Thomson Newspapers Co. v. Canada (Attorney General)

Thomson Newspapers Co. v. Canada (Attorney General)
Collection
Supreme Court Judgments
Date
1998-05-29
Report
[1998] 1 SCR 877
Case number
25593
Judges
Lamer, Antonio; L'Heureux-Dubé, Claire; Sopinka, John; Gonthier, Charles Doherty; Cory, Peter deCarteret; McLachlin, Beverley; Iacobucci, Frank; Major, John C.; Bastarache, Michel
On appeal from
Ontario
Subjects
Constitutional law
Notes
SCC Case Information: 25593
Decision Content
Thomson Newspapers Co. v. Canada (Attorney General), [1998] 1 S.C.R. 877
Thomson Newspapers Company Limited, doing business as
The Globe and Mail, The Evening Telegram, Winnipeg
Free Press and Times‑Colonist, and Southam Inc. Appellants
v.
The Attorney General of Canada Respondent
and
The Attorney General of British Columbia
and the Canadian Civil Liberties Association Interveners
Indexed as: Thomson Newspapers Co. v. Canada (Attorney General)
File No.: 25593.
1997: October 9; 1998: May 29.
Present: Lamer C.J. and L’Heureux‑Dubé, Sopinka,* Gonthier, Cory, McLachlin, Iacobucci, Major and Bastarache JJ.
on appeal from the court of appeal for ontario
Constitutional law ‑‑ Charter of Rights ‑‑ Freedom of expression ‑‑ Opinion surveys ‑‑ Elections ‑‑ Federal elections legislation prohibiting publication, dissemination or broadcasting of opinion survey results on last weekend of election campaign and on polling day ‑‑ Whether legislation infringes freedom of expression ‑‑ If so, whether infringement justifiable ‑‑ Canadian Charter of Rights and Freedoms, ss. 1 , 2 (b) ‑‑ Canada Elections Act, R.S.C., 1985, c. E-2, s. 322.1.
Constitutional law ‑‑ Charter of Rights ‑‑ Right to vote ‑‑ Access to information during election ‑‑ Restrictions on opinion survey results ‑‑ Canadian Charter of Rights and Freedoms, s. 3 ‑‑ Canada Elections Act, R.S.C., 1985, c. E‑2, s. 322.1.
The appellants brought an application for a declaration that s. 322.1 of the Canada Elections Act violates freedom of expression and the right to vote guaranteed by ss. 2 (b) and 3 of the Canadian Charter of Rights and Freedoms . The impugned section prohibits the broadcasting, publication or dissemination of opinion survey results during the final three days of a federal election campaign. The Ontario Court (General Division) denied the appellants’ application, holding that s. 322.1 did not violate a citizen’s right to vote and that, although the section infringed freedom of expression, it was justified under s. 1 of the Charter . The Court of Appeal affirmed the judgment.
Held (Lamer C.J. and L’Heureux-Dubé and Gonthier JJ. dissenting): The appeal should be allowed.
Per Cory, McLachlin, Iacobucci, Major and Bastarache JJ.: Section 322.1 of the Canada Elections Act, which applies only to “new” poll results, infringes s. 2 (b) of the Charter . The publication of opinion survey results is an activity that conveys meaning and therefore falls within the ambit of s. 2 (b). By prohibiting the broadcasting, publication or dissemination of opinion survey results during the final three days of an election campaign, s. 322.1 restricts freedom of expression.
Section 322.1 is not justified under s. 1 of the Charter . All the steps of a s. 1 analysis must be undertaken with a close attention to context. Characterizing the context of the impugned provision is important in order to determine the type of proof which a court can demand of the legislator to justify its measures under s. 1 . In the course of a contextual approach under s. 1 , the vulnerability of the group which the legislator seeks to protect, that group’s own subjective fears or apprehension of harm, the inability to measure scientifically a particular harm, and the efficaciousness of a remedy, are all factors which the court must take into account in assessing whether a limit has been demonstrably justified according to the civil standard of proof. Another factor to be considered is the nature of the activity which is infringed. The degree of constitutional protection may vary depending on the nature of the expression at issue. Here, the speech infringed is political information. Opinion surveys regarding political candidates or electoral issues are part of the political process and, thus, at the core of expression guaranteed by the Charter . The nature of the expression at issue suggests that a deferential approach is inappropriate in this case.
