University of New South Wales Law Journal
|
|
Home
| Databases
| WorldLII
| Search
| Feedback
University of New South Wales Law Journal |
|
A DELIBERATIVE DEMOCRATIC PERSPECTIVE
GRAEME ORR[*] AND RON LEVY[**]
Political opinion polling, especially about electoral matters, is ubiquitous. Ostensibly, polling takes snapshots of the mood of the electorate. Whether it is an accurate reading of that mood is, of course, unknowable. But polling results and trends are highly influential in making and shaping political policy, careers and even the fate of governments. In the lead up to elections, polling is widely assumed (and indeed marketed) for its predictive value. Occasionally, as in the United Kingdom (‘UK’) general election of 2015, such predictions miscue badly. Often it pre-empts debate about – and even drowns out – substantive issues. The question animating this article is: how, if at all, should such polling be regulated?
To ask that question in a liberal democracy is to invite debate about freedom of speech and the media. In the Anglophone legal world the debate tends to be foreclosed by an assumption that communicative liberties are trump cards. This may explain why the regulation of opinion polling has received little attention in most common law jurisdictions, even though the influence and extent of polling is as significant and controversial in those countries as elsewhere.
This article seeks to escape the confines of the ‘free speech rules’ rubric to consider the regulation of opinion polling by applying deliberative democratic concerns and arguments. As a normative theory, deliberative democracy is avowedly concerned with the quality of political discourse. Deliberative democrats are mindful of more jurisprudentially entrenched norms such as political liberty and equality, but are not enslaved to them. Deliberative considerations instead transcend these often clashing, rights-oriented principles to pay attention to how democratic judgments are made. In the case of opinion polling, the deliberative concern is with the role that statistical information derived from polling plays in the discourse and narrative of electoral debates.
The paradigm of polling in the electoral context involves quantitative surveys measuring voting intention and party leadership ratings. We are not concerned with ‘exit polling’, although it is capable of misuse. Exit polls chiefly measure actual voting behaviour rather than electoral opinion or intention. Whilst they sometimes also gather reflections on voter motivations, this has no immediate deliberative impact as it is rarely published until after voting closes.
Internationally, regulation of opinion polling during election periods is quite common even amongst liberal democracies. The literature on this regulation largely devolves, however, into a dichotomy between communicative freedom and electoral exceptionalism. ‘Electoral exceptionalism’ concedes priority to liberty but reasons that elections are exceptional occasions when other regulatory goals, notably integrity and equality in voting and between political parties, may displace liberty.[1]
Regulation falls into two broad categories: embargoes and disclosure. Embargoes, a more intrusive regulatory gambit, involve prohibiting the reporting, if not conduct, of electoral opinion polls during an election campaign. Disclosure, a lighter-touch regulatory approach, mandates the publication of key information such as sample size, margin of error and questions used, along with the initial reporting of poll results.
This article moves in four stages. The next Part offers an overview of the evolution of polling and of its perceived pathologies, to put the regulatory question in context. After that, Part III describes approaches to regulation found internationally. Aside from laissez-faire, these involve embargoes and/or disclosure. The judicial reaction to such regulation is also reviewed, revealing the classic split between judicial interventionism driven by strict readings of ‘free speech’ rights, and judicial reticence in accepting of the right of legislatures to accommodate other values.
Part IV of this article then brings in less well-recognised deliberative considerations. Here we argue that the dichotomy between free speech on the one hand, and deliberation on the other, is frequently a false one for speech in an electoral campaign context. Legal scrutiny (especially under rights guarantees) tends to imagine a competition between unfettered citizen expression, and the characteristics of good deliberation in a democracy – including access to information, time for reflection, etc. Yet, understood more thickly as democratic deliberation, democratic expression is a coherent compound of these elements. Free expression entails robust deliberation.
An artefact of legal ‘balancing’ or ‘proportionality’ tests – the notion of a clash between speech and deliberation – can be misleading and unduly reductive. As we will see, courts occasionally attempt to incorporate deliberative concerns into judgments about polling regulation. But the attempts have been inchoate, and seldom depart from presumptions that democratic speech and deliberation lie in tension. We note, therefore, how a basic conceptual confusion in common law countries’ legal treatment of polling regulation gets in the way of a rigorous appraisal of the merits of regulation. Often this means that legal tests undervalue the regulation’s rationales. After bringing deliberative democratic perspectives more squarely into the analysis of polling regulation, we evaluate alternative regulatory options accordingly.
As representative democracy took hold in the 19th century, it became increasingly necessary for politicians seeking election to ‘feel the pulse’ of their constituencies.[2] This can be equated with the need, identified by Jean-Jacques Rousseau, for politicians to understand the general will as a means to avoid misrepresenting ‘the people’.[3] James Madison, similarly, saw a need to build a representational equilibrium in the early United States (‘US’) republic so that factional politics would not lose sight of the interests of ‘the people’.[4] There is a social history, then, behind politicians surveying their constituents, whether motivated by electoral self-interest or more altruistic representative ideals.
Nineteenth century ‘surveying’, however, was qualitative to a fault. Politicians or their advisors would operate on fragmented information, heavily biased by interest groups and rife with methodological problems.[5] As the franchise grew, it became more difficult to understand what electors wanted or how one party was faring against another during campaigns. There was a need to shift opinion polling into a more systematic framework. That framework began to emerge first, in Anglophone polities, in the US in the early 20th century.[6] Local straw polls and mass postcard mail-outs, for example, were used to attempt to predict who would win elections.[7] But they were not methodologically robust.
In a famous turn during the 1936 US presidential election, George Gallup conducted a balanced survey of electors and predicted a comfortable victory for Franklin Roosevelt over Alf Landon (Landon had been pegged to win by Literary Digest style surveys which suffered from self-selection bias).[8] This turn led to a focus on analytical methods, drawing on market research into consumer attitudes and responses to commercial products.[9] This new style of polling articulated itself as ‘scientific’ polling.[10] Gallup and Elmo Roper, particularly, innovated with the statistical methodologies of political polling.[11]
Polling of course is no exact science: kinks and quirks are always being addressed and new methods experimented with. The literature abounds with cases of ‘opinion polling gone wrong’.[12] One of the best known examples involved pollsters predicting Thomas Dewey would triumph over Harry Truman in the 1948 presidential election – the reverse occurred.
While the market research industry is by now long-established, the
question of political opinion polling has become piquant in recent years, for interlocking reasons. One is the decreasing relative cost of polling.[13] Another is the intensification of media fascination with poll results and the easy narratives they breed. Combined, these forces encourage more polling. In itself, more polling is not necessarily bad, provided the polls have some scientific basis (unlike ‘vox pops’).[14] In electoral integrity terms, if nothing else, each poll can be a check on each other, and multiple poll results may be collated into larger data pools to generate more fine-tuned analyses and predictions. But a third reason for rising interest in polling is a general sense that political discourse is increasingly driven by superficiality and ideological reflexiveness, which opinion polls may breed or at least magnify.[15]
The desirability of the publication of some level of opinion polling is beyond question, particularly in relation to ‘issue’ polling. Issue polling asks questions about particular policies or themes, as opposed to asking respondents to rate leaders or to divulge voting intentions. The importance of polling of voting intentions for weak or emerging democracies should also be acknowledged, as repressing such polling may be a tactic to suppress popular opinion or a cover to make it easier to steal elections.[16] Polling is also useful to political parties as it helps them to adjust campaign strategies and undermine opposing positions, as well as being a tool for internal factional and leadership rivalry. Moreover, it is useful to media outlets and organisations to conduct polling. The numerical reductionism inherent in encapsulating public ‘opinion’ in a handful of percentages creates easily digestible headlines. These in turn generate interest in and help brand both media and market research firms alike as companies attuned to the pulse of community attitudes.
