Do Public Opinion Polls Taint the Spirit of Democracy?

The modernization of the democratic society in the last century has revolutionized the nature of discourse and interaction between political workers and the voters. The introduction of electronic voting machines and the wide spread use of media during election campaigns by both the electorate and political parties has transformed the nature of voting behavior. The public opinion poll is another mechanism that has dramatically influenced the democratic institution. A public opinion poll is defined as “a type of survey…designed to measure the public's views regarding a particular topic…” ("What Is Public Opinion Polling and Why Is It Important?"). The first known public opinion poll was conducted in 1824 by The Harrisburg Pennsylvania in the run up to the U.S Presidential Elections. Since then, public opinion polls have been widely used to predict the outcome of elections. Although public opinion polls are usually considered to champion the tenets of democracy by informing the voters about policy issues, a close analysis of the principle shortcomings of public opinion polls in the context of certain general elections in western liberal democracies reveals that public opinion polls function in a way that runs counter to the spirit of democracy because they are rife with definitional problems, they do not take into account the different metrics of voters’ preferences and they are wrongly utilized by political parties. Public opinion polls cannot accurately measure public opinion due to their various shortcomings on definitional parameters. The definition of public opinion by people who conduct opinion polls – pollsters – is premised on the assumption that society is comprised of an undifferentiated mass of people who have the same knowledge base and interest in political issues. This assumption is easily debunked by creating a principle distinction among the general population. ““The public…is made up of interest groups and a more detached and disinterested spectator-like body” (Blumer qtd. in Plowman 332). Plowman argues that it is the interested group, rather than the apathetic group, of individuals that asserts dominant control over public opinion (332). “Interest groups, which are backed by different power and prestige, [therefore] act through the available channels in the society, which means that they bring pressure to bear on those capable of making important decisions” (Plowman 332). Donsbach moreover argues that the public is composed of a number of sociological entities, ranging from pressure groups to people who have merely thought about contemporary political issues (20). In his opinion, “these varying conceptions [of public opinion] implicate a number of empirical phenomena – media use etc. – as criterial attributes [that opinion polls ignore]” (20). The direct effect of this misconception of the process of formation of public opinion is that pollsters give equal weight to all opinions, ignoring the aforementioned distinction that implies that every individual’s preferences towards casting the ballot varies dramatically. This shortcoming manifests in opinion polls the form of a non-response bias. A non-response bias is said to influence the results derived from survey research if the response rate of the sampled population is low. A low response rate, skewed dramatically towaeds one side, is said to be one of the contributing factors that resulted in the debacle of Literary Digest’s poll on the 1936 U.S Presidential Elections. Literary Digest distributed 10 million opinion ballots among a sample “…[that] was drawn primarily from automobile registration lists and telephone books” (Squire 126-127). “The magazine’s final count before the election gave Landon 1,293,669 votes (55%) and Roosevelt 972,897 (41%)…The actual results on election day gave the [Democratic incumbent, Roosevelt,]…61% of the vote and his Republic challenger [Landon,] only 37%” (Squire 127). Gallup’s 1937 post- election survey on the poll’s failure to forecast the correct result provides a possible explanation for this inconsistency. The result of Gallup’s survey reported that: “69% of the people who voted for Roosevelt did not return the vote ballot; whereas only 48% of the people who voted for Landon did not return the vote ballot” (American Institute of Public Opinion qtd. in Squire 131). This discrepancy in the preference of people to return the ballot was one of the contributing factors in producing a “wildly erroneous forecast of the vote percentages” (Squire 131). Similar problems regarding the voters registering inconsistent results, in relation to the eventual vote they cast, with opinion poll interviewers continue to exist. Moon delineates two possible theoretical explanations to substantiate the claim that voters might have lied to the interviewers conducting polls, which might explain reasons opinion polls were not able to accurately predict the outcome of the 1992 British General Elections. In 1992, “all of the polls significantly underestimated the size of the Conservative [Party’s] vote, and overestimated the Labour [Party’s vote]” (Moon 106). The Conservative Party gained the vote