How opinion polls poison politics

loading animation
Friday, 10 February 2012 15:20 by Ad astra
Imagine this – a political world free of opinion polls seeking voting intentions and leadership preferences. What would journalists write about? Would they, as they once did, revert to writing about genuine political issues, giving their readers the facts and their well-reasoned analysis of them? Would they still be capable? Some would, for example Ross Gittins, who tells it the way it is without fear or favour.

Reflect on what we have read from political journalists these last few weeks. Ask what proportion has been devoted to ‘leadership speculation’. Would it have been possible to write about leadership speculation in the absence of poll data? Is it not poll data that has been the catalyst for the speculation? If there were no data on voting intentions or the popularity of the leaders, what would journalists use to promote the idea that a leader was under threat? They may of course have heard ‘corridor whispers’ from ‘informed sources’ that there was dissatisfaction with the leader and a desire for change, which no doubt would be amplified in their journalist watering holes where groupthink prevails, but with nothing more substantial than that it would be pretty hard to mount a case that there was a serious leadership challenge. It is the so-called ‘hard facts’ derived from opinion polls that fuels contemporary questions such as: ‘How long can Julia Gillard withstand such poor poll numbers?’ or ‘How can Labor recover from such a low primary vote?’ or ‘How can nervous backbenchers in marginal seats support a leader who is so unpopular with the people?’ or ‘As she has failed to increase her popular support since last year, is Julia Gillard’s leadership terminal?’ or ‘As the polls show Kevin Rudd is more popular, is there gathering momentum to revert to his leadership?

It is dead easy for even the most inexperienced and incompetent journalist to conjure up such questions and fire them arrogantly and sometimes rudely at the PM, her ministers and even her backbenchers at doorstops. Without polling data, all they could say was that someone back-grounded them on the dissatisfaction, without revealing the source. There does appear to be some stupid enough in the Labor Party to background journalists, but without hard data that would simply be gossip.

This piece contends that it is polling data that does the damage, not the whispers. I can hear some of you saying that polling data is important and should be sought and used. My reply is “What makes it important two years out from an election?” Does it have useful predictive value so far out? No. Do journalists give poll data the mantle of predictive power? Yes. Does it feed into the journalist’s echo chamber as fodder for a juicy story? Yes. Does it provide intriguing copy for news editors? Yes. Does it create lazy journalism? Yes. Has it resulted in the deskilling of many of our journalists, even the experienced? It seems so.

Is poll data ever useful? Yes. Close to an election, it can give political parties and commentators a guide about how political fortunes are evolving. This can be helpful in pointing to what needs to be done to improve those fortunes. But way out from an election, voting intention and popularity polls are misleading, often poisonous, and usually create mischief.

But we know they won’t go away. They generate too much income for polling organizations and the news outlets they feed, and countless stories for the press, hungry for sensational tidbits. We are stuck with these pernicious objects. A large industry has developed around them. There are the pollsters whose opinions are eagerly sought – Martin O’Shannessy for Newspoll and John Stirton for the ACNielsen Poll. There is a coterie of journalists that ‘specialize’ in poll interpretation – Dennis Shanahan is one example. There are websites devoted to polls: The Poll Bludger and the brilliant analytic site Pollytics, with its focus on aggregated polls and trends rather than individual polls. There are hungry news writers ready to pounce on any commentary they make, which they can quickly convert into a news item that then dominates coverage in papers, radio and TV news, and current affairs programs for the next day or two. Take away regular polls: the weekly, the fortnightly, the monthly, and the ad hoc polls, and a gaping hole is left in the political news. So let’s not kid ourselves that polls way out from an election have any other purpose but an institutional one. They are of no genuine help to governments, and help oppositions only when the polls are adverse to the government. We are being led by the nose by these self-serving organizations, and it’s high time we realized how we are being duped.

