30 Oct 2012

Here Come The Polls!

By Ben Eltham
Why are we obsessed with political polls? Do they actually add a hard edge of data to speculation? Now that Labor's numbers are healthier, Ben Eltham turns to the polls - and the pundits who love them
The great American football coach Vince Lombardi once quipped that "if winning isn't everything, why do they keep score?"

It's a sentiment that will be readily understood in Canberra, where barely a week goes past without feverish speculation in the run-up to an opinion poll. The release of a poll is then generally followed by a post-orgasmic patter of political small talk, as the performance is dissected in statistical glee.

Since their invention as a form of political campaigning by George Gallup in the 1930s, opinion polls have slowly but inexorably moved to the centre of the political debate. Today they dominate the discussion of party politics in countries like the United States and Australia. Savvy poll bloggers have become some of the most influential commentators n the political media — voices like Nate Silver in the US, or Scott Steel and Peter Brent here in Australia.

The New Yorker's John Cassidy recently wrote a perceptive piece about the influence of Silver and his kin. In it, he pointed out that one of the most important things to understand about the modellers like Silver is this very influence. "People take it very seriously: he is widely regarded as a political astrologer who gets things right," Cassidy writes.

The same is true here in Australia. The political classes — the press gallery journalists, the bloggers, the professional campaigners and the politicians themselves — take the polls and the statistical models built on their data very seriously indeed. It's not just that they have predictive power — though quite clearly they do. And it's not just because opinion poll data offers a quantitative veneer to a business inevitably infected with bias and intuition. Perhaps the most important contemporary effect of opinion poll ubiquity is the signposting effect that poll trends can have for the political discussion more generally.

I'm talking about the general climate and the framing of political discussion that a clear trend on the polls tends to promote.

You can see the talismanic significance of poll data in the way politics is reported, particularly by the Canberra press gallery. A typical momentum shift is playing out in Canberra right now. What's happened? A series of opinion polls have shown a trend towards the government — the first such happy news for a beleaguered government in some time.

That's "changed the dynamic" — of whatever terms it is used by journalists like myself to gnostically refer to changes in political fortune. Again, the change is more about perceptions than it is about electoral reality or policy achievements, but it is no less important for that. For much of its second term in office, Julia Gillard's Government has appeared destined for defeat at the 2013 election. At some points the government has trailed the Opposition by an astounding 20 points in the two-party preferred analysis, portending an electoral wipe-out of historic proportions.

When a government (or for that matter, an opposition) trails in the polls by that much, it can be difficult for journalists and the general public to take it seriously. Seemingly every news article gets prefixed or suffixed with a one-liner about the party's dismal position in the polls. An air of inevitability sets in. Backbench morale can disintegrate. Leadership speculation surfaces.

On the other side of the coin, the party enjoying the big lead gets a boost in positive coverage. Little scrutiny is applied to their various statements. The election result can seem almost pre-ordained.

This is the position Labor has found itself in for most of 2011 and 2012, as poll after poll showed a government on the ropes. It was particularly marked around coverage of the carbon tax, which the opinion poll data showed was very unpopular. In contrast, Tony Abbott and the Opposition got a fairly easy ride, and plenty of coverage every time they decided to attack the government on carbon.

But in recent weeks, a ray of light has glinted. A number of recent polls have shown Labor closing the gap. The most recent Newspoll actually had Labor at parity. The latest Nielsen has Labor on 48-52. Any way you look at it, Labor is suddenly competitive again.

As usual, the commentary has been swift and speculative. Michelle Grattan argued that this means Gillard is now safe from the ever-lurking forces of Kevin Rudd. "Better polling and an altered political climate mean Gillard can shrug off questions about McKew's allegations," she writes. Dennis Shanahan made much the same point.

Attention now turns to the Liberal Party, with press gallery leader Phil Coorey reporting today that "several Liberal MPs" had told him that "Labor's recovery was now clearly a trend and Mr Abbott needed to broaden his approach beyond attacking the carbon tax.". As for poll blogger Brent, he's now arguing that the poll bounce will lead to the removal of Tony Abbott as opposition leader — which will in turn lead to the removal of Gillard as prime minister.

This is how the political circus moves on.

Polling doesn't even have to be public for it to be influential. One of the main revelations about Maxine McKew's new book about her time as a Labor parliamentarian, for instance, is that factional bosses were using private polling conducted by the party hierarchy to agitate for Kevin Rudd's overthrow. McKew argues that "a sense of crisis was actually created around Rudd's leadership and a principle tool in aiding and abetting that crisis was the use, or the misuse, of private party polling." The private polling reportedly showed that Kevin Rudd was perceived negatively in a series of marginal seats.

You can argue about whether that polling was accurate, and you can certainly argue about whether it was appropriate that party officials were apparently actively engaged in the white-anting of a sitting prime minster. But no-one appears to be rebutting McKew's central point about the persuasive power of those internal polls. The internal party polling changed votes in the Labor caucus. It helped undermine Rudd's leadership. That would not have been possible if poll data was not considered so important by the various players.