While the objective of providing a period of rest and reflection for voters prior to going to the polls is not a pressing and substantial objective, the objective of guarding against the possible influence of inaccurate polls late in the election campaign by allowing for a period of criticism and scrutiny immediately prior to election day is of sufficient importance to meet the first step of the s. 1 analysis. The purpose of this particular limitation on expression is to ensure that information which the evidence indicates has an important influence on the choice of at least some voters is presented according to the standards of accuracy which polls are normally expected to attain. To the extent that the votes of some might be distorted as a result of polls being presented in a misleading fashion, such a distortion is clearly a matter which the government may legitimately be concerned to remedy.
The three-day blackout period on the publication of polls will serve, to some degree, the purpose of preventing the use of inaccurate polls by voters by giving critics the opportunity to assess the methodological information made available by the pollster and to question the validity of the poll on that basis. To that extent, the ban is rationally connected to the purpose of the legislation. However, since s. 322.1 does not require the publication of methodological information, the most that could be achieved by the blackout period, if an opinion survey is released without such information, is that the validity of a poll could be undermined by pointing out the failure of the pollster to publish the methodology of the poll.
Section 322.1 does not minimally impair freedom of expression. The section is a very crude instrument in serving the government’s purpose. The social science evidence did not establish that Canadian voters are a vulnerable group relative to pollsters and the media who publish polls. The presumption should be that the Canadian voter is a rational actor who can learn from experience and make independent judgments about the value of particular sources of electoral information. While some voters clearly do consider polls to be of some value in making their electoral decision, no evidence has been adduced that voters have suffered from any misapprehensions regarding the accuracy of any single poll. Voters are constantly exposed to opinion poll results throughout the election and a single inaccurate poll result is likely to be spotted and discounted appropriately. This is not an appropriate case for the government to respond to the paucity of evidence by relying on the “reasoned apprehension of harm” test. First, the claims of widespread or significant harm based on a logical inference derived from surrounding factors are not compelling in the context of factors which, as in this case, refute such logical inferences. Second, the government is not dealing with a vulnerable group which is in danger of manipulation or abuse by the pollsters or the media because of an essential opposition of interests, or because of the nature of the speech itself. Nor is there a shared understanding amongst Canadians that a single inaccurate poll will mislead them to any undue extent.
Where the contextual factors indicate that the government has not established that the harm which it is seeking to prevent is widespread or significant, a deferential approach to the particular means chosen by the legislature to implement the legislative purpose is not warranted. In this case, s. 322.1 is not narrowly tailored to its objective. The ban is overbroad because it prohibits in the final three days of an election campaign the publication and use by voters of all those polls which would meet the usual standards of accuracy. The ban is also underbroad because it may not adequately disabuse voters of an erroneous impression left by a poll which did not disclose its methodology to critics or the public. The obvious alternative was a mandatory disclosure of methodological information without a publication ban. Although such a provision would still leave the door open to inaccurate poll results published immediately prior to the election having some impact, that possibility would be significantly reduced both by virtue of the reader’s initial access to those methodological data, and by the opportunity for rapid response by parties whose interests are prejudiced by the inaccurate poll. The failure to address or explain the reason for not adopting a significantly less intrusive measure which appears as effective as that actually adopted weighs heavily against the justifiability of s. 322.1. Finally, the experience of the international community is inconclusive.
The doubtful benefits of the ban are outweighed by its deleterious effects. The impact of s. 322.1 on freedom of expression is profound. The section imposes a complete ban on political information at a crucial time in the electoral process. The ban interferes with the rights of voters who want access to the most timely polling information available, and with the rights of the media and pollsters who want to provide it. Although it is conceivable that some indeterminate number of voters might be unable to spot an inaccurate poll result and might rely to a significant degree on the error, thus perverting their electoral choice, the government cannot take the most uninformed and naive voter as the standard by which constitutionality is assessed. A measure which decides that information which is desired and can be rationally and properly assessed by the vast majority of the voting electorate should be withheld because of a concern that a very few voters might be so confounded that they would cast their vote for a candidate whom they would not have otherwise preferred cannot be accepted. Given the state of the evidence adduced on this issue, the postulated harm will seldom occur. The benefits of the ban are, therefore, marginal. The deleterious effects, however, are substantial. The ban sends the general message that the media can be constrained by government not to publish factual information. As well, the ban interferes with the media’s reporting function with respect to the election. Further, by denying access to electoral information which some voters may consider useful, the ban interferes not only with their freedom of expression, but also with their perception of the freeness and validity of their vote. In sum, the very serious invasion of the freedom of expression of all Canadians is not outweighed by the speculative and marginal benefits postulated by the government.
