This article has to be considered in light of the broad anti-Corbyn brief prevailing in The Guardian at the moment.  The headline, "Anti-austerity message will not win over UK voters, poll shows".  This is as good an example as any of the way in which polling is used in class-democratic politics, so I will briefly unpack it.

First of all, the company hired by Jon Cruddas to carry out the research asked voters to support or oppose the following proposition: “We must live within our means, so cutting the deficit is the top priority.”  Most people agreed with it.  Now, any such statement is ideologically loaded.  What does it mean to "live within our means"?  How does "cutting the deficit" actually happen?  What relation does either imperative have to present or past government policies?  Few people are actually against 'cutting the deficit', but that is not what austerity does.  Everyone wants to live sustainably, 'within our means', but does that mean taxing the rich more, does it mean cutting spending on public services, does it mean living in mud huts, does it mean reducing the carbon footprint, or something else entirely?  The mere phrasing of the question, by implying a relationship between 'living within our means' and 'cutting the deficit' that is embedded in the austerian narrative, already frames a particular kind of debate.  A different question, loaded in the other direction, would yield a different response.  For example: "It is important that the users of public services and welfare pay for the banking crisis at all costs.  Agree or disagree?"  In fact, you don't need to take my word for it.  Just do a Google search for polling on spending cuts in the UK.  

Second, this isn't necessarily to criticise the polling company, because it is difficult to think of a single political statement with any degree of complexity that isn't loaded one way or another.  Ideology is complex, it contains ambivalent and downright contradictory elements, and what elements a particular statement touches on will vary depending on how it is phrased.  But it is to point out that things get very messy once you get beyond relatively simple questions about who you intend to vote for.  And in fact, even asking that comparatively simple question is an incredibly complex business, as the polling companies just found out at the last election.  Nonetheless, the company does engage in editorialising, in that it says: "the Tories did not win despite austerity, but because of it".  It is not that they're strictly wrong about this.  The Tories won just over a third of the vote, with the vote concentrated largely in richer, middle class areas.  There is no surprise there: this is probably exactly the level of hard political support for Osborne's austerity project that exists in this country.  Rather, the problem is that, by dint of the evidentiary basis adduced for the statement, it is misleadingly linked to a claim that there exists general political support for Osborne's austerity project.  Incidentally, as the Guardian report doesn't mention it, I will point out that the company was set up by a former Assistant Secretary General of the Labour Party, who ran Labour's 2001 election campaign in which it lost some three million voters, for a "pro-social purpose".

Third, after the polling company had asked its questions and got its predictably stilted answers and then added its own predictably stilted remarks to frame the findings, Jon Cruddas produced an editorial commentary on the research, distilling it down to the key points that he wanted to make.  For example, echoing the research company's editorial, he suggests: "The majority of voters in England and Wales supported the Tories’ austerity measures.  Voters did not reject Labour because they saw it as austerity-lite. Voters rejected Labour because they perceived the party as anti-austerity lite."  Not only is this contrary to the findings of the serious research conducted by the British Election Study, it is actually not at all supported in the research findings.  Nowhere in the process did a representative sample of voters respond to a poll question asking if they supported the Tories' austerity measures.  Nowhere in the research was it established that those voters who didn't support Labour but might have done, rejected them for being too anti-austerity.  It's simply not there in the research: it's made up.  Even if it was, Jon Cruddas is not a neutral player in this.  He is a leading member of the Labour Party, a former cabinet member, a doyen of 'Blue Labour', and someone who has argued for a right-wing, nationalist, socially conservative political strategy.  He has opposed any attempt to pivot Labour's agenda on an anti-austerity argument.  He played a key role as the 'policy review' guru in the failing 2015 election campaign.  Therefore, he has take some share of the blame if it does happen to be the case that the anti-austerity argument hasn't been winning.  As it is, he adopts faux naive posture, as if he had just descended from the clock tower, smoking gun in hand, and said, "where the fuck did all these dead bodies come from?"

Fourth and finally, a Guardian journalist took Jon Cruddas's summary of the research at face value, republished it with extensive quotations from both Cruddas and the research company, and then gave it the headline: "Anti-austerity message will not win over UK voters, poll shows".  Now that is what you would call a performative statement.  It makes a claim about a reality that it seeks to bring into being through its very utterance.  After all, there are lots of ways you could interpret the poll results.  They could be treated as a potentially useful snapshot of a particular bit of raw material, showing how certain ideological statements, phrased in a particular way, will produce a certain response among some voters.  You could even construct an argument that the majority of UK voters were not convinced by an anti-austerity message at the last election.  But what kind of message will win over UK voters is simply not determined, and not in the gift of a poll to demonstrate.  The journalist who wrote the article, Patrick Wintour, happens to be the same journalist who responded to Galloway's 2011 win in Bradford West by attributing it to a "fundamentalist call" to withdraw troops from Afghanistan and fight job losses, thus mobilising the "Muslim immigrant" population.  A racist travesty that the editors rightly spotted and deleted a few hours later.

I merely anatomise this process to point out something of the nature of hegemonic political strategies in a capitalist democracy.  There is a network of institutions - media firms, polling companies, political parties, businesses, think-tanks, lobbies, and so on - all of which contribute to the policymaking process.  And yet all of them spend a great deal of time effectively claiming that they're just reporting on it and responding to it.  They treat policymaking as if it is something that voters do, by some curious osmosis, merely by having opinions and the chance to vote once every five years.  The role of polling in this is supposed to be as a friend of 'the democratic process', a facilitator, enabling real, meaningful communication between party and base, guaranteeing the representative link.  What generally happens is that they are used as raw material for shaping opinion, for conducting debates, for sidelining undesired opinions, and for producing opinions that are desired.  The term for this is, 'manufacturing consent'.