Abstract
Introduction
Public opinion polls are a source of information and are widely covered by the media. Polls are a vital force in shaping and changing societies. For instance, they shape political strategies, media reportage, and people’s informal conversation about election (Daoust et al., 2020). By stating what the public view is on a given topic, opinion polls gain an element of potential influence among the public. They can provide quick “factual” information that is easily accessible and interpreted. A good poll will seek to give representative public views in a systematic way on social, political, economic, and cultural issues (Ireri & Wolf, 2010; European Society for Opinion and Market Research and World Association for Public Opinion [ESOMAR/WAPOR Report], 2012; Lee, 2016). They not only weigh support for political parties and candidates but also gage views on a wide range of social and economic issues (Rossini et al., 2018; Weaver & Kim, 2002).
However, polls are still viewed with mistrust by a certain segment of the public, who see them as elitist, or as a deceptive tactic to draw attention away from important topics such as governance, politics, and the economy. A systematic, scientific way of obtaining and conveying public opinion on a particular issue of public concern is through opinion polls. In Tanzania, pollsters have never projected the final results of an election with any accuracy and they have been criticized for their sampling techniques, the choice of particular regions in which to conduct their study, the kind of questions asked, framing of questions, withholding results and simply being bias. As a result, polls are frequently criticized by opposition parties, academics, politicians, civil societies, and the general public (Makulilo, 2012).
Central to the study is find out whether the residents of Mwanza city encounter the polls and if they have any impact on them. To achieve this, a list of open-ended questions was prepared to investigate whether Mwanza residents understood opinion polls, trust the polls results, if polls influence their voting behavior and the role of media in creating awareness about opinion polls. The questions were posed to 50 key informants and focus group discussion, which included politicians, civil societies, and community members. To elicit important background information, the participants were first asked about their attitudes, beliefs, or values in relation to opinion polls. When asked to whether they understood the meaning of opinion polls, the majority were aware and referred to them as
Media and Public Opinion Polls
In general, most people receive much of their information through the media (Sonck & Loosveldt, 2010). In this view, polls are regularly published in many countries through various media outlets (Rothmayr & Hardmeier, 2002). This view is also supported by Meffert and Gschwend (2011), who argue that political opinion polls results are largely used in the media. As explained by Dawson (2022, p. 2) that, “polls can serve as information shortcut to indicate the viability of candidates or parties.” Accessibility of information can be a factor in poll influence. This means that people who read polls are better positioned to predict electoral results but only if the polls are accurate (Daoust et al., 2020). In this regard, polls make news headline and are part of heated debate either on radio or television. Following this, publishing poll results in the media can influence voters’ preferences, whether polls were conducted scientifically or not.
Although polls are treated as a source of information sometimes they are of low quality and their limitations rarely discussed (Andersen, 2000; Meffert & Gschwend, 2011). The quality of media reporting is vital because of the increasing number of political opinion polls published during elections campaigns. Media reporting lacks technical details such as sampling, framing, size, margin of error, exact question wording, method of data collection, and poll sponsorship. For this reason, media coverage of political opinion polls and projections requires much attention to ensure balance, fairness, and objectivity for the public to be able to accurately assess and understand them (Ndati et al., 2014). There is need for pollsters to reconsider ways on how to publicize the surveys they conduct so as to heighten public awareness and dispel any air of secrecy or myths surrounding opinion polls (Ansu-Kyeremeh, 1999). This implies that the public is not well informed with regards to polls and media stories based upon them. Notably, in Tanzania, media have only analyzed immediate public reaction to polls. Moreover, they have been accused of not sufficiently analyzing poll results before disseminating them (Kiambi, 2021). This means that there are frequent allegations and denials over whether polls accurately reflect the opinions and expectations of the general public (Mutahi, 2008).
Opinion Polls
Opinion poll is the general public response to decisions made by decision-makers about a particular subject (Price, 2008). According to Pekar et al. (2022) political opinion polls are regarded as a major way to understand a nation’s attitude to different political and social issues. As a result, they create awareness among the public (Ireri & Wolf, 2010). According to ESOMAR/WAPOR Report (2012) must reveal the technical details of their survey. For instance, questions should be framed in a format which allows the respondents to provide correct information (Hillygus, 2011). This means that the principle of impartiality should be maintained. Poll researchers agree that questions must not be too general or insufficiently specific, use simple non-prejudicial language and avoid ambiguity or leading questions (Makulilo, 2012). This shows that the differences in findings between pollsters are often due to such technical details.
It is important for pollsters to reveal the technical details of the study such as nature of the questions, target population, timing, question wording, methodology, and who funds the polls are important for polls to be credible (Moyers, 2010). These measures would create greater transparency in the studies, hence, allowing pollsters to receive a higher level of trust from voters. It is desirable to release results as soon as possible. If pollsters take a long time, their findings can be overtaken by events. As a result, their projections become inaccurate. This implies that even polls that do not accurately reflect the public’s view may still impact upon voter behavior if they are viewed as credible. In this sense, any slight distortion in selecting a sample is harmful to the entire poll as it will not be representative of the population thus posing difficulty in generalization (Makulilo, 2012).