In light of the conclusion that s. 322.1 of the Canada Elections Act is an unjustified limit on free expression, it is unnecessary to determine whether the section constitutes an infringement of the right to vote protected by s. 3 of the Charter .
Per Lamer C.J. and L’Heureux-Dubé and Gonthier JJ. (dissenting): Section 322.1 of the Canada Elections Act does not infringe s. 3 of the Charter . A restriction on information would constitute an infringement of the right to vote under s. 3 only if it undermines the guarantee of effective representation. In the instant case, the short blackout period has no such effect. On the contrary, such a period assists effective representation.
While s. 322.1 limits freedom of expression within the meaning of s. 2 (b) of the Charter , it constitutes a reasonable limit demonstrably justified in a free and democratic society under s. 1 of the Charter . The objective of preventing the potentially distorting effect of public opinion survey results that are released late in an election campaign when there is no longer a sufficient opportunity to respond is a sufficiently important objective which meets the first step of the s. 1 analysis. Opinion polls on election issues influence voters’ decisions and it is important that the information the polls convey not be misleading or inaccurate. By providing for timely publication of poll results to allow scrutiny and criticism, s. 322.1 improves information to the public during election campaigns, enhances the electoral process and strikes a balance between the right to vote and freedom of expression. The social science studies which composed much of the evidence show that there exists a long‑standing concern about the publication of opinion survey results during election campaigns in Canada, including the problems associated with the undue influence, late publication and accuracy of polls. In enacting s. 322.1 , Parliament has responded to that concern. The Charter should not become an impediment to social and democratic progress and be made to serve substantial commercial interests in publishing opinion poll results, by defeating a reasonable attempt by Parliament to allay potential distortion of voter choice. Several studies and reports in the last 30 years on the publication of opinion survey results during election campaigns in Canada, as well as bills in the House of Commons, and legislation in other democratic countries, support Parliament’s reasonable finding that the concern at bar was serious.
The second step of the s. 1 analysis ‑‑ the proportionality test‑‑ is also met. First, the rational connection in this case is self‑evident. Opinion polls significantly influence voter choice and electoral campaigns. It follows that the publication of inaccurate, though authoritative, opinion survey results that go uncorrected may well lead to voters making misinformed decisions. Logically, there is a reasoned apprehension that voters will be deprived of the full exercise of their franchise. Its importance is measured by the significant influence of polls on voters and the prevalence of misleading polls. Ensuring that polls that cannot be adequately, publicly and independently evaluated as to their correctness because of insufficient time are not published clearly addresses this problem.
Second, s. 322.1 passes the minimal impairment analysis. Section 322.1 constitutes a genuine mediation between the rights of voters to receive information in a timely fashion, and the right of pollsters and publishers freely to provide the information they want. Not only does the legislation protect the rights of voters but it does so by serving one of the very purposes of freedom of expression ‑‑ informing while allowing for political debate and discussion ‑‑ and by striking a balance between two basic aspects of voters’ right to information ‑‑ the availability of accessible, unrestricted information and the timely availability of factual information that may be misleading so as to allow for scrutiny and criticism. In matching means to ends and asking whether rights are impaired as little as possible, a legislature mediating between the claims of competing groups is forced to strike a balance without the benefit of absolute certainty concerning how that balance is best struck. This Court should not second-guess the wisdom of a legislature in its endeavour to draw the line between competing credible evidence, once it has been established, on the civil standard of proof, that Parliament’s objective was pressing and substantial. At this stage, the question is whether there is a reasonable basis, on the evidence tendered, for concluding that a blackout on all opinion polls during the last weekend of an election campaign and during election day impairs freedom of expression as little as possible given the government’s pressing and substantial objective. Parliament is not bound to find the least intrusive nor the best means. This would be too high a standard for our elected representatives to meet. Here, although one can conceive of alternatives to the impugned measure, there is simply no equally effective alternative to the current short-term blackout for achieving the legislative objective. Based upon the current legislation and the reports and studies available, Parliament reasonably determined that a 72-hour period was necessary to allow meaningful scrutiny of poll results during an election campaign. The 72-hour blackout period is very shortlived and only affects one mode of expression which is not a primary source of information concerning relevant political facts. It mainly constitutes information as to the effect of relevant political information on potential voters. The scope of s. 322.1 prohibits polls of all kinds regardless of their scientific nature or quality because there is no clear cut line between reliable poll results and misleading poll results. Section 322.1 does not apply, however, to the discussion of previously released poll results. The legislation prohibits broadcasting, publication and dissemination and these expressions refer only to the initial release of poll results. Finally, both the motions judge and the Court of Appeal held that inaccurate polls at the end of an election campaign constitute a reasonable concern. Everyone is vulnerable to misinformation which cannot be verified. Our democracy, and its electoral process, finds its strength in the vote of each and every citizen. Each citizen, no matter how politically knowledgeable one may be, has his or her own reasons to vote for a particular candidate and the value of any of these reasons should not be undermined by misinformation. When Parliament identifies one matter of concern, it has no absolute duty to identify and regulate each and every factor. The government does not have to show that this concern is more serious or is causing more harm to the electoral process or to individual voters than any other potentially misleading information. There is no such standard under the Charter .