Its usefulness to citizens is less clear. The regulatory question for contemporary liberal democracies should be whether any pathologies involving the reporting of electoral opinion polling require regulation and, if so, what kind of regulation and at what cost. The plethora and frequency of electoral opinion polls has generated considerable concern, at several levels. There are many reasons why polling can never be an entirely ‘accurate’ product of an exact science; indeed it is part craft, part science. Polling can suffer from insincere responses from those surveyed.[17] It lends itself to cheating by those designing the survey (for example, through loaded questions designed to tilt results, or even to spread falsehoods, as in ‘push-polling’).[18] A lack of response from key sample populations may skew poll results.[19] There is also a lag between the design, execution and publication of opinion polls.[20] Even if pollsters’ methods were free from statistical bias, their efforts could not guarantee accuracy beyond a three per cent margin of error. This is because of both the phenomenon of ‘rogue’ polls (one in 20 samples will fall outside the margin of error, but there is no way of identifying the ‘rogue’) and because huge and unaffordable sample sizes are needed to reduce the margin of error further.
Over time, the cost of conducting polls has declined and statistical methodologies have generally improved. Margins of error have come down slightly, either through larger sample sizes or ‘polls of polls’ (where data from several comparable polls is agglomerated). Reliability has been enhanced by more sophisticated sampling and weighting, as has the forecasting of election outcomes through sub-jurisdictional polling of ‘battleground’ electorates. Modern elections, as a result, are experienced as if they were fairly predictable events: even uncertainty is assumed to be predictable, as tight races are labelled ‘too close to call’. Despite this faith in mathematical wizardry, polling remains problematic if understood as essentially a predictive tool. It can only capture present inclinations – a poll is not the moment of electoral judgment itself. Another problem is contemporary electoral volatility, as voters hew less staunchly to the established political parties.
Further, polling is itself an artefact – an artificial product – of a certain kind of politics. As with the observer-effect in quantum physics, in electoral politics polls can feed back into the very dynamic that generates that politics. The media, the chief sponsors of published polling, do not always conduct polling for altruistic or informational reasons. They also employ polls to charge the news cycle; polls have become an integral part of the headlines and chatter that dominate political discourse. Polls do not just reflect or report on a preset agenda. They are also means to (re)set that agenda and to apply pressure to politicians and representative government. By no means are polls just summative snapshots of stable demotic opinion; they also have a formative role influencing the development of that opinion. In short, polls are a significant element in political deliberation, at both elite and mass levels.
Jim Fishkin argues that polling’s empirical turn has helped to professionalise politics, but not in the ways that count. He reasons that what an ‘ordinary poll offers [is] a representation of public opinion as it is – even if that representation reflects no more than the public’s impressions of sound bites and headlines on the issue in question’.[21] It does not always take the public pulse, or make the will of the electorate easier to understand. Cold-calling of citizens measures knee-jerk and insincere responses as much as reflective and sincere attitudes.[22] Nor does it produce the equilibrium of political representation that Madison was after, as issue polling often just reflects partisan cues.
We can distil from this debate four pathologies of polling. Each of these criticisms is of particular concern to deliberative democracy. The first is that polling can be rather substanceless. It is a product of short-term electoral cycles, a valorising of the quantifiable over the qualitative and it is at the mercy of media spin. Polling may generate superficial conclusions[23] and skewed perspectives.[24] It often crystalises political issues in a binary way without offering alternative options, let alone arguments or solutions to problems.[25]
An excess of polling converts election campaigns into horse races.[26] As Simone Chambers notes, ‘[i]n the final weeks of the 2008 American presidential election, the public was inundated with polling data, all of it based on surveys asking some version of the question “who will you vote for on November 4?”’[27] Ideally campaigns would be a period of focus or contemplation of party programmes and manifestos. Antithetically to this ideal, polling generates a win/lose sporting atmosphere.[28] This metanarrative then swamps focus on first order questions of policy, candidate competence and so on. This criticism, as the Inter-American Commission on Human Rights has recognised, is ‘usually inscribed within a broader trend of mistrust of the relationship between the media and political processes’.[29]
From psychological and sociological perspectives, there is an analogy with debates about the regulation of sports betting and the live broadcast of odds (not coincidentally, betting on elections is re-emerging as a form of sports betting).[30] The analogy is not exact, since electors are participants and not merely spectators. But there is a shared concern that the experience of an enterprise – whether it be following sport or politics – is transformed in subtle but ontologically significant ways. The focus shifts from intrinsic aspects of the enterprise such as its moral or aesthetic dimensions and the strengths of the participants, to a focus on quantification and prediction of the outcome of the contest. In this shift, a complex sociocultural enterprise is reduced to a number.
A second, related concern is with a feedback-loop where the incessant
focus on polling may distort voting outcomes – whether that distortion comes in the form of a bandwagon effect favouring front-runners, the inverse underdog effect, or simply a self-reinforcing sense of inevitability, is another matter. A concrete manifestation is the phenomenon of ‘strategic voting’: electors may feel forced to hold their noses and vote not for their preferred candidate, but for some second-best candidate perceived as having a better chance at winning. This may then distort an election to the extent the election does not aggregate sincere
voter preferences, but instead reflects the voters’ presumptions and calculations about likely electoral outcomes. Some, like Wolfgang Donsbach, assert that
this phenomenon is both inevitable and democratically desirable.[31] But strategic conduct is contrary to the classic view of deliberation as ‘communicative reason’.[32] Whereas strategic electors plot how to help ‘their’ candidate, party or ideology win the political competition, more deliberative electors engage each other in discussion about the contest’s underlying substance in order to persuade or be persuaded.
A third potential pathology concerns the functioning of representative government. In Edmund Burke’s view (among the first on democratic deliberation in modern times), political representatives gain election to legislatures not to serve as mere conduits for constituent preferences, but to marshal their own expertise and deliberate carefully, with wide latitude, over their constituents’ needs.[33] Public opinion polling impinges on this ideal. In Simone Chambers’ words, ‘[p]oliticians use polls to find the median voter and then tailor their message to that voter. Re-election is the overarching interest when designing the message’.[34] A focus on polling surveys may then lead politicians astray from a deeper or longer-term consideration of the general interest,[35] skewing governments away from a Burkean trust/public interest model of representation and towards a delegate/‘what the public apparently demands’ model. This, it should be noted, is the elite-deliberation ideal of representative democracy. The more polling-friendly view recognises polls as a conduit of democratic feedback.[36]
A fourth and final risk is of politicians, lobbyists and an activist media manipulating poll results to their advantage. We see this when polls are used to disparage rival positions, but then discredited when it suits: something starkly apparent in the media coverage and party politicking that led to the deposing of three Australian Prime Ministers in 2010–2015 not through elections but by their own party caucuses. Polls themselves may be misrepresented, be misunderstood or even be misleadingly framed. This concern relates as much to the methods employed by pollsters and their publishers as the polls themselves. Is a given poll reflective of sentiments in the larger population? And can the choice of method give too much power to media organisations to sway election outcomes?
These concerns are, at heart, ones about the nature and efficacy of deliberation. Yet they also in turn implicate many of the wider democratic concerns that in recent decades have sharpened in the liberal-democratic world. We know, for example, that many electors are ‘switching off’ from politics due in part to the increasingly polarised, superficial and disputatious tenor of public discourse.[37] Some of these people may consequently pay less conscious attention to polls. But for others, polls prove catchier than the primary political
discourse they are turning away from, leading to a vicious cycle where the simple graphical and numerical reporting of polls comes to substitute for richer political dialogue. There is also the fear that citizens might disengage from those electoral contests seen as foregone conclusions (a fear heightened if electoral opinion polling data is taken to be an almost infallible predictor). On top of all this, assuming some of the problems just discussed are endemic, how can non-specialists detect these deficiencies without appropriate training? Polling is complex today. It is dependent on professional quantitative researchers with extensive methodological and statistical training. These researchers carry toolboxes beyond the reach or interest of most electors.