The history of political polling
Wikipedia has an interesting account of the history of polling:
"The first known example of an opinion poll was a local straw poll conducted by The Harrisburg Pennsylvanian in 1824, showing Andrew Jackson leading John Quincy Adams by 335 votes to 169 in the contest for the United States Presidency. Since Jackson won the popular vote in the full election, such straw votes gradually became more popular. In 1916, the Literary Digest embarked on a national survey... and correctly predicted Woodrow Wilson's election as president. Then, in 1936, its 2.3 million ‘voters’ constituted a huge sample; however, they were generally more affluent Americans who tended to have Republican sympathies. The Literary Digest was ignorant of this new bias. The week before election day, it reported that Alf Landon was far more popular than Roosevelt. At the same time, George Gallup conducted a far smaller, but more scientifically based survey, in which he polled a demographically representative sample. Gallup correctly predicted Roosevelt's landslide victory. The Literary Digest soon went out of business, while polling started to take off."

A brief history of polling in Australia, in the form of a PowerPoint presentation created during the Rudd era, shows that Sylvia Ashby began polling in 1908, followed by Roy Morgan. At the last election the main polls were by ACNielsen, Galaxy, Morgan and Newspoll. There are a few other minor polls conducted by news organizations. The Essential Poll is a relative newcomer.

The comments in this history are germane: Walter Cronkite is quoted as saying: “Australia: too many reporters, not enough news”. It insists that Australia has “too many reporters writing about opinion polls, too many opinion polls, and that too much attention is paid to opinion polls.”

It acknowledges that polls can affect behaviour via the “bandwagon” or “underdog” effect. It insists that they are “always beaten up by the journalists, can be paralysing, a tool of demagogues, pandering to irrational fears, have hidden agendas and biased questions, and are misrepresented.” But it concedes that “if read correctly, and modestly, polls are very, very useful tools.”

Without going into tedious detail, polls are only as reliable as the quality of the sampling and the size of the sample. Regarding the nature of the sample, some wonder whether the use of landline phones distorts sampling, as it may tend to sample less adequately the younger people who predominantly use mobile phones. Getting a sample that is truly representative of the opinions of the entire Australian electorate is the greatest challenge to pollsters. Because representative sampling in online polls is impossible, they are not only useless, but dangerously misleading.

A sample size of around 1,000 carries a margin of error of around 3%; with smaller samples (some may sample as few as 600), the margin of error rises. To reduce the margin to 1%, around 10,000 would need to be sampled, but this is too expensive for the pollsters. While pollsters acknowledge these sampling drawbacks, usually in fine print, they usually do not feature prominently in any commentary, so that readers tend to regard the figures as ‘gospel’ and attribute more significance to them than the figures warrant. Even minor deviations, within the margin of error, are given credence.

Are then the questions that are often added to those about voting intentions and popularity, worthwhile? In my view, if they are carefully worded, they can give useful information. For example, last Monday’s Essential Poll carried questions about the Aboriginal Tent Embassy, Aboriginal conditions and problems, the best party for handling Aboriginal issues; workers’ pay and productivity, IR laws, and trust in institutions. All these furnished useful information. Conversely though, poorly worded questions that lead those questioned to a predetermined position, can be dangerous, used as they sometimes are, with malicious intent to deceive. The problem is that the average person reading the poll is unaware of how manipulative the questions may be, and too often are.

Poll obsession
The obsession of political parties with poll results is a serious problem. They appear to attribute to them the status of holy writ. We have seen the Labor Party seemingly make major reversals of policy based on opinion polls and their opinion-seeking cousins, focus groups. The decision to defer action on an ETS by the Rudd Government seems to have been the initial reason for its loss of popularity and Kevin Rudd’s decline. We saw the delighted faces on Labor politicians after the recent less unfavourable ACNeilsen Poll, and can remember the long faces when the polls were steadily getting worse. We can be sure that the look on Labor faces after the next Newspoll will reflect its numbers.

