Tag Archives: polling companies

How Polling Works

Following on from my critique of polls and polling companies, I have produced this flow chart that explains how opinion polling works.  Polls don’t exist to measure public opinion, they exist to shape it.  The media and the polling companies enjoy a symbiotic relationship in which each sustains the other.  One supplies a narrative and the other responds by producing a poll to support that narrative.  The media company then produces a story that reinforces the initial narrative,  which uses the poll as ‘evidence’.  You may need to click on the image to view it properly.

how-polls-workEven when polls are patently nonsensical or illogical,  their ‘findings’ are lauded by commentators and their followers, who cite them as evidence to support one narrative or another.  One such poll is the recent ComRes poll , which claimed that “most people” think the Tories would do a better job at running the National Health Service this winter than Labour.  The same poll also claimed that “most people” thought the Health Secretary, Jeremy Hunt, was doing a “poor job”.  Make of that what you will.  In any case, the media’s commentators seized on it and cherry-picked its ‘findings’.  The negative narratives that had already been produced to put the Labour Party in a bad light were thus reinforced by this shoddy poll.

Wash, rinse and repeat: that’s how the process works.

Advertisements

3 Comments

Filed under Journalism, Media, Opinion polls, propaganda

But… But… The Polls Say…

I can’t count the number of times people have said to me on Twitter or Facebook that the polls have “told them” (as if the polls are some present day Oracle at Delphi speaking especially and directly to them) that Jeremy Corbyn is ‘unelectable’. This usually happens when you demolish their narrative (I won’t dignify their discourse with the word ‘argument’) that only a Blairite or a similar stuffed shirt would make a better Labour leader. They base this notion on the fact that he (Blair) won three General Elections in a row. That the Blair-led Nu Labour party won those elections is irrefutable, however as I pointed out in a previous blog, Labour lost 5 million voters in the space of 13 years. Of course, that fact is also ignored because it reveals an uncomfortable truth: the policies of Nu Labour and its variants Blue Labour, and the unfortunately coined ‘Brownism’, are unpopular with many people. So why do people persist in citing polls as some kind of ‘evidence’?

For eons, humans have sought to master nature. One way in which people have tried to achieve a mastery over  powerful unseen forces is by attempting to predict future outcomes.  For some, tarot cards do the trick and for others, it means consulting their horoscopes in the papers.  Sometimes, the future will be divined from random signs that have their origins in folklore: bones scattered on the ground and animal entrails thrown onto a fire have both been used with little or no success.

Polling deals with numbers, so it is seen as being more scientific and less susceptible to human fallibilities. I mean, numbers don’t lie, surely? Well, they do. It all depends on how numbers are interpreted and who is doing the interpreting. Sadly, polling companies don’t employ people who have been produced in an ideological vacuum and free of discourse. They may make all kind of plausible claims that their ‘research’ (sic) is ‘rigorous’ but this is done to throw people off the scent. I mean, how objective was Lord Ashcroft’s polling? At least he declared his political position from the outset. Polling companies don’t do that and will claim to be ‘objective’, but as many academic researchers will tell you, it isn’t possible to be totally objective.  This is why qualitative researchers use self-reflexivity.  Pollsters don’t bother with such things because they see themselves as the impartial interpreters of signs and that’s their weakness.  Thus, we can regard them, quite literally, as the self-appointed high priests of psephological divination. In the eyes of the mass media, therefore, they are uncritically accepted as politically-neutral soothsayers; mere observers of a history to come. Their legitimation having come entirely from their claim of being impartial.

But it’s not just the numbers, it’s how people arrive at their responses . This is rarely, if ever, discussed. Polls exist, not to gauge public opinion, but to shape it.  Thus, the questions that are asked of respondents are equally important as the numbers themselves.   Since Jeremy Corbyn became leader of the Labour Party, the right-wing press and the Nu Labourites of the Parliamentary Labour Party have persisted with the narrative that Corbyn is “unelectable”.  There is no basis for this claim and it seems to be based entirely on antipathy towards him, rather than his policies or ability to connect with voters (which is also disputed).  Narratives like this and “Corbyn has failed to reach out to working class voters” are trotted out frequently as kinds of truths.  But if you start to subject these narratives to scrutiny, they quickly fall apart.  Polls may start with a statement like “It has been said that Jeremy Corbyn is unelectable”.  To this, a question will be added that reads something like “how likely are you to vote for a Labour Party led by him”?  The polling companies prompt respondents to react in a certain way.  Thus the narrative has been planted in their minds from the outset.  The narrative will also be repeated in the mass media as a kind of Truth.