Some of the problems with polling, just discussed, are not static, but have generated various responses. In the mid-20th century, for instance, pollsters in the US overhauled their methodologies to identify better and worse polling practices. This has been followed by ongoing developments in statistical inquiry and sampling selection. Yet pollsters, media and governments have struggled to address the deeper pathologies of polling. This is unsurprising, given polling remains a largely self-regulated product of political and lobby groups, market research firms and media outlets competing for business, attention and influence. There is no neutral, state-based expert agency to lead the way, as there is in other vital predictive and data-gathering fields such as meteorology or economic and social statistics. The market research industry has developed ‘codes’, both to guide its members on ethical questions such as honesty and privacy,[38] and to assist the media and others to interpret polls.[39] It does this both in the name of self-regulation and to ward off state regulation. But it is still possible, for instance, to skew public perception through techniques that fall within acceptable norms (like the massaging of ‘rogue’ samples or the partial and loaded wording, ordering and selection of questions). Alongside industry bodies, civil society groups like Article 19, an organisation committed to media and informational freedom, monitor and agitate against regulation of polling.[40]
In contrast, government regulation has intervened to play a role in many countries. More than 90 national jurisdictions ban or delay exit polling, embargo the publication of electoral opinion polling during campaign periods or require the initial publication of polls to be accompanied by key information about their conduct.[41] Such regulation is often challenged as a restriction on freedom of speech. Proponents of regulation defend embargo periods (also known as ‘blackouts’) as important times when electors can ‘cool off’ and seriously consider their vote without being influenced by mass media, political parties or special interests. Some even argue that polling companies should be registered and subject to audits to ensure that polls are conducted professionally and ethically.
As was just noted, many countries legislate some limits on electoral opinion poll reporting: half of all democracies in recent surveys.[42] This proportion has been fairly stable in the past decade or so.[43] Whilst embargoes on publishing electoral polls during the campaign period are common enough, the trend has been towards relatively short embargoes, prohibiting publication only during the last few days of the election. That trend has in some instances been the result of judicial intervention.
The longest embargo currently in force appears to be Luxembourg, with a one-month blackout.[44] Of major democracies, several such as France, Italy, South Korea and Taiwan have had substantial embargoes extending longer than the final week of the campaign period. Italy prohibits opinion poll reporting during the final 15 days of the campaign (down from a previous embargo of 28 days).[45] South Korea’s embargo presently covers 21 days (down from 23 days).[46] Taiwan and Switzerland have embargoes lasting ten days (in the Swiss case, this is up from seven days). Spain retains an embargo covering five days before polling day. Many Latin American countries embrace embargoes: Argentina’s lasts over a week, Bolivia’s a week and Mexico’s four days. France currently restricts publication not only of polling, but campaigning generally, but now this covers only the final two days of its national campaigns (down from seven days).[47]
In contrast, Germany restricts only the reporting of exit polling before polls close. Japan has no embargo. Amongst common law countries, the US and Australia have no restrictions, even on exit polling.[48] New Zealand bans the conduct of exit polls; the UK merely bans their reporting prior to the close of polling.[49] And Canada embargoes only new polls on election day itself (down from three days, following judicial intervention).[50]
When it comes to mandating disclosure of information about each new poll, a majority of countries in one survey required publication of basic data about poll coverage, sample, currency and type of interview – although in four-fifths this was the result of self-regulation rather than state law.[51]
Exit polling is generally permitted, but typically (around 70 per cent of permissive countries) its publication is merely restricted until after polls close. Exit polling however, as noted earlier, occupies a different space from opinion polling. It can play an electoral integrity role – it is harder to steal an election through stuffing ballot boxes or manipulating the count if there are reliable exit polls. But exit polling does not impact on deliberation as such. It is primarily used in established democracies by media outlets seeking to win the race to call electoral contests before the final votes are counted. Scientific exit polling is not necessarily conducted even where permitted (for example, in Australia exit polling has only really been used for academic research into voter behaviour).[52]
In countries where legislation is subject to judicial review under constitutionalised principles of free speech, courts have taken different approaches to the validity of embargoes on polling. The overall tendency, however, has been to accentuate free speech considerations over all others, and to sublimate deliberative considerations. Although unfortunate, this is not surprising, since most bills of rights explicitly recognise liberty of political communication, without also expressly mentioning other values such as political equality, press responsibility or the quality of political discourse. Even in Australia, where there is no explicit ‘free speech’ protection, the implied freedom of political communication has been invoked as a trump card, and deliberative values downplayed.[53]
In Canada, a three-day, pre-election embargo on new polls was struck
down by the Canadian Supreme Court in Thomson Newspapers.[54] The applicant newspaper relied on the Charter of Rights and Freedoms protections relating to media freedom and to voters. The Canadian government only sought to defend the law on the narrow ground that a short ban insulated voters from potentially inaccurate or deliberately misleading last minute polls. The Supreme Court rejected this as paternalistic. The majority accentuated free speech rights. The minority was open to deliberative arguments, but felt constrained to rely on the narrow defence offered by the government, namely that late polls could not be scrutinised for misleading qualities.[55] As a result Canada now only embargoes publication of new polls on election day; just as it bans most election advertising on election day, to create a period of repose, a lull before the storm of the election results.[56]
The French Cour de cassation has also held a 10-day embargo on
publishing polls to be an undue breach of freedom of expression.[57] It suggested such blackouts could be discriminatory, since results might be published in nearby countries or on the web, and the media might still be informed by polling though the general public would not be aware of it. The Colombian Constitutional Court invalidated a 30-day ban, in part reasoning that those surveyed had a right to have their responses publicly expressed.[58] But it left leeway for the legislature to regulate a shorter ban. The Colombian Constitutional Court subsequently accepted a ban on polling day itself, reasoning that on election day ‘any voice that is not the voice of the people shall remain silent’.[59] The Philippines Supreme Court invalidated a 15-day embargo.[60]
In contrast, the United Nations Human Rights Committee upheld a (23-day) campaign length ban in South Korea.[61] A journalist had challenged the law, after being fined approximately US$500 for publishing new poll results, seven days out from polling day. The South Korean Constitutional Court upheld the law, citing the potential for both bandwagon and underdog effects. The journalist then argued that: (a) the effects were speculative and might cancel each other out; and that (b) suppressing electoral polls created an unfair divide between those with access to such information and those without. The majority of the Human
Rights Committee nonetheless upheld the ban, as a reasonable measure to ‘provide the electorate with a limited period of reflection, during which they
are insulated from considerations extraneous to the issues under contest’.[62] A dissenting opinion, from the US delegate to the Committee, argued that whilst ‘[s]ome might welcome an interval in which elections were not discussed as a horse race’ a ban also hobbles the ability of the media to discuss the impact that candidates’ or parties’ positions or actions might have on the race. Opinion polls, the dissentient argued, could act as ‘part of the conversation between candidates and citizens’.[63]
The weakness of most of these judgments is their failure to engage with deliberative considerations. This reflects a lack of legal imagination, rooted in a narrow conception of the enterprise of electoral politics. In this jurisprudence, individualised liberties are championed in a relative vacuum. The interests of the media, in publishing polls they commission, tend to be conflated with the question of how and whether polls represent ‘expression’ of the people who happen to be surveyed. This question would be particularly important in a country like Australia, where constitutional protection of ‘political communication’ derives not from any explicit protection of free speech, but the institutional importance of a dialogue amongst electors, and between the government and the governed. The publication of polls on issues during a term of government falls within this protection. But it is far from clear that voting intention polls, conducted whilst a government is in caretaker mode, amount to political communication in the relevant sense.