Polls play an important role in determining the eventual outcomes of elections by shaping the behavior of both voter and politicians (Pekar et al., 2022). This implies that choices are influenced by the perceived chances of particular candidate winning. Moreover, “motivating” effects can take place when electorates who had not intended to vote are persuaded to vote. On the other hand, polls may have a “demotivating” effect, when they convince voters that their candidate or party will not win. A “free will” effect may also occur when voters cast their votes to prove the polls wrong (Moyers, 2010). This means that the effects of political polls can be restricted to certain sub-groups of voters (Dawson, 2022).
According to Marsh (1985), the term “bandwagon” refers to a situation in which some individuals adopt a majority position out of a desire to be in the majority. This view is also supported by scholars such as Rothschild and Malhotra (2014) and Wolf (2020), who argue that most people believe that opinion polls can affect voters since the media hype regularly attracts them. Therefore, releasing poll result can encourage people to jump on a bandwagon. However, the “underdog effect” denotes the possibility of the expected loser gaining greater support following the poll’s release (Traugott & Lavrakas, 1996). Thus, the effect of bandwagon and underdog theory. This is because many election campaigns, particularly those at the local level, are characterized by a dearth of information as to the candidates and their position on issues. As a result, voters may be highly susceptible to influence or manipulation by any informational stimuli that they receive. However, some countries such as Canada, Brazil, Poland, and Portugal prohibit the release of poll information within a given period ranging from a day to a month before Election Day (ESOMAR/WAPOR Report, 2012). This is because they want to protect voters against abuses and manipulation in their vote preferences by giving them a quiet period in which they can reflect for a few days before casting their votes. In Tanzania, as many other African nations, the lack of specific legislation to regulate electoral opinion polls (Ndati et al., 2014). This can pose a challenge in ensuring the accuracy and reliability of poll results. However, the Statistics Act in Tanzania is an important piece of legislation that establishes a framework for collection,analysis and dissemination of official statistics.
Opinion Polling in Tanzania
Opinion polling is a modern phenomenon in Tanzania (Makulilo, 2012). The rise of opinion polls can be attributed to media freedom and increase in democratic space. For instance, Article 18 of Tanzania’s Constitution guarantees the right to seek, receive, or impart information or ideas (URT, 1977). As a result of this progress, market research companies such as Twaweza Tanzania, IPOSOS Synovate, Research Education for Democracy in Tanzania (REDET), and the Tanzania Citizens’ Information Bureau (TCIB) started to conduct opinion polls gage the public’s sentiment on political, economic, and social issues in Tanzania. However, a number of shortcomings in the public opinion polls conducted in Tanzania have been identified.
The polls have been accused of leaning toward particular political parties or candidates and they have never accurately predicted election results. For example, in 2010, two polls conducted by Research Education for Democracy in Tanzania (REDET) and SYNOVATE put the ruling party
Reporting of electoral poll results have failed to indicate in which particular regions the polls was conducted. For example, TCIB polls concerning the presidential election, conducted between 27th September and 10th of October 2010 in Tanzania mainland excluded Zanzibar from their sampling. This is problematic since the United Republic Tanzania has 31 regions, five of which are in Zanzibar. The methodologies used by pollsters have also been criticized. If not scientifically conducted, polls can be discredited as partisan (Magalhaes, 2005). For instance, carrying out a study in only one region, where one political party or candidate is perceived to be strong makes poll results partisan due to the target population. This raises the question whether the target population used in studies by pollsters are a true representative of the broader population (Makulilo, 2012).
Another drawback is that some pollsters do not disseminate their findings. For instance, REDET did not publicize their findings for both March and September polls in 2010. This results in a situation where polls can be manipulated or conducted by a biased source, hence distorting the polls results. Makulilo (2012) explains that pollsters should not be afraid to report their findings if they use scientific methods to produce them. This implies that if polls are misrepresented, election outcomes could result from voters unduly influenced to vote for a party they would not otherwise have voted for. In this sense, the results are not a reflection of public views (Lee, 2016).
Methodology
In this study, the methodology used was mainly qualitative. A descriptive research design was used to investigate public perception of political opinion polls and their influence on people’s voting behavior. The study was conducted in Mwanza City and participants included both male and female residents of Mwanza City, aged 18 years and above, who are eligible and registered voters. It used purposive sampling technique to select participants for the key informant interviews and focus group discussion. The key informant interviews and focus group discussions were used as a means of collecting data. There were 10 key informants, comprised of opinion leaders in Mwanza City such as politicians, members of civil societies and community leaders. The key informants were selected largely due to their understanding of Tanzanian political context. During the interview, open-ended questions were used to allow informants to give a full answer and to get the relevant information which is appropriate for the research. The interviews were recorded and lasted for between 30 and 45 min.