Third, the salutary effects of s. 322.1 concerning both the right to vote and freedom of expression outweigh its deleterious effects. The salutary effect of s. 322.1 is to promote the right of voters not to be misled in the exercise of their right to vote. Section 322.1, however, deprives some voters, who rely on polls to make their decision, of late campaign opinion poll results. This deleterious effect is quite limited when one considers the delay between conducting the poll and ultimately publishing its results. As to the effects of the measure on freedom of expression, s. 322.1 has a positive impact, promoting debate and truth in political discussion since it gives voters the opportunity to be informed about the existence of misleading factual information. Although s. 322.1 precludes the media from publishing polls on the last weekend of the election campaign and on polling day, this ban causes only minimal impairment to freedom of expression because of its very short duration and because of the lack of satisfactory alternatives available to tailor the measure to the legislative objective.
Cases Cited
By Bastarache J.
Distinguished: R. v. Butler, [1992] 1 S.C.R. 452; Libman v. Quebec (Attorney General), [1997] 3 S.C.R. 569; Irwin Toy Ltd. v. Quebec (Attorney General), [1989] 1 S.C.R. 927; RJR‑MacDonald Inc. v. Canada (Attorney General), [1995] 3 S.C.R. 199; R. v. Keegstra, [1990] 3 S.C.R. 697; referred to: Dagenais v. Canadian Broadcasting Corp., [1994] 3 S.C.R. 835; Reference re Provincial Electoral Boundaries (Sask.), [1991] 2 S.C.R. 158; Sauvé v. Canada (Attorney General), [1993] 2 S.C.R. 438; Haig v. Canada, [1993] 2 S.C.R. 995; Eur. Court H.R., Mathieu‑Mohin and Clerfayt case, judgment of 2 March 1987, Series A No. 113; Bowman v. United Kingdom (1996), 22 E.H.R.R. C.D. 13; R. v. Oakes, [1986] 1 S.C.R. 103; Edmonton Journal v. Alberta (Attorney General), [1989] 2 S.C.R. 1326; Rocket v. Royal College of Dental Surgeons of Ontario, [1990] 2 S.C.R. 232; Ross v. New Brunswick School District No. 15, [1996] 1 S.C.R. 825; Towne Cinema Theatres Ltd. v. The Queen, [1985] 1 S.C.R. 494; Vriend v. Alberta, [1998] 1 S.C.R. 493; Harvey v. New Brunswick (Attorney General), [1996] 2 S.C.R. 876; R. v. Laba, [1994] 3 S.C.R. 965; Butler v. Michigan, 352 U.S. 380 (1957).
