One of the most important principles behind my polling is transparency. All the data from my polls is published for all to see. This is important as it shows the research is done properly, and allows anyone who is interested to get the maximum possible value from the work.
The slightly more uncomfortable but nonetheless crucial side of transparency is that people can see when a mistake is made. Unfortunately that was the case last week in my poll of Doncaster North.
I am grateful to Anthony Wells for bringing to my attention an error in the way this survey was weighted. A mistake at the data processing stage meant that the poll included too many Conservative voters and not enough Labour voters. This made opinion in Doncaster North look considerably less favourable towards Ed Miliband than it actually is.
Corrected data tables are below. In a nutshell, Labour lead UKIP by 29 points in Doncaster North, not twelve; Miliband leads Cameron as best PM by 14 points, not one point; Miliband’s constituents would rather see him as PM than Cameron; they give him the highest ratings of the four main party leaders, not the third highest; and they trust Miliband and Balls more on the economy than Cameron and Osborne, not the reverse.
I hope readers will appreciate that I always aim to conduct my research to the highest standards, that data is always checked, and that instances like this are very rare. As ever, I welcome questions and observations on my research, and will always seek to correct errors that inevitably crop up from time to time.
Meanwhile, my apologies to all readers and especially to Ed Miliband.
Note: The data was further corrected in February 2015 to show Labour with a 30-point lead on 55%. Updated data tables are below.