When pollsters and the media that promulgates their results see the power they are able to exert over politicians, political parties and their policies, they relish their capacity to twist political arms and thereby exercise remote control over policy and the parties that develop them. This is just as Rupert Murdoch and his News Limited Empire would have it. If political parties were less preoccupied with every new poll, if they gave them less credence, the power of the polls, and the poisonous influence they now have, could be contained. The Gillard Government is trying to discount their importance, but the media is determined to counter this move so that the profound influence polls have over the political course of events continues.

The disturbing reality is that the media wants to influence politics, wants to mould public opinion, wants to foster its preferred party and leaders; in short wants to call the political shots, usually to fulfill its commercial agenda. Often this is at the bidding of wealthy vested interests that fund the media through advertising or who sit on media boards. We saw that flagrantly demonstrated during the minerals tax and carbon tax debates. We saw Gina Reinhart insist that Andrew Bolt be given his extreme right wing Bolt Report after she joined the Channel Ten Board. Opinion polls are simply one facilitatory device to aid and abet these agendas, which are sometimes overt, but too often covert.

Past polls feed future polls
One poll feeds into the next. As poll results are so widely publicized, those responding to the next poll are significantly influenced by prior polls. As few want to be the ‘odd man out’, and with groupthink operating powerfully, all except the real thinkers go along with the crowd, herd-like. This is why it is so hard to shift opinion polls; it needs either something radical to happen, or gradual attrition of entrenched attitudes as the facts overwhelm the prejudice. The Coalition knows how well its three word slogans have stuck in people’s minds and how that has influenced their response to opinion polls. So the slogans will persist, and its media supporters will continue to promulgate them and the disingenuous rhetoric it uses to justify them.

Opinion polls have a poisonous effect when misused, when used to pursue political agendas. Are they really necessary? Should they be curtailed?

Following the last piece on Edward de Bono’s Six Thinking Hats, let’s attempt to summarize this piece using that method:

The facts – The White Hat
Polls of voting preferences and popularity can be useful, especially as an election approaches.
The further out from an election they are, the poorer the predictive value they have; two years out they are of dubious value, and can have a self-perpetuating poisonous effect on public opinion.
Their value can be in guiding the strategy of political parties, politicians and political outlets.
To be valid and reliable, a representative sample and an adequate sample size are essential.
Many polls are inadequate in this regard and therefore of doubtful value, but this is not always stated.
Polls of political attitudes and values can be valuable provided the questions are carefully worded so as not to lead those being questioned.

The dangers – The Black Hat
Polls that have an unrepresentative and inadequate sample are dangerously misleading.
Polls of voting intention are misleading far out from an election.
Popularity polls can be illegitimately used to drum up discontent and dislodge ‘unwanted’ leaders.
Polls of attitudes and values can be misleading if poorly worded, or if designed to achieve a pre-determined outcome. Wrongly used, they can improperly influence or even form public opinion.
Polls can be used as a manipulative device by vested interests in the media and elsewhere.

The positives – The Yellow Hat
Polls of voting intention and popularity can be valuable close to an election.
Polls of attitudes and values can be valuable as a guide to those developing policy if carefully worded in a neutral way not designed to lead respondents to a predetermined outcome.

How do people feel about polls? - The Red Hat
Political watchers feel passionately about each new opinion poll, awaiting each with keen anticipation or trepidation.
Political operatives feel disappointed and anxious if the polls are adverse, and gleeful if positive for them.
Some political watchers feel angry at the way polls are misused.
Vested interests feel positive about polls when they are serving their interests, and angry when they are not.
Most people are indifferent.

Creative ideas about polls – The Green Hat
Would we not be much better informed if well structured professionally-designed polls of attitudes and values comprised the majority of polls, and those directed to voting intentions and popularity were fewer in number and confined to the last six months before an election?
Could an expanded set of attitudes and values be explored instead of the stereotyped ones so commonly assayed?
Could a more reliable and valid polling method be designed that had respectable and worthwhile predictive value?

So there it is. Let’s have your opinion about opinion polls.

Do you share my concerns and views about them?
Do you regard them as poisonous, as I do?
How would you like to see them conducted?
What suggestions do you have to make them more useful and valid than the present set we have thrust upon us every few days?