of the 42.8% of the electorate, even though the various polls predicated it to gain only approximately 38% of the vote; whereas, the Labour Party gained the vote of the 35.2% of the electorate, even though the various polls predicated it to gain only approximately 40% of the vote (Moon 106 Table 21). Nick Moon mentions the hypotheses of ‘spiral the silence’ and ‘cognitive dissonance’ to answer the “profitable line of enquiry…: if people were lying [why] were they lying in the same direction [?]” (Moon 127). “The basis of the [spiral the silence] theory is that people who hold views perceived to be unpopular are often reluctant to express them…” (Moon 128). Moon argues that within the context of the 1992 British General Elections during which the provision of tax credit(s) and better public services, such as health, were central policy issues being raises prior to election day on May 3, voters – who would have preferred the Conservative Party’s policies that would have made them better off individually – might have felt guilty about choosing to ignore the collective benefit to society that Labour Party’s election campaign promised (128). John Major, the Conservative Party’s candidate for the seat of Prime Minister, was aware of “what might have been happening, for in one his…speeches in the final days…he actually told people that there was no reason for them to feel guilty about voting conservative” (Moon 128). Moon also makes use of the theory of cognitive dissonance to further advance his claim. “The theory has it that there is a tendency for individuals to seek consistency among their cognitions (i.e. beliefs)” (Moon 129). Moon argues that if the voters being interviewed were lying to the interviewers, they were in effect lying to themselves because they either genuinely desired the Conservative Party’s financially beneficial tax policies or found it impossible to place the interests of the society above their own respective interests on election day – even when they knew that a vote for the Labour Party was the correct thing (128). Moon’s assertions reduce the to essential claim that voters are prone to changing their opinions even till the penultimate day. This claim is further substantiated by advancing a principle distinction between the notion of an opinion and the intention to vote. “To ask how someone will vote is to ask, not for an opinion, but for an intention” (Plowman 337). Public opinion polls at best measure people’s intention to tvote by asking them such questions as: ‘Which party do you intend to vote for?’ This question measures the intention of the voters by giving equal weight to the vote of the individuals, not to their opinions. Opinions, however, are incapable of being validated by opinion polls. “Many people, [for example] do not really have opinions at all on many issues, although some of these will still answer questions [when asked to be interviewed], perhaps in response to the pressure to have an opinion” (Plowman 339). These myriad of definitional problems, however, do not apply to a special form of an opinion poll: the exit poll. Exits polls are conducted on the election day “in order to enable a media client to predict the result of an election a…few hours before the official result is known” (Moon 134) Exit polls share the distinct advantage of being a valuable source of information for the electorate. Moon argues that exit polls enjoy such crucial advantages over pre-election opinion polls as that of interviewing only those people who have voted – rendering void the notion of speculating the voters’ choices – and that of timing (interviewing people after the elections) – eliminating the possibility of failing to account for any last-minute swing in voters’ decisions (141-142). The multifarious definitional problems with public opinion polls, which do not apply to the special case of exit polls, render them an imperfect tool for accurately measuring public opinion and sentiment prior to election day. Public opinion polls are not a true reflection of the preferences of the voters. The questionnaire of the public opinion polls is designed under the assumption that the voters are cognizant of policy issues that direct political debates. Questions 9 to 15 in a sample opinion survey conducted by the Sunday Times in 1997 asked the interviewee questions ranging from Britain’s domestic policies – for example, which party has the best policy of improving the National Health Service (NHS) – and Britain’s regional policies – for example, What do you feel about the disagreements of the Conservative Party on Europe? (Moon Figure 7). Opinion polls not only assume that the sampled population of the electorate is aware of policy issues but also implicitly imply that the electorate casts their votes solely on the basis of such issues. Although voters do take into account the stances of political parties on policy issues, they also cast their votes after taking other metrics into consideration. Although various polling organizations such as Gallup take into account public perception of political workers by regularly publishing the poll on approval ratings, they do not fully account for the myriad of voting preferences