YouGov Welsh poll

The latest YouGov poll of Wales for ITV is now up on their website here. Topline voting intentions in Wales, with changes from the last poll way back in July are:

Westminister: CON 22%(-1), LAB 51%(-3), LDEM 9%(+5), PC 10%(nc), UKIP 7%
Assembly const: CON 21%(+2), LAB 46%(-4), LDEM 10%(+3), PC 17%(nc), UKIP 5%
Assembly list: CON 14%(+3), LAB 26%(-9), LDEM 11%(+3), PC 26%(+6), UKIP 13%(+1)
European: CON 23%, LAB 44%, LDEM 7%, PC 14%, UKIP 9%

The list vote figures for Labour and Plaid are particularly striking, though all the usual cavaets about not getting too excited over a single poll and eyecatching results normally turning out to be wrong should apply.

The poll also asked how Welsh respondents would vote in a referendum on membership of the EU, findng 42% would vote in favour, 35% would vote against (and suggesting that Wales is slightly more pro-European than the country as a whole).

A quick update on the latest voting intention polling. TNS-BMRB’s weekly poll has topline figures of CON 29%(-2), LAB 38%(-3), LD 11%(+1), UKIP 12%(+2), Others 10% (changes are from a week ago).

Meanwhile this morning’s YouGov poll for the Sun had topline figures of CON 32%, LAB 41%, LDEM 12%, UKIP 8%.

This morning’s Times had some claims about polling conducted by UKIP in Eastleigh. I would advise totally ignoring any claims about “private polling” from political parties unless they cough up the tables so you can see if they were playing a straight bat. More often than not party claims about their “polling” in elections actually means their canvas returns. I’ve dropped a line to the various polling companies just to check none of them have any tables to release under the BPC disclosure rules, but thus far no one seems to have done anything.


This week’s results for the YouGov/Sunday Times poll are online here. Topline voting intention stands at CON 32%, LAB 43%, LDEM 12%, UKIP 9% – so very much in line with the typical YouGov Labour lead of about 10 points. There’s no sign of any remaining effect from the EU referendum pledge here.


Almost three quarters of people blame food manufacturers (26%) or meat processors (46%) the most for the horsemeat scandal, rather than retailers (11%) or the government (6%). While a majority of people think that there is more the government could be doing to keep the food chain secure, broadly speaking the government is seen to have handled the horsemeat scandal well – 47% say they have handled it well, 39% badly.

68% of people do not think there is any actual health risk from horsemeat getting into the food chain and 37% say that, if it was properly sourced, they would be prepared to eat horsemeat.

Relatively few people say that they will substantially change their behaviour as a result of the horsemeat scandal – only 5% say they might change which supermarket they use to they buy their groceries, only 13% that they will reduce the amount of meat or beef that they will buy. However, a third of people say that they will reduce the amount of *processed* meat they will buy. In reality all these are likely to be gross overestimates: it is much easier to say in a survey that you will change your behaviour than it is to do so in real life – in practice most people will probably continue as usual.

Eastern European Immigration

On the general principle of the freedom to work and live anywhere within the European Union, 33% of people think it is a good thing, 56% a bad thing.

On balance immigration from western European countries like France and Germany is seen as a positive thing (39% think it has had a positive effect on Britain, 16% a negative effect, 31% neither). Immigration from Eastern Europe and from outside the European Union are both seen as having had a negative effect on Britain by a majority of respondents.

70% of people think that the rules on immigration into Britain from the EU should be tougher, almost the same as the 73% who think the rules on immigration into Britain from outside the EU should be tougher. On the specifics of the extension of the right to live and work across the EU to Bulgarian and Romanian citizens, 20% of people think there is no problem with this and Britain should welcome them, 19% think it will have a negative impact on Britain but we have no choice but to meet our legal obligations, 48% think Britain should limit the right of Bulgarian and Romanian citizens to live and work in Britain, even if it means breaking the law.


There is little support for fox hunting being legalised. Only 23% want to see the ban lifted, compared to 65% who would like it to remain. This includes 50% of Conservative voters.

29% of people who describe the area they live in as “urban” say the number of foxes in their local area has increased in recent years, but the overwhelming majority, 92%, say that they have never been attacked or felt threatened by a fox. Nevertheless there is significant minority support for a cull of urban foxes – 38% would support a cull, but 41% would oppose it.

Long term care and inheritance tax

52% of people say they support the government’s plans on capping the cost of long term with only 21% opposed. 50% of people say that it is right that the plans to reduce inheritance tax were shelved to fund the long term plans, 26% would rather they had been funded in some other way.

Asked a straight choice of whether they’d prefer inheritance tax to be reduced, or the cost of long term care to be reduced, far more people choose the later – 57% to 18%. This is particularly the case for older voters, people over the age of 60 say they would prefer a cut to long term care costs over a reduction in inheritance tax by 66% to 13%


Finally, 76% of people support the principle of withdrawing benefits from unemployed people who refuse to work. On the more specific recent court case, 55% of people think the government should be able to withdraw benefits from unemployed people who refuse to do unpaid work experience, 34% think they should not.

Descriptions matter, hence politicians and campaigning groups often go to great lengths to try and frame the language that policies and causes are described in, trying to get policies they support referred to in inherently positive terms and their opponents policies with inherently negative terms. Think of the fantastically successful efforts of supporters of the very dull sounding financial transactions tax to have their cause consistently referred to as a “Robin Hood Tax”, or attempts by those opposed of estate taxes in the USA to get people to call the target of their dislike the “Death Tax”.

