting can be very common.

So what do local election results tell us?

Well they give some indication of party organisation.  A badly organised party will not  usually manage a large number of victories.  They give some indication of activist levels as volunteers need to be found to stand and again found to campaign. And they give some indication of local roots and knowledge.

But what they don’t do is tell us who will become an MP.

In 2001, had local election results been an indicator of national success I would have become an MP in Liverpool.  This on its own should be a warning to those who put too much prediction weight on the colour of the local council.

Published by

Paula Keaveney

Paula Keaveney

Paula Keaveney is a Senior Lecturer in  Politics.

A former journalist and PR professional, her research interests include political communications, public affairs and PR and marketing in the charity sector. She is the Chair  of the  Political Marketing Group of the Political Studies Association. She is also a former leader of the opposition on Liverpool City Council.

Leave a Reply

Your email address will not be published. Required fields are marked *