Two pollsters, three polls, and UKIP shares between 16% and 24%
With all eyes on UKIP polling shares following their by election successes the online survey by ComRes for the Indy on Sunday and Sunday Mirror carried out a test to see whether, as many purple enthusiasts argue, their shares are understated by firms that don’t specifically prompt for the party.
So the ComRes sample was split in two with the first using the conventional approach and the second including UKIP in its main party prompts.
The problem with this is that the sample sizes became so small, down to 782 in one case, that the margin of error increases substantially especially when trying to analyse the UKIP voting subset.
In fact the difference between the two approaches can almost all be explained as standard margin of error.
- With that caveat a big move was apparent between the two ComRes polls. The views of women barely changed when UKIP was prompted – men, however increased their support by 8%
Read into that what you will! Maybe prompting says more about how men and women respond to online polling than it does about UKIP support.
Another difference was that non-2010 voters amongst UKIP support amounted to 7% in normal poll, but 13% in the prompted one.
Meanwhile the latest YouGov, with a later fieldwork period than ComRes, has UKIP down 3% from dizzy heights of last week to a more normal looking 16%.