The counterarguments permitted to topple these liberties have, to date, been limited to a certain type of anti-corruption argument. Thus, the dissenting judgment by Kapunan J justifying the 15-day embargo on polls in the Philippines cited misleading polls, the ‘junking’ of ‘losing’ candidates by parties and electoral cheating (it is easier to steal an election if one knows in advance the scale of vote-rigging required).[64] Within these judgments, however, are seeds for a debate about the deliberative rationale for and against regulating polling. On the one hand lies a claim that a period of repose to focus on the first-order issues of the campaign is more valuable than an abstract liberty to publish anything, anytime. This is consonant with electoral systems that limit campaign expenditure or broadcast advertising, especially in a blackout period for tranquil reflection prior to polling day.[65] Electoral opinion polling is often a reflection of crude responses to personalities and party brands. Overdone, its horse race narrative corrodes the debate about the substance of issues. Free elections and pure freedom of expression are not the same thing.
On the other hand, there is a counterargument that polling provides useful information. It generates, if not a dialogue, at least a kind of democratic toing and froing between electors en masse and political figures over issues. In any event, behind-the-scenes polling will occur and any suppression of poll results only creates a divide between the broader public and those ‘in the know’ (for example, party apparatchiks, journalists on their drip, or those able to source polling information from outside the country especially via foreign internet sources).[66] Such withholding of information is often labelled a ‘paternalistic’ approach ‘that does not coincide with the idea of democracy’.[67]
In 1999, the Council of Europe affirmed a broad interest in the publication of opinion polls. But it qualified this finding by recognising some fundamental weaknesses in the informational value of baldly presented polling data.[68] The value of any particular poll, and hence of polling collectively, the Council recognised, can only be assessed if polls are accompanied by sufficient information to enable the public to judge the value of the poll.[69] This includes the identity of the body that commissioned the poll, the organisation that conducted it, the broad method and questions used, and the dates of the survey, its sample size and margin of error. In theory this goal may be achieved by formal law or by media self-regulation. A relatively early French law, from 1977, gives a taste: election period polls had to be notified, along with key information about each poll, to an ‘Opinion Polls Commission’. Italy similarly requires publication of key information. Two recent examples of national law are notable: the Canadian and the Kenyan.
Just prior to its 2000 election, Canada enacted a provision requiring disclosure of key information about electoral polling (this type of regulation
had been endorsed in Thomson Newspapers, as more proportionate and
hence justifiable than a ban or embargo).[70] The bodies that commissioned and conducted the poll must be revealed. So too must the wording of the questions, the pool and sample size and the margin of error; a methodological report must also be made publicly accessible. Where a survey lacks statistical validity, that fact must be published instead. The duty falls both on the ‘first person who transmits the results of an election survey during an election period’ and anyone who republishes them within 24 hours.[71] The law is predicated on an assumption that individual polls have a limited shelf life; the cumulative effect of polls is ignored.
A similar model of disclosure law was recently legislated in Kenya. The Publication of Electoral Opinion Polls Act 2012 (Kenya) requires those reporting on polls within the first day of their release also to report the poll’s source, size, margin of error and (in print reports) wording of questions. These requirements apply during the year prior to an election.
Most scholars trace the beginnings of the modern deliberative literature to Jürgen Habermas and John Rawls, although, especially since the 1990s, a host of others have also contributed to the field. Deliberative democracy theory assumes those subject to collective decisions should have voices in the process.[72] But it also aims for decision-making to be characterised by an exchange of reasons in which participants persuade each other based on what Habermas termed the ‘force of the better argument’.[73] For his part, Rawls stipulated (among other relevant notions) that democratic speakers ideally should put ideas to each other using relatively rational and widely understood forms of argumentation that others ‘may reasonably be expected to endorse’.[74] Such deliberative analyses and conditions of collective decision-making were long overshadowed by the view that democracy is primarily a process of preference aggregation. However, following its rise, deliberative democracy research – which has normative, empirical and often practical institutional strands – casts doubt on this old certainty. Some prominent expectations of deliberative democratic governance include that it should be characterised by relatively high levels of rational persuasion, open-mindedness (rather than more adversarial ‘agonism’),[75] other-regarding,[76] civility and cooperation,[77] reflection,[78] equality and inclusivity,[79] broadly-sourced information,[80] policy holism (that is, awareness of the need for policymaking trade-offs)[81] and other relatively rigorous forms of debating and reasoning.
Deliberative democracy forces us to focus on the quality of political deliberation and debate. In the process, freedom of speech is understood not as a constitutionally entrenched guarantee or end in itself, but an element of a thicker conception of expression.[82] Free speech is not an unbridled right clashing with and trumping other interests. Rather it is a value accommodated as part of a broader democratic space ordered to help foster the ultimate goal of productive and informed deliberation.
Proponents of unregulated polling, like Thomas Petersen, assert that it is ‘inconsistent to demand ... that voters make their decision as rationally as possible, based on factual information, but then ... to forbid the publication of such factual information.’[83] This looks like a deliberative argument resting on a claim that all polls are informative. The argument’s apparent simplicity is seductive. Yet it hides naïve assumptions about deliberation, for example the implicit claim that all information is good however tangential to the task at hand, and that the condition of deliberation is the simple provision of information.
Since polling, at best, involves a society holding a mirror to itself, it is a recursive exercise. Information about attitudes may feed back on itself. Kurt Lang and Gladys Engels Lang posit that ‘there are three ways in which polls can influence the public opinion they purport merely to measure’.[84] The chief effect they admit is that advocates of positions labelled, thanks to polling, as distinctly ‘minority’ views may be discouraged, ensuring that those views remain minority positions. A recursive effect also arises because polls may arouse interest in the respondents themselves. This is a particular concern when polls are tendentious – especially with push-polling, the sowing of misleading information in the guise of polling.[85] But we are focusing here on reputable polls. The numbers of people surveyed in any one poll or even year is still a small subset of the general population so the effect of polling on those polled is small.
A third form of direct influence of polls on opinion is the vaunted bandwagon effect. This involves electors being infected by enthusiasm for a frontrunner party or position, naïvely assuming that popular positions are necessarily the most rational ones, or simply deciding to back a ‘winner’ because it is psychologically more rewarding than backing a ‘loser’.[86] Lang and Lang detect limited evidence for this effect in voting outcomes; a meta-analysis of research suggests that evidence for electoral bandwagon is confined to apathetic voters or may be balanced out by an ‘underdog’ effect.[87] This argument about an underdog effect muting any bandwagon effect is not, however, a good argument from the point of view of better deliberation. It is only an argument about macro-electoral effect. Deliberation is concerned less with outcomes per se, and more with the discussion and reflection involving individuals and the social whole.
Ultimately, the effects of polling remain a live question for empirical study. Whether polls have an undue influence on voting behaviour directly, and
hence electoral outcomes, remains an open and disputed question.[88] As a result, proponents of embargoes on polling advocate a precautionary principle, whilst proponents of a broad liberty for pollsters insist on a stricter constitutional logic. To a deliberativist, however, the question of direct effect on voter behaviour and electoral outcomes is somewhat beside the point. The deeper question is the health of public discourse. As Susan Herbst observes:
Conversation ... is fundamental to the construction of a democratic public sphere ... In a way, polls may make many political discussions superfluous, since they give the illusion that the public has already spoken in a definitive manner.[89]
It is true that many surveys of voting intention are prefaced with words such as ‘if an election were held today ...’. In reality, injecting such a contingency into a survey muddies, rather than clarifies, the question of what is being measured. Polls purport to take a present-day pulse. Yet they involve play-acting, by inviting respondents to imagine a counterfactual. This play-acting is rendered all the more complex because respondents are aware that their responses will feed into a recursive political dynamic. Sincerity in such a situation easily gives way to partisan responses to issues or to inflating complaints in the hope that they gain traction. Such manoeuvres by respondents highlight the intricate dynamic of polling outside an electoral campaign. However blunt polling may be, issue-oriented polling is a legitimate influence on policy-making. It acts as a routine part of the democratic centrifuge, counterbalancing the elite tendency of representative governments that are subject otherwise to only occasional, formal electoral accountability.
When an election is called, this dynamic shifts. The question of voting intention is no longer hypothetical. Party platforms and pitches are set in place, and hence no longer subject to the democratic centrifuge. Pollsters assert that this renders polling less volatile and more reliable in a predictive sense. To a deliberativist, however, this misses the point about the nature of discourse.