By Gonthier J. (dissenting)
Re C.F.R.B. and Attorney‑General for Canada, [1973] 3 O.R. 819; Reference re Provincial Electoral Boundaries (Sask.), [1991] 2 S.C.R. 158; R. v. Keegstra, [1990] 3 S.C.R. 697; R. v. Oakes, [1986] 1 S.C.R. 103; Libman v. Quebec (Attorney General), [1997] 3 S.C.R. 569; RJR‑MacDonald Inc. v. Canada (Attorney General), [1995] 3 S.C.R. 199; Irwin Toy Ltd. v. Quebec (Attorney General), [1989] 1 S.C.R. 927; Ross v. New Brunswick School District No. 15, [1996] 1 S.C.R. 825; R. v. Zundel, [1992] 2 S.C.R. 731; Ford v. Quebec (Attorney General), [1988] 2 S.C.R. 712; RWDSU v. Dolphin Delivery Ltd., [1986] 2 S.C.R. 573; R. v. Butler, [1992] 1 S.C.R. 452; Canada (Attorney General) v. Somerville, [1996] 8 W.W.R. 199; Harvey v. New Brunswick (Attorney General), [1996] 2 S.C.R. 876; R. v. Edwards Books and Art Ltd., [1986] 2 S.C.R. 713; McKinney v. University of Guelph, [1990] 3 S.C.R. 229; Dagenais v. Canadian Broadcasting Corp., [1994] 3 S.C.R. 835.
Statutes and Regulations Cited
Canada Elections Act, R.S.C., 1985, c. E‑2, ss. 213(1) [rep. & sub. 1993, c. 19, s. 106], 255, 256, 322.1 [en. idem, s. 125].
Canadian Charter of Rights and Freedoms, ss. 1 , 2 (b), 3 , 33 .
Criminal Code, R.S.C., 1985, c. C‑46, s. 126(1) .
Constitution Act, 1982, s. 52 .
Convention for the Protection of Human Rights and Fundamental Freedoms, First Protocol, Europ. T.S. No. 9, Art. 3 .
Election Act, R.S.B.C. 1996, c. 106, s. 235.
Ontario Rules of Civil Procedure, R.R.O. 1990, Reg. 194, rule 14.05(3)(g.1).
Authors Cited
Ansolabehere, Stephen, and Shanto Iyengar. “Of Horseshoes and Horse Races: Experimental Studies of the Impact of Poll Results on Electoral Behavior”, Political Communication, vol. 11, No. 4, 1994, 413-430.
Cameron, Jamie. “The Past, Present, and Future of Expressive Freedom Under the Charter ” (1997), 35 Osgoode Hall L.J. 1.
Canada. Committee on Election Expenses. Report of the Committee on Election Expenses. Ottawa: Queen’s Printer, 1966.
Canada. Privy Council. White Paper on Election Law Reform. Ottawa, 1986.
Canada. Royal Commission on Electoral Reform and Party Financing. Reforming Electoral Democracy: Final Report, vol. 1. Ottawa: Minister of Supply and Services Canada, 1991.
Canada. Senate. Debates of the Senate, April 29, 1993, p. 3117.
Canada. Senate. Standing Committee on Legal and Constitutional Affairs. Proceedings, Issue No. 41, May 6, 1993, pp. 41:14‑41:15, 41:17‑41:18.
Feasby, Colin C. J. “Public Opinion Poll Restrictions, Elections, and the Charter ” (1997), 55(2) U.T. Fac. L. Rev. 241.
Herbst, Susan. Numbered Voices: How Opinion Polling Has Shaped American Politics. Chicago: University of Chicago Press, 1993.
Hogg, Peter W. Constitutional Law of Canada, vol. 2, loose‑leaf ed. Scarborough, Ont.: Carswell, 1992 (updated 1997, release 2).
Hoy, Claire. Margin of Error: Pollsters and the Manipulation of Canadian Politics. Toronto: Key Porter Books, 1989.
Johnston, Richard, et al. Letting the People Decide: Dynamics of a Canadian Election. Montreal: McGill‑Queen’s University Press, 1992.
Lachapelle, Guy. Polls and the Media in Canadian Elections: Taking the Pulse. Ottawa: Royal Commission on Electoral Reform and Party Financing, 1991.
McAllister, Ian, and Donley T. Studlar. “Bandwagon, Underdog, or Projection? Opinion Polls and Electoral Choice in Britain, 1979‑1987”, Journal of Politics, vol. 53, No. 3, August 1991, 720-41.
APPEAL from a judgment of the Ontario Court of Appeal (1996), 30 O.R. (3d) 350, 92 O.A.C. 290, 138 D.L.R. (4th) 1, 37 C.R.R. (2d) 225, [1996] O.J. No. 2829 (QL), dismissing the appellants’ appeal from a judgment of the Ontario Court (General Division) (1995), 24 O.R. (3d) 109, [1995] O.J. No. 1375 (QL), dismissing an application for a declaration that s. 322.1 of the Canada Elections Act is unconstitutional. Appeal allowed, Lamer C.J. and L’Heureux-Dubé and Gonthier JJ. dissenting.