that influence voting behavior. Opinion polls, for example, are said to completely ignore the dimension of personality politics. Social Psychologists have asserted that “individual personalities of voters…[have] become decisive for political choice” (Caprara et al. 2. The effect of personalities of individuals on voting behavior has been empirically verified by comparing the effects of personality traits and basic values. Personality traits of individuals are studied using the Five Factor Model that “provides a…quantifiable description of the main surface tendencies of personality” (Caprara et al. 6) that are classified into emotional stability, agreeableness, energy, conscientiousness and openness to experience; basic values are derived from “universal requirements of human condition: power….universalism etc.” (Caprara et al. 7). The results of the experiment conducted by Caprara et al. concluded that the voters’ voting decisions correlate with their personality traits. The Center-left coalition party in Italy (the country in which the experiment was conducted) in 2000, for example, “advocated social welfare, concern for justice, equality…such [an overarching] policy is most expressive of universalism values (emphasizing understanding, appreciation etc.)” (Caprara et al. 11). The trait of universalism indicated the highest correlation – of + 0.28 – indicating a strong positive correlation with the voting preferences of centre-left voters (Caprara et al. 17 Figure 2). Caprara et al. suggested a possible explanation for this phenomenon by asserting that the “political issues have become increasingly complex and politics units increasingly interdependent…” (2) as political parties now mainly occupy the center position on the political ‘left-right’ spectrum. Italy’s context in which the experiment was conducted lends support to this claim. “The established order of Christian Democrats, Socialists and Communists collapsed in the 1990s after 40 years. Two main coalitions replaced them, center-left and center right” (Caprara et al. 9). “…Right and left in Italy have [now] lost most of their traditional ethos…[as] both coalitions champion the principles of liberal democracy” (Caprara et al. 9). The evolution of political discourse and political parties has subsequently influenced voting patterns that are not accurately captured by the current construction of opinion polls. Public opinion polls are thought to assume the role of an unbiased medium of communication that informs the otherwise uninformed voters. Opinion polls are presumed to encourage the voters to evaluate their personal opinions by providing them adequate information about the prevailing aggregate opinion. It could be argued, however, that rather than merely measuring public opinion and informing the voters about it, public opinions polls shape public opinion by influencing the voters to vote for a certain candidate. A major way public opinion polls are said to influence voters’ choice is through the existence of the process of the Bandwagon Effect. The Bandwagon Effect is defined as “the phenomenon of a popular trend attracting even a greater popularity.” In the context of the publication of opinion polls, the Bandwagon Effect is said to exist when the polls urge the voter to vote for the candidate shown to be leading in the electorate’s opinion. The empirical evidence to support the claim is ambiguous as proponents of the different sides of the issue have produced analyzed empirical data to support their respective claims. George Gallup, founder of Gallup, attempted to debunk as early as 1940 the theory regarding the existence of a Bandwagon Effect. He asserted that if such an effect exists, “a substantial rise in the popularity of the winning candidate could logically be expected [after the publication of opinion polls].” Gallup and Rae analyzed the data from various US states to prove their claim. “In Kentucky, surveys were conducted over a period of four months…on the Barkley-Chandler campaign for the Democratic senatorial nomination” (Gallup and Rae 245). Gallup and Rae tabulate that the results showed that the approval rate for Barkley decreased from 67% to 59% between April 10 and August 5; whereas approval rate for Chandler increased from 33% to 41% between the same time period (245). “According to the bandwagon theory, Senator Barkley should have continued to gain strength from the time of the first survey when he had the support of about two-thirds of the …voters” This trend, however, was not observed. On the contrary, recent analysis of data suggests the existence of the Bandwagon Effect. Skalaban concluded that after controlling for such variables as political interest and party identification, the outcome of the 1980 U.S Presidential Election was influenced by the Bandwagon Effect by increasing the voters’ tendency to vote for Ronal Reagan by an average of 9.2% (147). Additionally, during the 1992 U.S. Presidential Election, an experiment conducted with college students revealed the existence of the same effect. Some students were given the results of student and national polls indicating that Bill Clinton was in the lead; whereas, other students were