It is not the undue influence of any particular poll, but the effect of the unconstrained flow of polling that is potentially pernicious. Polling is reported as ‘factual’ with little incentive for the media, pundits or pollsters themselves to nuance reporting with explanation of each poll’s limitation. As Sally Young relates, the media is fixated on drama, and hence polls are routinely spun to create a sense of change and instability, even if the ‘change’ is well within the margin of error.[90] The plethora of polls in most western countries also generates a one-dimensional metanarrative. This narrative sublimates politics – understood as a first-order dialogue and argument over policy, vision and leadership – to a conception of politics as a game or a second-order object of study. This prioritises two objectively narrow perspectives over any wider interests. The ‘game’ metaphor reduces political rivalry to a relative poll ranking and narratives about inter-party strategy, while the ‘object of study’ approach reduces political commentary to the psephological concerns of media observers and social scientists.
The excessive focus on polling risks campaign discourse being sidetracked into a hall of mirrors, with everyone reflecting on how everyone else is leaning, without deliberating on why. Parties are left to worry about bandwagon or reverse bandwagon effects, which certainly affect the morale and motivation of their activists, even if they are not substantial determinants of voting behaviour. Several studies have detailed voter dissatisfaction with the ubiquity of opinion polls during elections and referendums.[91]
The rationale then, for regulation of published polls is heightened during election periods compared to the years beforehand. During times of policy formation, opinion polls are part of the communicative interplay from governed to governors. But contrary to naïve libertarian (and occasionally judicial) presumptions, election campaigns are not febrile political free-for-alls. As Frederick Schauer and Rick Pildes, amongst others, have reasoned, election periods are politically ‘exceptional’.[92] This is not as paradoxical as it may seem at first glance. Election campaigns are unique occasions and are typically structured to prioritise goals such as a measured focus on the parties vying for power. As a result, public funding and restrictions on campaign finance, requirements of broadcasting balance, and limits on anonymous speech and donations, are commonplace.
A genuine paradox however arises in that relatively short embargoes on electoral opinion polling – of the kind most common today, covering just the last few days of the campaign period – are harder to justify on deliberative grounds than an embargo lasting the whole of that campaign.[93] Short bans reflect an aesthetic touch. They form a brief period of repose prior to polling day in which undecided voters focus on whether and how to vote, and campaign machines and other citizens consider the logistics and ritual of election day itself.[94] Deliberative concerns with electoral opinion polling extend beyond such limited horizons.
It is not, of course, completely irrelevant from a deliberative perspective for citizens to be concerned with the fluctuations in opinion polling. To a purist who thinks elections should be determined only by an analysis of contested policy and its appeal to voters’ individual values and interests, this might sound odd. But elections are not so simplistic. The nature of electoral choice is ‘over-determined’ – an election is not a referendum focused on a single issue.[95] Inevitably, and reasonably, myriad factors motivate voting behaviour. These range from ideology and loyalty to party, personal and class identity, the multiplicity of policy proposals, overall assessments of economic and social welfare and security, and perceptions of trust and competence.
Moreover, good deliberation is partly ‘other-regarding’. In other words, in a democracy people should consider the views and interests of others.[96] That consideration need not be limited to conversations with others; as Robert Goodin argues deliberation typically involves citizens being ‘imaginatively present’ to each other and to legislators.[97] This is one of the normative aspects of the deliberative approach: democracy is not reducible to the utilitarian sum of nominal self-interests, but is an attempt to reconcile competing interests and accommodate a variety of values. This is not to concede that opinion polling is an infallible source of information about fellow citizens’ opinions. Nor does it offer direct insight into the reasons for those opinions. But it offers much more reliable evidence than, say, letters to the editor or ‘vox pops’, which are not statistically sampled and can easily be distorted by editorial curation.[98]
Given a mass and often segmented electorate, opinion polling can thus be a deliberatively useful source of reliable information about the plurality of values and their spread among different subgroups (poll results are often broken down by age, gender and geography). For citizens to adapt their positions on an issue as they become aware of the acceptance of rival positions is not capitulation but moderation. Opinion polling may not reveal the justifications for those alternative positions, but it brings those positions to light and measures fluctuations in their relative support. Issues-based opinion polling, in particular, can also be a means for interest groups to raise awareness of neglected issues and thereby broaden the deliberative and policy agenda.
There are good reasons, then, to distinguish between issues-based opinion polling and polling based purely on voting intention or leadership ratings. Issues-based polling can assist deliberation by politicians and bureaucrats in framing policy, and by citizens in considering such policy debates. It can do this by providing information about (shifting) values and adjusting the agenda. Pure electoral opinion polling, however, especially during the campaign period, is a reductive and over-used device. Its publication can legitimately be prohibited or embargoed once the campaign period commences, to help ensure there is a balanced and deliberative focus on the first-order electoral issues and questions. This would address the pathologies of wall-to-wall polling stoking substanceless, horse race narratives during elections, as well as the problem of polls feeding back onto electoral opinion. Yet it would still permit polling during the term of government, and hence not forsake the deliberative democratic value that polls can have in educating both citizens and politicians/bureaucrats as elite deliberators alike, about public opinion on developing issues.
The deliberative benefit of publication of either type of polling – issues-based throughout the political cycle, and electoral opinion polling outside the campaign period – also needs to be strengthened. This can be done through laws requiring disclosure of key information about the source and methods used in those polls, as with the laws in Canada and Kenya. This would enable deliberation about the value of each poll, and enhance the integrity of polling itself by minimising its manipulation by both the sponsors of polls and the wider media.
These are normative recommendations for reform. We are also arguing here against the one-dimensional assertion by courts (and industry bodies) that ‘free speech’ trumps any legislative decisions to regulate, whether by election period embargo or disclosure requirements. Of course, as we noted earlier, banning the publication of electoral opinion data risks creating a class of political insider who has access to it, who may selectively leak such privileged information to the media.[99] But that already occurs, often for mischievous strategic purposes. Under our proposal, the embargo would equally apply to the off-hand publishing of such leaks as to the publishing of formally commissioned electoral surveys.
Disclosure, as we noted earlier, is a ‘soft’ form of regulation, compared with embargoes. Disclosure regimes do not of course guarantee every member of the general public will become savvier about the limitations of opinion polling. But at a minimum, mandated disclosure would direct the attention of journalists and commentators to those limitations. Pollsters, themselves no fans of regulation, do not deny there is a problem to be addressed. An industry survey of some 60 countries with widespread polling found that only 31 per cent of pollsters thought that journalists had a fair or better grasp of polling; in contrast 43 per cent thought the quality of journalistic treatment of poll data was ‘low’ or ‘rather low’.[100]
Besides often simplistic appeals to liberty, those who defend opinion
polling as a benefit for good government tend to rely on arguments about equality. Polling, in this view, is an inherently democratic channelling of
broad-level public opinion. Provided it is conducted scientifically and not tendentiously, polling may act as a counterweight to lobby groups who have the ear of politicians and bureaucrats.[101] It can also act as an antidote to media misrepresentation of the zeitgeist. To the optimist, then, polling acts as a conduit whereby collective sentiment is publicly manifested and communicated upwards to elites. It possesses value because of ‘who’ it includes.
Those who are wary of the value of opinion polling for good government tend to rely on arguments about quality. The essential weakness of polling is the very thing that makes it seem democratic: random sampling, by cold-calling, across a large population. To many of those surveyed, polling questions lack salience and are asked so rapidly and baldly that they measure a gut-level reaction at best. Ideally, pollsters should seek to measure this (through ‘no opinion’ options), but pollsters often try to force or massage responses into
one box or another.[102] For the pessimist then, polling acts as a conduit for unconsidered, reflexive and not genuine ‘opinion’. It lacks value because of ‘how’ it is conducted.