This article in the New Yorker asks if polling is destroying democracy.  If polls are being commissioned by the newspapers and broadcasters, then questions need to be asked, not only of their validity but of their purpose.  Last week, the Daily Express produced a story from a survey that claimed “Most people want to go down the pub with Boris Johnson”.  My first question was “who did they survey” and my second question was “who commissioned this rubbish”? Perhaps the most important question is “who is this story and survey for”? It tells us nothing and if The Express commissioned this poll, then it begs the question of why it’s still in business as a serious (sic) newspaper.

The failure of the polling companies to predict the future was brought into sharp relief by Brexit, Donald Trump’s victory and last year’s UK General Election. Their fallibility was laid bare for all to see.  “Ah, but what about the margin of error”? What about it? Whenever polls are criticized, especially in the case of their claims of Corbyn’s apparent unelectability, the margin of error canard is deployed as an appeal to authority.  Crucially, those who defend polls never consider the fact that those questioned in these surveys may be Tories who won’t vote for Corbyn or who may not even vote at all.  They may even change their views between now and election day. Some respondents may even lie. Apparently, these variables are factored into polling but how accurate is this margin of error? Not very, by the look of things.

YouGov is often cited by polling experts and watchers as being the most accurate of the polling companies, but this company was founded by Tories, Stephan Shakespeare and Nadhim Zahawi.  The latter still has questions to answer over his involvement in a Jeffrey Archer charity in which millions of pounds, apparently destined for Kurdish refugees mysteriously vanished.  Shakespeare is a twice-failed Conservative Parliamentary candidate and former member of the Socialist Workers’ Student Society.

The latest YouGov poll repeats the by now familiar “Labour is x points behind the Tories”. Polling companies and the mass media work hand-in-glove with each other.  The latter produces a constant stream of negative stories and the polling companies respond by producing a poll, which reinforces the claims of the former.  Sometimes the poll will be commissioned by persons or organizations known or unknown.  In any case, they feed each other.

 

6 Comments

Filed under allegations of bias, Media, propaganda

Opinions Are Not Sacrosanct

I’ve lost count of the times someone on Facebook or Twitter has told me that I should “respect” their opinion. I always tell them, “I don’t give a fuck about your opinion” and they, like some needy child, will always wail, moan and stamp their feet and accuse me of all kinds of things. As someone once said “Opinions are like arseholes, everyone has one”. Those who offer their opinions en lieu of a reasoned argument need to bear this in mind.

Opinions also tend to be confused for coherent and well-constructed arguments. There is nothing well-constructed about opinions, they are the product of an emotional reaction to someone or some thing, or they’ve been produced by the media and repeated without question . Yet, if you use the word ‘argument’ in your response, these people will confuse it with the word ‘quarrel’ or ‘disagreement’. An argument or a thesis is based on reading/study and is supported by the production of evidence. Opinions require no evidence. Hence, they are useless in a debate.

So what’s the difference between an opinion and an argument?

An example of an opinion is “I like McDonald’s burgers because they’re nice”. That statement is based on how a person feels about McDonald’s in relation to Burger King or any other fast food outlet. It is entirely subjective. There is no evidence to support the statement.

An example of an argument is “I hate McDonald’s because they refuse to recognise unions on their premises and have sacked unionised workers. Therefore, I boycott them”. Here, I have not only made a statement, but I have supported it with some evidence in the form of a link and given you my reasons for disliking McDonald’s. If I wished I could have added more evidence to support my thesis.

Recently, anti-Corbynites, Tories and Kippers have shouted in near unison that “Jeremy Corbyn is unelectable”. There is no evidence offered and the opinion is based almost entirely on narratives produced by opinion-formers in the mass media. If you attempt to press them to offer an explanation for this narrative, they will almost always mention a recent poll. But are polls evidence? No. Why do I say that? Because polls are not peer-reviewed and those who interpret the data aren’t self-reflexive. Moreover, the way in which respondents are led to a conclusion is rarely, if ever, discussed. Polling companies will preface a question with a statement like “It has been said that Jeremy Corbyn is unelectable”. Already, the notion that Corbyn is “unelectable” has been fixed in the respondent’s mind and they will offer a Pavlovian response on cue.  Remember, these are called opinion polls for a good reason.

Polls don’t exist to measure public opinion. On the contrary, they exist to shape it.

4 Comments

Filed under Opinion polls, Research, Society & culture

Polling Companies and the Conservative Party

Political parties, especially the Tories,  have a morbid fascination with polls. They see the polls and the companies that produce them as some sort of Delphic Oracle. What interests me isn’t the Tory fascination with polling companies but their involvement in them, since polling companies are always at pains to tell the general public that they are politically neutral. Yet, as any qualitative researcher will tell you, it is not possible to be 100% objective and put one’s ideology or cultural baggage to one side. The researcher must act self-reflexively. Bourdieu and Wacquant discussed this at some length in An Invitation to Reflexive Sociology. The researcher must consider their own position. Yet this idea of self-reflexivity appears to have escaped the pollsters. I have discovered that a number of Tory MPs are being paid by polling companies and there is no indication why they are being paid. Polling companies don’t deal in human vagaries: they and those who pay for the polls (usually the press and politicians) are interested in abstract numbers from which they hope to divine the future. That is the nature of quantitative research methodologies. They’re not particularly interested in why a certain number of people feel a certain way, because that’s the business of the qualitative researcher. The questions of what, how and why are of little interest to them. On Sunday I posted the following Tweet.