not exposed to the results of the polls. Several students who had previously intended to vote for George H. W. Bush changed their minds after seeing the poll results (Morwitz and Pluzinski 58-64). Although there is no conclusive evidence on the issue of the existence of the Bandwagon Effect, the analysis of poll data from different time periods provides evidence that undercuts the claim that public opinion polls tend to only inform the electorate. Political parties rely too heavily on public opinion polls by using them as tool to help direct their strategies. Political parties conduct private polls to gauge public opinion regarding their campaigns. Political parties in liberal democracies have been conducting private polling for more than half a century. “Labour was the first British party to conduct private polling, with a survey conducted by Mark Adams in 1946” (Moon 174). Since 1946, political parties in Britain – and across the world – have been spending vast sums of money to conduct private polls. Kavanagh states that “in the 1992 general election the Conservatives spent £ 250,000 and Labour [spent] £ 200,000” (Moon 172). The purpose of private polls is for political parties to be able to constantly remain cognizant of public opinion. Barack Obama and Mitt Romney, for example, used to opinion polls to attune their respective campaign strategies in the lead up to the 2012 U.S Presidential Elections. Given that the “…Pew Research Center, in its opinion poll [published a week prior to election day], said that Obama has edged Romney in the final days by holding a… 3 per cent lead over Romney…” (“Obama, Romney focus on swing states in final campaigning”), both candidates directed all their efforts in the last days to the ten different swing states – including Ohio – in “last minute efforts to break into each other’s vote banks” (“Obama, Romney focus on swing states in final campaigning”). The function of opinion polls in the context of helping direct campaign strategy is beneficial. Political parties, however, rely heavily on opinion polls both before and after the elections. They have come to rely on public opinion polls to provide them an avenue for analyzing public opinion both before and after formulating public policies. The assumption behind this rationale is that the voters possess adequate knowledge to influence public policy via opinion polls. Margolis argues that the respondents of opinion polls are assumed to have perfect knowledge about public policy issues, making them better informed than those people who choose not to respond to opinion polls (64). Furthermore, if people aggregately express an opinion through polls, it does not necessarily imply that there exists a logical link between their opinion and their political behavioral manifestations. Opinion polls, as mentioned earlier, are able to accurately predict only the voters’ intention to vote, which is captured by the “the closeness of the actual to the predicted election outcomes” (Margolis 65) published by, for instance, Gallup. It is more difficult, however, to demonstrate a link between “the expressed opinion on policies concerning…energy, environmental pollution…and subsequent political behavior” (Margolis 65). This lack in congruence between opinion and action is readily manifest in the form of negative populist sentiments against newly instituted policies. In Nigeria, for example, despite the fact the a Gallup poll showed President Goodluck Jonathan to have an approval rating of 81% (“Rheault and Tortora”), Nigerians readily rioted against the government when a higher gasoline price was imposed. “Until…2011, gasoline prices in Nigeria averaged $1.70 a gallon…but cheap oil came at a heavy price: Nigeria was spending $8 billion a year to subsidize gasoline…. The IMF…agreed that cutting subsidies would be a vital first step towards fixing Nigeria’s finances.” (Agrawal). “It took [the president, however, only] sixteen days to cave. The government dropped prices way back down, to $2.27 a gallon from $3.50]” (Agrawal). This example explicitly portrays the degree of disparity between the expression opinion and subsequent political behavior, supporting that claim that political parties’ over reliance on public opinion polls results in their inability to institute effective public policies. Modern political thought emphasizes making the institution of democracy more participatory by encouraging more people to go the polling stations on election day to cast their votes. Public opinions are assumed to in part achieve objective as they seemingly provide information to the voters’ about the aggregate public opinion. The public and political parties’ manifest over reliance on public opinion polls, however, taints the spirit of democracy because the aforementioned democratic entities have come to over emphasize the ability of public opinion polls to predict the nature of political discourse prior to the elections with impeccable veracity. The effective result of the increased dependence on public opinion polls is that public opinion is delineated in a grossly erroneous manner that misrepresents the prevalent political mood and environment during the election campaigns.