This does, of course, pose rather a problem for pollsters. If even the language used to describe a policy is politically contentious how do you ask an unbiased question on it? You can’t ask about a policy without referring to it, yet just the language you choose to describe it is coming down on one side or the other.

Sometimes it is relatively easy – there is a non-contentious neutral term in the middle. For example, supporters of same-sex marriage tend to refer to it as “equal marriage”. Its opponents tend to refer to it as “redefining marriage”. The impact of the wording is clear, ComRes questions asking about redefining marriage tend to show a majority against, Populus questions that have asked about gay marriage in terms of giving gay and lesbian people equal rights have tended to produce the most positive results. The obvious solution which most polls have taken is to take a less contentious term in the middle, like gay marriage or same-sex marriage, which is not so value-laden or innately associated with one or the other side of the argument.

Other times it is more difficult. The government’s planned changes to housing benefit are officially called the “under occupancy charge”, but have been referred to by Ed Miliband and some of the press as the “bedroom tax”. There is not an obvious neutral point, just the way the government refer to it, or the way the opposition refer to it. Prima facie it is better to ask the question about the “under occupancy charge” – it is, after all, its official name and not an inherently positive or negative term, while “bedroom tax” is a pejorative term of abuse for the policy. However, it’s not always clear cut, if we go back to the 1980s, for example, after a while the community charge was almost universally called the poll tax – it would have been almost obtuse not to call it the poll tax when asking about it, since that was what everyone called it. Generally speaking, I’d say one should avoid value-leaden slang terms for policies, or terms solely associated with champions or opponents of a cause… but it is a matter of judgement.

In the particular case of the ComRes/People poll today on the charge, ComRes have gone for the solution of referring to it using both terms – “From April, unused bedrooms in social housing will be subject to an under-occupation charge or ‘bedroom tax’ meaning housing benefit will be reduced for working age households if they are deemed to have spare rooms.” (though such neutrality in language is then somewhat undermined by then using the term “bedroom tax” six times in the rest of the survey)

The reporting of the poll, incidentally, goes on to illustrate just why one needs to be wary of “agree or disagree statement” polls. As I’ve written at length before now agree/disagree statements risk biasing answers in the direction of the statement, and often produce apparently contradictory answers within the same survey (the post here includes some cracking examples). This isn’t necessarily a problem if the survey includes statements in both directions, because looking at the whole you can get a rounded picture, but for those more interested in pushing an agenda than discerning the truth it particularly lends itself to partial reporting and cherry-picking.

For example, in the ComRes poll today they found people agreed with the statement “David Cameron should abandon the ‘Bedroom tax’ entirely and think of other ways to save money.” by 45% to 37%. However, asked if they agreed or disagreed with a statement in the opposite direction “It’s only fair that people who have spare bedrooms in council or housing association homes should receive less housing benefit”, 46% agreed and 36% disagreed. The two statements are not necessarily contradictory, but taking either one of them in isolation without reference to the other creates a very different impression of what public opinion on the issue is. The People’s write up does at least mention that second statement briefly in passing, much of the interpretation of the poll elsewhere ignores it completely.

ComRes’s monthly online poll for the Independent on Sunday and Sunday Mirror is out and has topline figures of CON 31%, LAB 36%, LDEM 8%, UKIP 14%.

The changes from last month are bit complex. As regular readers will know, there has been a bit of a back and forth in ComRes’s methodology. For most of last year ComRes treated likelihood to vote for minor parties differently to how they did it for the main parties – for the big three they included people who said they were 5/10 or more likely to vote (weighted proportionally), for minor parties they only included those who said they were 10/10 certain to vote. In their December online poll they experimented with treating all parties the same on turnout, producing a substantial jump in UKIP support. In January they used their old method, which dropped UKIP back down by 4 points. Unfortunately this wasn’t flagged up in media reporting of the polls, giving the impression of UKIP increasing in December and then dropping back down in January after Cameron’s EU referendum pledge, when actually much of the movement was due to methodological reasons.

Anyway, the back and forth seems to be behind us – ComRes have now shifted to treating all the parties the same when it comes to taking account of likelihood to vote:

“In recent months we have been exploring the best way to treat smaller parties when calculating voting intention. We have experimented with including smaller parties in voting intention scores only if respondents are certain to vote; this has been on the basis that, comparing polling against actual 2010 results, we were concerned that the traditional method was over-stating smaller parties. However, with UKIP the game has changed and we therefore propose that from now on supporters of smaller parties will be included if respondents are 5/10 or above in terms of likelihood to vote, as is the case with the major parties. We will however continue to review our methodology as the general election approaches”

The effect is simply to increase the reported level of support for minor parties, with the Conservatives, Labour and Liberal Democrats all decreasing proportionally. Last month’s ComRes figures were CON 33%, LAB 39%, LDEM 11%, UKIP 10%, so a naive comparison would suggest a significant increase for UKIP with everyone else suffering. However, ComRes have released changes on what last month’s figures would have been using the new method, implying that last month’s figures would have been CON 32%, LAB 37%, LDEM 11%, UKIP 13% and that today’s poll actually shows very little difference for the Conservatives, Labour or UKIP.

That is interesting in itself – last month’s ComRes poll was the height of the post-referendum boost for David Cameron. Polls since them for other companies have shown whatever benefit Cameron accrued from his referendum boost fading away again, but ComRes have him consolidating it. As ever, that could be a sign of a Conservative advance, but in the absence of other polls confirming it will probably turn out to be a blip.

UPDATE: Full tabs are here