Andrew Hawkins of ComRes fired back.

My reply was blunt.

I was using my phone, so I didn’t have access to the necessary information to properly rebut his aggressive response. A couple of years ago, I was investigating Priti Patel’s business interests after I’d discovered she had close ties to the United Arab Emirates and to Bahrain, in particular. My research began after Patel had claimed she was a champion of Human Rights. I also knew Patel was involved in the right-wing Free Enterprise Group, which advocates among other things, abolition of the minimum wage to “grow the economy”. So I knew she wasn’t being entirely straight with her interviewer. I started my investigation by looking at the Searchthemoney website, and was astonished to discover that Patel had received £75 on 11 occasions between 2011 and 2013. It wasn’t clear why she was given this money nor was it clear what she’d done to deserve it.

I’d then noticed that many other Conservative MPs had also been paid money by ComRes. You can see the list here. Philip Davies, for example, was paid the same amount of money on more than 18 separate occasions between 2011 and 2014. YouGov is another polling company that’s cited for its supposedly rigorous methods. Yet this company was founded by Tories Stephan Shakespeare and Nadhim Zahawi. The company’s public face is Peter Kellner, who often appears on television to explain how the polls work. However Kellner isn’t as non-partisan as he seems to be. According to Lenin’s Tomb, Kellner intervened in the 2010 Labour leadership contest.

Firstly, Kellner uses figures relating the division of ABC1 and C2DE voters among the electorate to support his point that the number of ‘working class’ voters is declining precipitously. If he is right, then the proportion of ‘working class’ voters dropped from 51% to 43% between 1997 and 2010. That’s a rapid rate of employment change, though – given the way New Labour allowed manufacturing industries to collapse and shed employment – not all that incredible. However, the conception of ‘class’ deployed by Kellner is the old, misleading ‘social class’ model preferred by market researchers. His ‘classes’ (ABC1 = middle class vs C2DE = working class) are based on the National Readership Survey classifications derived from official statistics. As he revealingly puts it, according to his conception the middle class are those who work primarily with their brains, the workers primarily with their hands.

Kellner, according to Richard Seymour,  was a member of the Labour Party in the 1970s but soon swung behind Tony Blair in 1997. It’s a position that he’s maintained ever since. He’s also married to Baroness Ashton, a dyed-in-the-wool Blairite. YouGov also bungs money to Tory politicians. Caroline Dineage, for example, has accepted £280.00 from the company, while Philip Davies collected the cool sum of £1,030.00. Ipsos Mori has also slipped Tory MPs money and it seems as though there isn’t a single polling company that doesn’t do this. If there are any Labour MPs being paid by these companies, I haven’t managed to track them down yet.

Many political opinion polls are conducted as part of what is known to market researchers as an ‘omnibus survey’. The polls are often tacked onto the end of some survey about chocolate or soft drinks, or dropped into the middle of the survey on home insulation.

If polling companies want us to take their polls seriously, then perhaps they should tell us why they donate money to Tory MPs. They should also try and behave more self-reflexively. Say what you like, but I know what I’m getting from an Ashcroft poll; he’s a Tory donor and everyone knows it.

The main polling companies are members of the British Polling Council, which oversees standards in the industry. The founding members of this council are YouGov, Mori, NOP and ICM.  The BPC was founded to establish best practice in the industry to ensure validity and reliability. Prior to the formation of the BPC, polling companies did as they pleased and there was little, if any, accountability. Even so, there are questions that need to be answered but will the BPC or the polling companies answer them or will they obfuscate?

When it comes to objectivity I’m with Hunter S Thompson, who wrote the following about ‘objective’ journalism.

“So much for Objective Journalism. Don’t bother to look for it here–not under any byline of mine; or anyone else I can think of. With the possible exception of things like box scores, race results, and stock market tabulations, there is no such thing as Objective Journalism. The phrase itself is a pompous contradiction in terms.”

To adapt the Merovingian in The Matrix, objectivity is an illusion created between those with power and those without.

Reference

Bourdieu, P. and Wacquant, L.J.D. (1992). An Invitation to Reflexive Sociology. Chicago: University of Chicago Press

11 Comments

Filed under Conservative Party, General Election 2015, Government & politics