Equality of voice or quality of opinion: this dichotomy reflects an old conundrum for mass electoral democracy. Who will be heard, and with what weight?[103] Pathologies of opinion polling, perceived and real, have led in recent decades to regulatory debates and measures. The general public is not especially knowledgeable about the debates surrounding polling. Polls are beloved of political insiders, especially given their strategic usefulness. Opinion polls can play an important, if problematic (sometimes shallow, sometimes distorting) role in contemporary politics.
Our concern here has not been with every type of poll and every context in which they are used, but with the specific question of electoral opinion polling and deliberative arguments for and against its regulation. The issue is now on the agenda, even in Anglophone democracies. A UK Lord has proposed establishing a ‘Political Opinion Polling Authority’ to regulate polling of voting intentions. The proposed authority, established by but independent of parliament, would make rules on disclosure and pre-election embargoes and even approve sampling methods and advise on good practice in the framing of questions. Whilst the Bill is unlikely to be enacted, it reached its third reading stage in the House of Lords in late 2015.[104]
We have sought to bypass hoary debates entrenched in reasoning from the values of liberty, equality and integrity, and instead to locate opinion polling and its regulation as a question for deliberative democracy. Put simply, deliberative democracy provides a thicker, less conceptually tenuous reading of what free expression requires. Law informed by deliberative motivations seek to understand and shape the rules of politics in ways that accommodate deliberative aspirations. Better deliberation is not merely a legitimate public aim (which regulators, whether legislative or judicial) should ‘balance’ against political freedoms. Rather, it makes better sense of such political values than if we imagine them as rights subsisting in a vacuum.[105]
A key problem is the sheer frequency of and media obsession with electoral opinion polling. The weight of such polling frames electoral politics as a race and reduces it to a simplistic statistical measure. It would be naïve to think that electoral politics does not involve a contest: indeed the framing of elections is already infused with belligerent metaphors.[106] But it is not just an antagonistic contest for victory at the polls; underlying it is meant to be a richer contest over political ideals and ideas.
[*] Professor, TC Beirne School of Law, University of Queensland.
[**] Senior Lecturer, College of Law, Australian National University.
The authors thank Greg Dale, Samuel Rutherford and Dr John-Paul Gagnon for their research help. They also acknowledge the support of the Australian Research Council (Discovery Project, ‘The Law of Deliberative Democracy: Theory and Reform’, DP 130100706).
[1] Fredeck Schauer and Richard H Pildes, ‘Electoral Exceptionalism and the First Amendment’ (1999) 77 Texas Law Review 1803, 1805–7.
[2] See generally Sidney A Pearson, ‘Public Opinion and The Pulse of Democracy’ (2000) 42 Society 57. See also Michael Saward, The Representative Claim (Oxford University Press, 2010).
[3] Kevin Inston, ‘Representing the Unrepresentable: Rousseau’s Legislator and the Impossible Object of the People’ (2010) 9 Contemporary Political Theory 393, 400–1.
[4] Harvey C Mansfield, ‘Democracy and Populism’ (1995) 32 Society 30, 30, citing James Madison, ‘The Federalist No 10’ in Terence Ball (ed), The Federalist with Letters of ‘Brutus’ (Cambridge University Press, first published 1787, 2003 ed).
[5] See generally L John Martin, ‘The Genealogy of Public Opinion Polling’ (1984) 472 Annals of the American Academy of Political and Social Science 12.
[6] Katherine A Bradshaw, ‘“America Speaks”: George Gallup’s First Syndicated Public Opinion Poll’ (2006) 31 Journalism History 198, 199.
[7] Archibald M Crossley, ‘Straw Polls in 1936’ (1937) 1(1) Public Opinion Quarterly 24, 24–5.
[8] Dominic Lusinchi, ‘“President” Landon and the 1936 Literary Digest Poll: Were Automobile and Telephone Owners To Blame?’ (2012) 36 Social Science History 23, 24.
[9] For a discussion of market research methodology in the early 20th century, see generally Roger S Mason, ‘Breakfast in Detroit: Economics, Marketing and Consumer Theory, 1930 to 1950’ (1998) 18 Journal of Macromarketing 145; Gerben Bakker, ‘Building Knowledge about the Consumer: The Emergence of Market Research in the Motion Picture Industry’ (2003) 45(1) Business History 101.
[10] Martin, above n 5, 22–3.
[11] Sarah E Igo, ‘“A Gold Mine and a Tool for Democracy”: George Gallup, Elmo Roper, and the Business of Scientific Polling, 1936–1955’ (2006) 42 Journal of the History of Behavioral Sciences 109.
[12] See generally Adam J Berinsky, ‘American Public Opinion in the 1930s and 1940s: The Analysis of Quota-Controlled Sample Survey Data’ (2006) 70 Public Opinion Quarterly 499; Amy Fried, Pathways to Polling: Crisis, Cooperation and the Making of Public Opinion Professions (Routledge, 2012).
[13] Including via electronically managed surveys of online sample groups or ‘robo-polls’ using automated telephony.
[14] A ‘vox pop’ is a survey inviting readers to respond online, or viewers to respond by SMS. Vox pops involve no statistical sampling and are easily swamped by supporters of one viewpoint or another encouraging each other to take part, sometimes multiple times.
[15] See Stephen J A Ward, Ethics and the Media: An Introduction (Cambridge University Press, 2011) 108–9 on how the modern media and political discourse may enhance participation but not deliberation.
[16] Humphrey Taylor, ‘The Value of Polls in Promoting Good Government and Democracy’ in Jeff Manza, Fay Lomax Cook and Benjamin I Page (eds), Navigating Public Opinion: Polls, Policy, and the Future of American Democracy (Oxford University Press, 2002) 315.
[17] An oft-cited example is the supposed ‘shy-Tory’ effect, whereby some respondents intending to vote for the incumbent Conservative Party hid that fact, skewing expectations and discourse in the 1992 UK general election. Additionally, interviewers themselves may exhibit bias: Leo P Crespi, ‘The Cheater Problem in Polling’ (1945) 9 Public Opinion Quarterly 431.
[18] On push-polling, see below n 85 and accompanying text.
[19] Contemporary difficulties involve reaching people who only own mobile phones, who live in gated communities, or who are overloaded with spam calls/emails and may be experiencing wavering interest in electoral politics.
[20] Charles K Atkin and James Gaudino, ‘The Impact of Polling on the Mass Media’ (1984) 472 Annals of the American Academy of Political and Social Science 119, 123.
[21] James S Fishkin, ‘Consulting the Public through Deliberative Polling’ (2002) 22 Journal of Policy Analysis and Management 128, 128. Fishkin’s response is the ‘deliberative poll’, which records changes in opinion after sample groups engaged in deliberative debate on particular policy issues: at 128. We do not consider deliberative polls here, as they are not intended to measure and report raw public opinion but to be a transformative counterpoint to such opinion.
[22] Compare Zaller’s influential work on how ‘opinions’ are often constructs of the polling process, rather than any actual ‘preference’ held prior to the survey itself: John R Zaller, The Nature and Origins of Mass Opinion (Cambridge University Press, 1992).
[23] Ibid; see also Fishkin, above n 21, 128–9.
[24] See generally Robert Weissberg, ‘The Problem with Polling’ (2002) 148 Public Interest 37; Elias Walsh, Sarah Dolfin and John DiNardo, ‘Lies, Damn Lies, and Pre-election Polling’ (2009) 99 American Economic Review 316.
[25] Angus Campbell, ‘Polling, Open Interviewing, and the Problem of Interpretation’ (1946) 2(4) Journal of Social Issues 67, 68. See generally Drew Altman and Brodie Mollyann, ‘Opinion on Public Opinion Polling’ (2003) Health Affairs: Web Exclusive W276 <http://content.healthaffairs.org/content/early/2002/08/14/hlthaff.w2.276.full.pdf+html> .