Informed observers tend to decry the influence of opinion polls on long-term policy reform, as George Megalogenis did in his excellent Quarterly Essay at the end of 2010. Megaologenis' point — that the constant drumbeat of poll data proves irresistibly distracting for the politicos and the media — is a sound one. But it is also a fact of contemporary political life. The constant spin cycle of political commentary demands a way to keep score. Opinion polls provide that. And when they move, the political system moves with them — to the next endless merry-go-round.

It's not quite meaningless: after all, we're talking about who's running the country. But it is certainly wearying. Like negative campaigning and personal attack tactics, the endless opinion poll commentary does not tend to improve the health of our democracy. It increases cynicism and degrades political engagement amongst voters.

But the opinion poll merry-go-round is not going to go away. Politics is the ultimate zero-sum game. Only one person can be prime minister. Only one major party can form government. And the people who care about that want to know who's winning. Even if we take away the scoreboard, the players are still going to keep count.

Log in or register to post comments

Discuss this article

To control your subscriptions to discussions you participate in go to your Account Settings preferences and click the Subscriptions tab.

Enter your comments here

Posted Tuesday, October 30, 2012 - 15:44

Thanks for a thoughtful piece. The conclusion is especially hard to refute, for mine, and yet that makes journalistic responsibility more important in discussions of these numbers. Part of what infuriates me is the way journalists report these numbers as though they are probably specific and probably accurate, when usually they are only accurate about a probability (95%) that they are approximate (give or take about 3%). So a two party preferred 'surge' of 4 or 5% is reported as a 'gamechanger,' when the maths says it may well be recording no change at all.

I note the ABC radio news has started talking about the 3% 'margin of error,' but even this exaggerates the reliability. For every poll result, there is a 5% chance that it misrepresents general attitudes by more than 3%. Nowhere in the Australian media do you hear or read the Canadian Broadcasting Commission's standard line that '19 times out of 20 these figures should be accurate to within 3%.'

That said, some of the longer term work being done through monthly 'polls of polls' (e.g. Andrew Catsaras) and the like are much more meaningful — but the media usually treats them as a sideshow, and keeps its freaks in the big top.

I wish this comment didn't seem so boring and worthy. I wish I knew more about statistics than I do. But most of all, I wish more journalists made more of an effort to help their readers and audiences understand the numbers, what they do and don't show.

Tom Clark
Victoria University (Melbourne)

Posted Wednesday, October 31, 2012 - 06:56

"the endless opinion poll commentary does not tend to improve the health of our democracy. It increases cynicism and degrades political engagement amongst voters."

I wonder whether this is, and whether it should be, the case. In a democracy government is meant to serve the will of the people, and this will is best assessed using precisely the methods utilized by posters. If the public feels that government is listening and responsive to their opinions (as measured through polling) then surely this should increase, not decrease, political engagement. It could even be argued that, in an era of modern statistics and cheap, instant communication, representative democracy is hopelessly outdated and ought to be replaced or supplemented by a system more directly reliant on the kinds of tools utilized by pollsters. Perhaps, for example, the Senate could be gotten rid of, and representative, informed samples of the populace could exercise a power of veto over legislation? That would certainly be a lot more democratic than our current system.

I do understand the feeling of weariness and pointlessness which the article identifies. However I don't think that this is a result of polling, but rather a result of the adversarial, personality driven character of our political climate. Any tool can be turned to good or bad ends, and I don't think that it is fair to see polls as the problem when, for example, politicians use them as a justification for rolling their leader, or when journalists superficially refer to them while they are busily obsessing over the personalities and electoral prospects of politicians rather than writing about the actual issues of the day.

Polls could and should raise the level of debate and make our society more democratic. If this isn't happening then perhaps we ought to be polling more, asking better questions, and analyzing the results less superficially? If the populace is disengaged and politics is degenerating into a popularity contest then it seems to me that this is very likely not because of polling but in spite of it.

That said, I am VERY sympathetic to the comments of Tom Clarke above. Certainly any journalist who does not have a clear understanding of what a confidence interval is has absolutely no business even mentioning the word "poll" in their writing. Perhaps the real problem is not an inflation of the importance of polling but the simple fact that most journalists and politicians don't seem to have much interest in learning how to read them. This does seem both a waste and a shame, as the basic statistics required for this kind of thing really isn't even remotely complicated.

K Brown
Posted Wednesday, October 31, 2012 - 09:39

The salutary lesson in the recent turn-around in the polls is that it has occurred following a solid year of "damn the torpedos - full speed ahead" Labor value policies being passed and tabled for future action. The public are at last seeing a National vision emerging, a Labor narrative developing and the political leadership that they have been craving and they like it. The "real Julia’’ is starting to look as though she does believe in something other than bland cliches like "reward for hard work" and fence-sitting on same-sex marriage. Principle and political courage shape the polls not the reverse.