W. Ian C. Binnie, Q.C., and Michael J. Bryant, for the appellants.
Roslyn J. Levine, Q.C., and Gail Sinclair, for the respondent.
Joseph J. Arvay, Q.C., for the intervener the Attorney General of British Columbia.
Sydney L. Goldenberg and Stephen L. McCammon, for the intervener the Canadian Civil Liberties Association.
The reasons of Lamer C.J. and L’Heureux-Dubé and Gonthier JJ. were delivered by
1 Gonthier J. (dissenting) -- I have had the benefit of the reasons for judgment of my colleague, Justice Bastarache. I refer to his summary of the judgments below and statement of the issues. While I agree with his statutory interpretation of s. 322.1 of the Canada Elections Act, R.S.C., 1985, c. E-2, with his views as to the scope of s. 3 of the Canadian Charter of Rights and Freedoms , and with his position that there is an infringement of freedom of expression, within the meaning of s. 2 (b) of the Charter , I am not in agreement with his reasons and his disposition of the appeal with respect to the justification of the infringement. In my view, the said infringement is justified under s. 1 of the Charter .
I. Introduction and Historical Background
2 In recent political history, polls have had a substantial impact on the strategies of candidates and the policies of governments. They have become a permanent feature of Canadian politics. It is said that polls tend to reduce the level of political discourse to the lowest common denominator: principles are sacrificed for percentage points (C. C. J. Feasby, “Public Opinion Poll Restrictions, Elections, and the Charter ” (1997), 55(2) U.T. Fac. L. Rev. 241, at p. 244). They tend to preempt the discussion of issues and short-circuit the democratic process:
Conversation, as theorists from Tarde to Habermas have argued, is fundamental to the construction of a democratic public sphere, and polls do not seem to generate interpersonal communication. In a way, polls make many political discussions superfluous, since they give the illusion that the public has already spoken in a definitive manner. [Emphasis in original.]
(S. Herbst, Numbered Voices: How Opinion Polling Has Shaped American Politics (1993), at p. 166.)
To the extent that media coverage of election campaigns focuses more on polling results, it focuses less on the merits of the candidates and their positions and tends to distract voters’ attention from substantive issues pertaining to the good government of the country. The reliance on polls has become so pervasive that some commentators characterize election campaigns as “horse races” (S. Ansolabehere and S. Iyengar, “Of Horseshoes and Horse Races: Experimental Studies of the Impact of Poll Results on Electoral Behavior”, Political Communication, vol. 11, No. 4, 1994, 413-430). The following quotation strikingly illustrates the developing situation:
The 1980s is a decade in which media polling has become an epidemic; in which polling has become not just a political tool, an early-warning guidance system, but an occasional substitute for policy itself; in which democracy itself was undermined by the pervasive influence of polls and pollsters who, for all practical purposes, replaced elected representatives, including cabinet ministers, and traditional political strategists as the major determinants of political action. On the strength of their computer printouts, pollsters came to be viewed as sainted public oracles, supposedly capable not only of probing our innermost thoughts and feelings, but of predicting our future actions as well.
. . .
Polls have become a perplexing reality in the political process; still a legitimate, albeit imperfect, measure of public attitudes, they are afforded an unhealthy level of credence in the affairs of state by political players and the public alike.
(C. Hoy, Margin of Error: Pollsters and the Manipulation of Canadian Politics (1989), at pp. 39-40.)
The problem becomes more acute when some voters consciously use survey results to make decisions, despite the fact that some polls may be inaccurately conducted, misrepresented by the media or misunderstood by the public. The accuracy of opinion poll results may be deceptive and the credibility owed to them exaggerated. The closer the polling day, the less time there is to assess, scrutinize, and possibly correct polls. While the “horse race” turn that election campaigns take is unfortunate, voters have the right to choose the information that they want to rely on in deciding how to vote. If a voter wants to vote strategically, he may rely on poll results to make his decision. However, he will not be well served by an inaccurate or misleading poll. The expression he tries to convey through his vote could be unsubstantiated or misinformed.
