— Mike Smithson (@MSmithsonPB) March 14, 2013
Why GE2015 could be closer than it now looks
One of the biggest challenges when analysing voting intention polls is the very high level of “will vote – don’t knows” that we see.
These totals are generally higher in ICM polls because the form of questioning is much less judgemental than with some other firms and respondents appear more comfortable to say they don’t know. Generally these get “netted off” and and are not included in the headline figures.
This can be distorting because one of the features of current polling is the high levels of don’t knows.
Thus in the latest poll out this week the DK proportions of 2010 voters were 14% LAB, 28% CON and 29% LD. Adding on the won’t vote and the refusers and you get to CON 33, LAB 22 and LD 35.
Given that these were all actual voters three years ago then the absence of a voting intention could be signficant.
In the pie chart above I feature the responses of 2010 CON voters. The striking feature for me is how few are actually saying another party. 81% say CON or don’t give a voting intention.
As we get closer to the election, particuarly in the 80 or so seats that matter, this segment will surely be the main target of Lynton Crosby and his team.
For the latest polling and political betting news