Tony Abbott and the Coalition don't have two policies to rub together. Promises to repeal Labor's policies ain't policy. Politics led by public polls and focus groups is morally and intellectually bankrupt in the eyes of the public.

Posted Wednesday, October 31, 2012 - 12:56

The key passage in the article for me was "Michelle Grattan argued that this means Gillard is now safe from the ever-lurking forces of Kevin Rudd. "Better polling and an altered political climate mean Gillard can shrug off questions about McKew’s allegations" - and presumably about ANYTHING.

That encapsulates what is wrong with public discussion in scientifically-illiterate, anti-science, US- and Zionist-perverted, greed-driven, look-the-other-way Australia today under the disgraceful, pro-war, pro-Zionist, US lackey, pro-coal, pro-gas, pro-iron ore, anti-environment, human rights-violating Gillard Labor Government.

"The polls" won't make the following gross abuses of truth and humanity go away:

1. Gillard Labor is proposing to excise Mainland Australia from the "migration zone" i.e. it will be "disappearing" a whole 7,617,930 square kilometre (2,941,299 square mile) country.

2. Northern Territory Aboriginals have been excised by the Liberal-Laborals (Lib-Labs, Coalition and Labor) from the protection of the 1975 Racial Discrimination Act i.e. Australia has become a de jure as well as de facto Apartheid state.

3. Australia violates the UN Rights of the Child conventions not just over this "migration zone" excision and Aborigines but also over its complicity in under-5 infant deaths in Iraq (post-1990) and Afghanistan (post-2001) that have totalled 2.0 million and 2.9 million, respectively, 90% avoidable and in gross violation of Articles 55 and 56 of the Geneva Convention relative to the Protection of Civilian Persons in Time of War that demand that an Occupier must provide life-sustaining food and medical requisites to its conquered subjects "to the fullest extent of the means available to it".

4. Under Gillard Labor Australia is a world leader in annual per capita greenhouse gas (GHG) pollution. The WBGU, which advises the German Government on climate change, estimated in a 2009 report that for a 75% chance of avoiding a 2 degree centigrade temperature rise the world can emit no more than 600 billion tonnes of carbon dioxide (CO2) between 2010 and zero emissions in 2050. Australia's Domestic plus Exported greenhouse gas (GHG) pollution is so huge that it used up its "fair share" of this terminal budget in 2011. Indeed the bipartisan Lib-Lab (Liberal-Laboral) policies of a derisory "5% off 2000 GHG pollution by 2020" and unlimited coal, gas and iron ore exports coupled with Australia's huge resources mean that Australia is committed to polluting THREE (3) times the whole world's terminal pollution budget. Will the world permit us to do this? (for details and documentation see "2011 climate change course": https://sites.google.com/site/300orgsite/2011-climate-change-course ). Labor's "We are tackling climate change for a clean energy future" is an immense LIE.

5. Under Gillard Labor, Labor has now slashed university research grants (this cutting $500 million over 4 years) and is rushing (against Green and Coalition objections) to get the Senate to pass the Defence Trade Controls Bill (2011) which, according to Universities Australia (representing 39 Australian universities), “As currently drafted will significantly impact the training and research conducted by universities” and will interfere with the 3 key university purposes of teaching, research and informing the public. Universities committed to or subject to censorship are unfit for students, including the over 240,000 full fee-paying overseas students crucial to Australia’s $16 billion pa education export industry (see "Censorship by omission and commission by Australia’s ABC & The Conversation" : http://bellaciao.org/en/spip.php?article22270 ) .

Regardless of the polls, utterly betrayed decent Labor voters will vote 1 Green and put Labor last until it reverts to decent values.

Peace is the only way but Silence kills and Silence is complicity.

Posted Wednesday, October 31, 2012 - 13:59

I cant even remember the outcomes I predicted after an election, let alone the multitude of conflicting predictions from the past Polls. It seems to me that they are rhetorical and designed to influence an election and I have no doubt that, to a degree they do. Hindsight is also hindered, as predicted outcomes are washed off the street or obscured by ticker tape, after the event. Memories are replaced as opposite outcomes are trotted out and after a brief break the machine automatically turns on again, to inadequately fill the tiny space between Kardasian adventures and the newest hurricane to hit the USA.

Defense, Education, Health, Welfare, Energy prices and our long forgotten wars are all brushed aside as the drug and alcohol soaked electorate struggles to put the key in the ignition drive to the polling station and vote for the best of a bad bunch that few of us have ever met and will never get within striking distance of.

"In a democracy government is meant to serve the will of the people,"
Pan. This is the first I've heard of this, I was under the impression that it was meant to serve the will of the few people who control corporations. The rest of us can whistle Dixie while buying the products and breathing the fumes.

Posted Wednesday, October 31, 2012 - 14:10

I enjoyed this article Ben, even though I don't ponder over the use of populism. However popular politicians have gotten their dirtiest policies through the parliament in the best or worst of times & on most occasions, it's come back to bite them in the anal rectum but given them the highest ratings at the time.