[26] Thomas Fitzgerald, ‘Public Opinion Sampling’ (2002) 39(6) Society 53, 53.
[27] Simone Chambers, ‘Rhetoric and the Public Sphere: Has Deliberative Democracy Abandoned Mass Democracy?’ (2009) 37 Political Theory 323, 343.
[28] Atkin and Gaudino, above n 20, 124.
[29] Eduardo Bertoni, Annual Report of the Inter-American Commission on Human Rights 2005, Organization of American States Doc OEA/Ser.L/V/II.124, Doc 7 (27 February 2006) vol III, 154.
[30] See generally Graeme Orr, ‘Betting on Elections: History, Law and Policy’ (2014) 42 Federal Law Review 309.
[31] Wolfgang Donsbach, ‘Who’s Afraid of Election Polls?: Normative and Empirical Arguments for the Freedom of Pre-election Surveys’ (Research Report, The Foundation for Information, European Society for Opinion and Marketing Research, 2001) 13–14.
[32] Jürgen Habermas, The Theory of Communicative Action: Reason and the Rationalization of Society (Thomas McCarthy trans, Beacon Press, 1984) vol 1, 285–6 [trans of: Theorie des Kommunikativen Handelns: Handlungsrationalität und Gesellschaftliche Rationalisierung (first published 1981)].
[33] Edmund Burke, ‘Speech to the Electors of Bristol’ in Phillip B Kurland and Ralph Lerner (eds), The Founders’ Constitution (University of Chicago Press, 1987) vol 1, 391, 392.
[34] Chambers, above n 27, 338.
[35] Rick Henderson, ‘Polling Error’ (1993) 25(1) Reason 11, 11.
[36] See below nn 91, 103 and accompanying text.
[37] Andrew Russell, ‘The Truth about Youth? Media Portrayals of Young People and Politics in Britain’ (2004) 4 Journal of Public Affairs 347, 348.
[38] International Chamber of Commerce and ESOMAR, International Code on Market Research and Social Research (Code, December 2007) <http://www.esomar.org/uploads/public/knowledge-and-standards/
codes-and-guidelines/ICCESOMAR_Code_English_.pdf>; World Association for Public Opinion Research, Code of Ethics (Code, 1 December 2011) <http://wapor.org/wapor-code-of-ethics/> . ESOMAR is an international organisation representing market and social opinion researchers.
[39] ESOMAR and World Association for Public Opinion Research, Guideline on Opinion Polls and Published Surveys (August 2014).
[40] See discussion below in Part III(A) on comparative regulation.
[41] ACE Electoral Knowledge Network, Blackout Period for Release of Opinion Poll Results (26 November 2014) ACE Project <https://aceproject.org/epic-fr/CDMap?question=ME062&questions=all&set_language=en>.
[42] Thomas Petersen, ‘Regulation of Opinion Polls: A Comparative Perspective’ in Christina Holtz-Bacha and Jesper Strömbäck (eds), Opinion Polls and the Media: Reflecting and Shaping Public Opinion (Palgrave Macmillan, 2012) 47, 57–8, updating the estimate in Frits Spangenberg, The Freedom to Publish Opinion Poll Results: Report on a Worldwide Update (Foundation for Information, 2003). See also Robert Chung, ‘Freedom to Publish Opinion Poll Results: A Worldwide Update of 2012’ (Research Report, University of Hong Kong, December 2012) reporting 46 per cent of 83 countries having some form of embargo on election period publication: at 6.
[43] See Chung, above n 42. For a summary of the restrictions in place in a selection of countries in 2003, see Article 19, Comparative Study of Laws and Regulations Restricting the Publication of Election Opinion Polls (Report, January 2003). A pair of industry group surveys at the turn of last century identified 30 democracies with polling embargoes (out of 66 and 78 countries surveyed, respectively): Wolfgang Donsbach and Uwe Hartung, ‘The Legal Status of Public Opinion Research in the World’ in Wolfgang Donsbach and Michael W Traugott (eds), The SAGE Handbook of Public Opinion Research (SAGE Publications, 2008) 431, 433; Donsbach above n 31, 7.
[44] Loi électorale du 18 février 2003 [Electoral Law of 18 February 2003] (Luxembourg) JO, 6 March 2009, 469, art 97.
[45] Donsbach and Hartung, above n 43, 438.
[46] Chung, above n 42, 7.
[47] Loi nº 77-808 du 19 juillet 1977 relative à la publication et à la diffusion de certains sondages d’opinion [Law No 77-808 of 19 July 1977 relating to the Publication and Distribution of Some Opinion Polls] (France) JO, 20 July 1977, 3837, art 11.
[48] Tim Bale, ‘Restricting the Broadcast and Publication of Pre-election and Exit Polls: Some Selected Examples’ (2002) 39 Representation 15, 16, 20.
[49] Electoral Act 1993 (NZ) s 197(1)(d)–(e). Andrew Geddis notes the ritual element to this limit, which lends ‘an extra air of excitement to the media’s election-night coverage’: Andrew Geddis, Electoral Law in New Zealand: Practice and Policy (LexisNexis, 2007) 184. In the UK, see Representation of the People Act 1983 (UK) c 2, s 66A.
[50] See current version of the Canada Elections Act, SC 2000, c 9, s 328. The intervention was Thomson Newspapers Co Ltd v A-G (Canada) [1998] 1 SCR 877 (‘Thomson Newspapers’), discussed below.
[51] Chung, above n 42, 10. Compare also industry best-practice guidelines for information that should accompany published voting intention polling data, in ESOMAR and World Association for Public Opinion Research, above n 39, 7–8.
[52] See, eg, Ian McAllister, ‘Elections without Cues: The 1999 Australian Republic Referendum’ 36 Australian Journal of Political Science 247.
[53] Most obviously in Australian Capital Television v Commonwealth [1992] HCA 45; (1992) 177 CLR 106.
[54] [1998] 1 SCR 877. The case and issues involved were prophesied in Howard Kushner, ‘Election Polls, Freedom of Speech and the Constitution’ (1983) 15 Ottawa Law Review 515.
[55] Thomson Newspapers [1998] 1 SCR 877, 902, 926 (unscrutinisable polling late in the campaign), 888 (averting to deliberative arguments) (Gonthier J).
[56] Canada Elections Act, SC 2000, c 9, ss 323–4.
[57] Cour de cassation [French Court of Cassation], 00-85329, 4 September 2001 reported in (2001) Bull crim n°170, involving the Le Parisien newspaper.
[58] Corte constitucional [Colombian Constitutional Court], Decision No C-488/93, 28 October 1993.
[59] Corte constitucional [Colombian Constitutional Court], Decision No C-089/94, 3 March 1994. This was a reasonable finding but one which undermined the earlier rationale that polls were acts of expression of those surveyed.
[60] Social Weather Stations Inc v Commission on Elections, Supreme Court of the Philippines, GR No 147571, 5 May 2001, reported in (2001) 409 Philippine Reports 571.This source may be found at the following address: <http://elibrary.judiciary.gov.ph/thebookshelf/showdocs/1/52161>
[61] Human Rights Committee, Views: Communication No 968/2001, 84th sess, UN Doc CCPR/C/84/D/
968/2001 (23 August 2005) 4 [2.1], [3.1], 8 [9] (‘Kim Jong-Cheol v Republic of Korea’). Similarly, the Supreme Court of Justice of Paraguay has upheld a 15-day ban: Corte Suprema de Justicia [Paraguan Supreme Court of Justice], Decision No 99/1998, 5 May 1998. The case involved Teledifusora Paraguay, and may be found at the following address: <gestor.pradpi.org/download.php?id_doc=1544>.
[62] Kim Jong-Cheol v Republic of Korea, UN Doc CCPR/C/84/D/968/2001, 7 [8.3].
[63] Ibid 10 (Committee Member Ruth Wedgwood).
