Labourleave "poll" in Stoke

Several newspapers last night reported a "poll" commissioned by Labourleave in Stoke on Trent. It claimed to show UKIP on 35%, Labour on 25% and the Tories on 10%.

Labourleave have today put up this document. It is fair to say it is light on methodological detail.

There is no sign of who did the fieldwork, how the data was weighted or even what mode it was conducted by. We do not have any information about the demographics of the achieved sample. Worryingly it doesn't even specify that it was specifically Stoke Central though I can only assume it was. All we have is a sample size of 182. In a random sample this would give a huge margin of error of plus or minus 7 percentage points (despite the 4% it claims in the document)

My understanding is it comes from Labourleave convassing their own database of contacts in Stoke (though there has also been a suggestion that it was a Facebook poll). Obviously something like that brings a heavy risk of bias depending on who they have on their database and what skews may be present. With all those concerns, one can put very little weight upon the results. Even if details are forthcoming and it turns out it was actually conducted and weighted in an appropriate way, the tiny sample size renders it of limited use.

For now - at least until more detail is forthcoming - ignore.