3 It should be noted that overall results may not reveal the actual impact of opinion polls on individual voters. For instance, the choice of a voter who wants to support the leading candidate may be cancelled out by the vote of someone who wants to support the trailing candidate, while both were misled by the same poll. Our democracy, and its electoral process, finds its strength in the vote of each and every citizen. Our Charter provides that each citizen has the right to vote. One should not find solace in the thought that two mistakenly cast votes may perchance cancel each other out.
4 To quote the finding of Somers J., the motions judge below, “[t]he effect of polling has been a long-standing concern amongst those involved in the study of elections and with those who actively partake in the elections themselves” ((1995), 24 O.R. (3d) 109, at p. 121). In 1966, a government-appointed Committee on Election Expenses __ the Barbeau Committee __ in the course of its work on the limitation of election expenses, identified opinion polls as an area of concern. In its report, at p. 51, the Committee recommended that the publication of opinion surveys be completely banned throughout election campaigns as it considered “their uncontrolled use for public purposes improper”. The Government recognized that this was an issue of some concern in its White Paper on Election Law Reform in 1986. The White Paper pointed out that up to that time, more than 20 private members’ bills had been introduced in the House of Commons with the purpose of either prohibiting the publication of opinion surveys or controlling the methodology of such surveys published during campaigns. Although none of these bills were passed, they were reflective of the public concern over this issue, since the sponsors of the bills represented different regions of the country, as well as the various political parties which have been represented in the House of Commons over the past 30 years. The Government argued that certain precautions should be taken to protect the public against abuses in this area. In so doing, it recommended that methodological information be included in the publication or broadcast of opinion surveys during an election campaign.
5 By an Order in Council dated November 15, 1989, a Royal Commission on Electoral Reform and Party Financing, chaired by Mr. Pierre Lortie (the “Lortie Commission”), was appointed to inquire into and report on the appropriate principles and process that should govern the election of members of the House of Commons and the financing of political parties and of candidates’ campaigns. It held hearings across Canada and considered the issue of opinion surveys in some depth. During the hearings, at least 90 briefs were submitted regarding the influence of public opinion surveys. Seventy percent of these briefs favoured government regulation of such polls during elections.
6 The Commission retained Professor Guy Lachapelle, Associate Professor in the Department of Political Science at Concordia University, to research and study this issue. His research study, entitled Polls and the Media in Canadian Elections: Taking the Pulse (“Lachapelle Study”), was published in 1991. Somers J. found the Lachapelle Study, “because of its breadth and thoroughness to be the most credible evidence put before the court” (p. 116). Among other things, Lachapelle surveyed the briefs filed with the Commission concerning opinion polls. At para. 8 of his affidavit filed in the instant case (Case on Appeal, at pp. 92 et seq.), Professor Lachapelle stated:
The basic argument of most of these briefs was that polls have undue influence on elections, especially as a potential influence on voters. However, no definitive conclusion about the actual impact of public opinion surveys could be drawn from these briefs. [Emphasis added.]
Professor Lachapelle admitted that the actual impact of public opinion surveys was, and continues to be, a matter of controversy even among researchers. He went on to say:
Therefore, the academic literature is highly divided on the impact of public opinion surveys during and outside election periods. Further, this literature cannot demonstrate that polls do, in fact, have a major impact on the outcome of elections. Even the concept of strategic voting cannot be scientifically proven. However, we cannot reject or ignore the hypothesis that an impact does exist, and has consequences for voting behaviour. [Emphasis added.]
(Lachapelle’s affidavit, at para. 21.)
The foregoing debate appears centered on the effect of polls on the outcome of elections. Of greater import in the case at bar is the effect of polls on individual voters. The Lortie Commission Final Report (Reforming Electoral Democracy (1991), vol. 1) found, at p. 455, that “the publication of opinion polls during election campaigns is controversial”. However, it reached the conclusion that “[n]otwithstanding the frequent assertion of pollsters that their data have minimal influence on voters, recent research provides strong support for the proposition that published opinion polls can significantly influence campaigns and voters” (p. 457). It also found that “[r]ecent Canadian research supports the conclusion that published campaign opinion polls create the conditions for a ‘politics of expectations’ that includes both strategic voting and bandwagon effects” (p. 458). “In short, the argument that published polls do not influence voter choice or affect the conduct of campaigns is simply untenable” (p. 458). At para. 18 of his affidavit, Professor Lachapelle mentioned that research on the 1988 federal election shows that the behaviour of some voters is guided partly, but not solely, by the published and broadcast results of public opinion surveys.