[64] Social Weather Stations Inc v Commission on Elections, Supreme Court of the Phillipines, GR No 147571, 5 May 2001, reported in (2001) 409 Philippine Reports 571. This source may be found at the following address: <http://elibrary.judiciary.gov.ph/thebookshelf/showdocs/1/52161>
[65] A factor relied on by the majority of the Supreme Court of Justice of Argentina in Asociación de Teleradiodifusoras Argentina v Gobierno de la Ciudad de Buenos Aires, Corte Suprema de Justicia de la Nación [Argentinian Supreme Court of Justice], Decision No A682.XXXVI, 7 June 2005. This source may be found at the following address: <http://servicios.csjn.gov.ar/confal/ConsultaCompleta
Fallos.do?method=verDocumentos&id=585381>
[66] This reasoning – that prohibiting publication of polls created political inequality – was developed most rigorously in Tribunal Constitucional del Perú [Peruvian Constitutional Court], Decision No EXP nº 02-2001-AI/TC, 4 April 2001, [12]. This source may be found at the following address: <www.justiciaviva.org.pe/jurispu/libertaexp/07.doc>.
[67] Bertoni, above n 29, vol III, 155.
[68] Committee of Ministers, Council of Europe, Recommendation No R (99) 15 of the Committee of Ministers to Member States on Measures Concerning Media Coverage of Election Campaigns (adopted 9 September 1999).
[69] Ibid appendix pt III [2].
[70] Thomson Newspapers [1998] 1 SCR 877, 963 (Cory, McLachlin, Iacobucci, Major and Bastarache JJ).
[71] Canada Elections Act, SC 2000, c 9, ss 326–7.
[72] Jon Elster, ‘Introduction’ in Jon Elster (ed), Deliberative Democracy (Cambridge University Press, 1998) 1, 8; Amy Gutmann and Dennis Thompson, Democracy and Disagreement (Harvard University Press, 1996) 14.
[73] Jürgen Habermas, ‘Reconciliation through the Public Use of Reason: Remarks on John Rawls’s Political Liberalism’ (1995) 92 Journal of Philosophy 109, 124; Elster, above n 72, 8.
[74] John Rawls, Political Liberalism (Columbia University Press, revised ed, 2005) 137.
[75] Gutmann and Thompson, above n 72, 57–9, 110–19; Mark E Warren and Hilary Pearse, ‘Introduction: Democratic Renewal and Deliberative Democracy’ in Mark E Warren and Hilary Pearse (eds), Designing Deliberative Democracy: The British Columbia Citizens’ Assembly (Cambridge University Press, 2008) 1, 6; Chantal Mouffe, ‘Deliberative Democracy or Agonistic Pluralism?’ (1999) 66 Social Research 745, 752.
[76] Zsuzsanna Chappell, Deliberative Democracy: A Critical Introduction (Palgrave Macmillan, 2012) 7.
[77] Ibid 83–7; David M Estlund, Democratic Authority: A Philosophical Framework (Princeton University Press, 2008) 177.
[78] Chappell, above n 76, 8.
[79] James Bohman, ‘Deliberative Democracy and the Epistemic Benefits of Diversity’ (2006) 3 Episteme 175, 175–6; Joshua Cohen, ‘Deliberation and Democratic Legitimacy’ in James Bohman and William Rehg (eds), Deliberative Democracy: Essays on Reason and Politics (Massachusetts Institute of Technology, 1997) 67, 69, 72.
[80] Gutmann and Thompson, above n 72, 43.
[81] James Fishkin, ‘Deliberation by the People Themselves: Entry Points for the Public Voice’ (2013) 12 Election Law Journal 490.
[82] See also Ron Levy and Graeme Orr, The Law of Deliberative Democracy (Routledge, 2016) ch 4.
[83] Petersen, above n 42, 65. Petersen was a president of the World Association for Public Opinion Research.
[84] Kurt Lang and Gladys Engel Lang, ‘The Impact of Polls on Public Opinion’ (1984) 472 Annals of the American Academy of Political and Social Science 129, 129.
[85] See generally George Williams, ‘Push Polling in Australia: Options for Regulation’ (Research Note No 36, Parliamentary Library, Parliament of Australia, 1997). In the US, see Evan Gerstmann and Matthew J Streb, ‘Putting an End to Push-Polling: Why It Should Be Banned and Why the First Amendment Lets Congress Ban It’ (2004) 3 Election Law Journal 37.
[86] Sibylle Hardmeier, ‘The Effects of Published Polls on Citizens’ in Wolfgang Donsbach and Michael W Traugott (eds), The SAGE Handbook of Public Opinion Research (SAGE Publications, 2008) 504, 508–9.
[87] Ibid 506–7.
[88] See, eg, Guy Lachapelle, ‘Polls and the Media in Canadian Elections: Taking the Pulse’ (Research Studies No 16, Royal Commission on Electoral Reform and Party Financing, Canada Communication Group and Dundurn Press, 1991). See also Donsbach, above n 31, 22–31.
[89] Susan Herbst, Numbered Voices: How Opinion Polling Has Shaped American Politics (University of Chicago Press, 1993) 166. See also Colin CJ Feasby, ‘Public Opinion Poll Restrictions, Elections and the Charter’ (1997) 55 University of Toronto Faculty of Law Review 241.
[90] Sally Young, ‘Media Reporting of the Next Federal Election: What Can We Expect?’ (2012) 58 Papers on Parliament 73, 78.
[91] See, eg, Robert Anderson, ‘Reporting Public Opinion Polls: The Media and the 1997 Canadian Election’ (2000) 12 International Journal of Public Opinion Research 285; Claes H de Vreese and Holli A Semetko, ‘Public Perception of Polls and Support for Restrictions on the Publication of Polls: Denmark’s 2000 Euro Referendum’ (2002) 14 International Journal of Public Opinion Research 367.
[92] Schauer and Pildes, above n 1.
[93] Cf Feasby, above n 89, 266.
[94] On the ritual of voting, see Graeme Orr, Ritual and Rhythm in Electoral Systems: A Comparative Legal Account (Ashgate Publishing, 2015).
[95] Graeme Orr, ‘Deliberation and Electoral Law’ (2013) 12 Election Law Journal 421, 423.
[96] Chappell, above n 76, 4.
[97] Robert E Goodin, ‘Democratic Deliberation Within’ (2000) 29 Philosophy & Public Affairs 81, 98–9.
[98] Donsbach and Hartung, above n 43, 436.
[99] Donsbach, above n 31, 9.
[100] Chung, above n 42, 20.
[101] Frank Teer and James D Spence, Political Opinion Polls (Hutchinson, 1973) 10–11.
[102] W Lance Bennett, ‘News Polls: Constructing an Engaged Public’ in Robert Y Shapiro and Lawrence R Jacobs (eds), The Oxford Handbook of American Public Opinion and Media (Oxford University Press, 2011).
[103] A conundrum underlying both historical and ongoing questions for electoral system design, such as multiple votes and vote-weighting, the breadth of the franchise and whether balloting should be compulsory or voluntary.
[104] Regulation of Political Opinion Polling Bill [HL] 2015–16 (UK). The Bill predates the inaccuracies in polling prior to the 2015 Westminster election. Compare Australia, where to date ‘there has been no serious or influential proposal for any official regulation of [polls]’: Stephen Mills and Rodney Tiffen, ‘Opinion Polls and the Media in Australia’ in Christina Holtz-Bacha and Jesper Strömbäck (eds), Opinion Polls and the Media: Reflecting and Shaping Public Opinion (Palgrave Macmillan, 2012) 155, 155.
[105] We develop this case in detail in Levy and Orr, above n 82, chs 4–6.
[106] Indeed, debate per se is often framed with war-like metaphors: positions are ‘attacked’ and ‘defended’. On the role of metaphor in cognition, see George Lakoff, ‘The Contemporary Theory of Metaphor’ in Andrew Ortony (ed), Metaphor and Thought (Cambridge University Press, 2nd ed, 1993) 202.
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/UNSWLawJl/2016/12.html