7 In his study, Professor Lachapelle mentioned several potential effects of opinion polls on voters at election time:
· the bandwagon effect (electors rally to support the candidate leading in the polls);
· the underdog effect (electors rally to support the candidate trailing in the polls);
· the demotivating effect (electors abstain from voting out of certainty that their candidate will win);
· the motivating effect (electors vote because the polls alert them to the fact that an election is going on);
· the strategic effect (electors decide how to vote on the basis of the relative popularity of the parties according to the polls); and
· the free-will effect (electors vote to prove the polls wrong).
(Lachapelle Study, supra, at pp. 13-14.)
The bandwagon effect is the best documented result of exposure to poll news. Studies found that, while the bandwagon effect may occur in response to any poll, elections provide the most fertile ground for the growth of poll-driven opinion. It is said that the bandwagon effect was observed in British and American elections, and in Quebec during the 1988 federal election (I. McAllister and D. T. Studlar, “Bandwagon, Underdog, or Projection? Opinion Polls and Electoral Choice in Britain, 1979-1987”, Journal of Politics, vol. 53, No. 3, August 1991, 720-41, at p. 736; Ansolabehere and Iyengar, supra, at p. 427; R. Johnston et al., Letting the People Decide: Dynamics of a Canadian Election (1992), at p. 200).
8 In the course of his work, Lachapelle studied the brief of the Quebec Federation of Professional Journalists to the Lortie Commission. In its brief, the Federation conceded that a 48-hour blackout prohibiting all publication or broadcast of opinion polls before the vote would be an acceptable compromise and would respect every individual’s right to have time to reply (Lachapelle Study, supra, at p. 27).
9 At the end of his study, Lachapelle concluded that “there are significant shortcomings in the media treatment of polls” (Lachapelle Study, at p. 154). He listed several factors which could seriously impact on the accuracy and reliability of opinion poll results:
1. The Order and Wording of Questions Asked: For instance, according to Claire Hoy’s statement in 1989 (Hoy, supra), Angus Reid differs from the other polling organizations in that it asks which of the three leaders is the most popular before asking which of the parties the respondent supports. In addition, Reid asks the question on voting intentions at the end of the questionnaire, whereas Environics, for example, asks it at the beginning (Lachapelle Study, at pp. 90 and 95);
2. The Different Methods of Distribution of the Undecided Voters: In the 1988 federal election, no polling organization made a distinction between respondents who intended to refrain from voting, intended to spoil their ballot, did not know how they were going to vote, or refused to answer, although the first two instances are hardly cases of indecision; they express a clear opinion. The treatment of undecided respondents varies considerably and may actually misrepresent reality (Lachapelle Study, at pp. 95-96);
3. Data-Collection Methods: During the 1988 federal election campaign, the polling organization Gallup Inc. conducted its surveys both by telephone and in face-to-face interviews, separately or together. This poses numerous problems, especially when the results are presented together without comparing telephone results with those obtained through personal interviews (Lachapelle Study, at p. 97). It is difficult for the elector to grasp all the methodological nuances of sampling (Lachapelle Study, at p. 117). The reader requires a certain attentiveness to recognize that Gallup’s methodology varied from one poll to the next, although the press reports contain the information (Lachapelle Study, at p. 116);
4. Margin of Error: The margin of error for a particular question may be higher than the average margin because not all the interviewees respond to all the questions. This is not expressly mentioned in the polls (Lachapelle Study, at p. 104);
5. Timeframe for the Interviews: For a cross-Canada survey, some organizations take up to seven days to interview their respondents, which may reduce the accuracy and reliability of the “snapshot” (Lachapelle Study, at pp. 113-15 and 116);
The last poll of the 1988 campaign, broadcast two days before election day, did not mention the dates of the interviews. As the study put it: “How could even moderately perspicacious citizens be expected to react and exercise their right of rejoinder? Given the amount of time left, this right was rendered ineffective.” (Lachapelle Study, at p. 116.)
10 Predicting regional voting intentions from a cross-Canada survey is also subject to considerable risk, though it became widespread in recent elections. The validity of certain local polls can also be questioned when the interview periods extend over several weeks if not months (Lachapelle Study, at p. 145).
11 In spite of their serious impact on the reliability of polls, these factors, just to name a few, are often overlooked by the public when considering poll results but are carefully taken into account by analysts.
12 On th