Central to focus group discussions was to explore participants perception, experience, and understanding in relation to political opinion polls and election. For the five-focus group discussions, participants were purposively selected to represent the low, middle, and higher-income segments in Mwanza City. Each group consisted of six participants (three males and three females), who were drawn purposively to ensure equal representation from each social class. The focus group discussions were limited to 1 hr so as not to overly inconvenience participants. The collected data were translated from Kiswahili to English and transcribed. Data was grouped into common themes. As such, similar views were grouped into categories that reflected the several themes related to electorate’s attitude, values, and beliefs.
It is important to note that the key informant interviews and focus group discussions were conducted between November 2013 and March 2014, when Jakaya Kikwete was president (2005–2015) and media partially free. However, the media faced significant restrictions under John Pombe Magufulu’s administration. Because of this, media reported stories in generality and avoided statistics in reporting their stories during Magufulu’s era (Media Council of Tanzania, 2019). However, since Magufulu died in March 2021, things have begun to open again. Under the current President Samia Suluhu Hassan, many media and statistical restrictions that had been imposed are being removed, and polling data is starting to appear in the media. This makes consideration regarding circulation and reception of polling data timely.
Results and Discussion
Do the Residents of Mwanza City Discuss Opinion Polls?
The findings obtained show that Mwanza residents are aware of political opinion polls and refer to them as
The Role of Media in Creating Awareness About Opinion Polls?
Media primarily affect the collective opinion, a societal level judgment. Both traditional and new media are used in creating awareness about political opinion polls in Tanzania. In the case of public opinion poll news, newspapers were commonly mentioned as a major source of information when compared to radio and television. This is because newspapers can be read by more than one reader and used for future reference. According to Meffert and Gschwend (2011) reading polls improves one’s ability to predict election results. As such, the general public view polls as instruments used by the media and politicians to sway public opinion (Theuri, 2015). Additionally, to prove their case during the arguments at
Although covered in radio and television, they cannot make reference. People use newspapers to influence other voters by showing them the published poll results. The public attention is drawn to news about public opinion polls by media sensationalism (Wolf, 2020). The results indicate that the level of awareness among voters about political opinion polls has increased as the media use their results as a source of news. This view is supported by Kim et al. (2000), who explain that media regularly include significant polls in their coverage. As a result, people usually utilize polls while talking about vital topics. Another journalist explained, “Most Tanzanians don’t understand the polls. However, because of the newspaper, radio, television and social networks, people now discuss the poll results. This has increased awareness among many people who have access to such facilities.” This suggests that even those who may not read polls or question their reliability may be influenced when they decide for whom they will vote. The results indicate that pollsters should evaluate themselves in order to see how many Tanzanians are aware of the polls. They also suggested that polls should be done frequently, not only when the general election is approaching. This will increase the level of accuracy of the information provided by the respondents taking part in the study, as they will understand what polling is all about and levels of voter mistrust will be reduced.
Media coverage tends to be data rich, but reporting on polls’ findings is conducted poorly by journalists. This is because most do not understand the technical details of the polls; for example, the methodology used to collect data, sample size, sampling procedure, response rate, the undecided category, and statistical data analysis. These elements are needed in a story to give the readers enough tools for evaluating the poll results. As noted by one journalist, “This will increase the level of accuracy of the information provided by the respondents taking part in the study, as they will understand what polling is all about and levels of voter mistrust will be reduced.” However, journalists are untrained in interpreting raw data. Hence, prone to misinterpreting it and jumping to false conclusions. As one journalist lamented, “They just copy and paste the findings as presented by the pollsters. Pollsters should arrange for a seminar to educate regular journalists and editors.” Another journalist added, “Otherwise, people will just see them as numbers and not understand anything.”
Additionally, meeting deadline pressure is a problem for journalists as they lack time to do adequate analysis for their audience to understand the impact of poll results from a layman’s perspective. Polls are exposed through the media, which is a good thing if the media are not biased. The credibility of polls is largely dependent on the mass media that cover them (Kim et al., 2000). For instance, polls are regularly used by politicians and interest groups to frame messages (Daoust et al., 2020). This implies that opinion leaders use mass media more to follow which topics are under discussion than for specific content. Therefore, the citizens are more directly influenced by the opinion leaders who help interpret mass media on their behalf. However, some media houses are owned by politicians and favor the government. They, therefore, may not disclose some pertinent facts in their coverage of the poll’s findings. The media play a major role influencing voters in Tanzania. As a major source of information, the journalists should ensure poll-related stories are covered in a fair, balanced, and objective manner to avoid being partisan. In general, findings obtained demonstrate that communication played a significant role mostly in helping to inform the voters on the intended message of the polls.
Do the Residents of Mwanza City Trust Political Opinion Polls?
The results show that the public is fearful to speak its mind. For instance, those that are employed by the government will not express themselves freely in order to protect their jobs. They are afraid of being victimized yet the public take them to be opinion leaders. One key informant stated, “Some people run away from the pollsters while others are reluctant to give them information. Pollsters should think of how to publicize their studies so as to create public awareness among the citizens. This will build trust between the public and pollsters.” Another key informant lamented, “It is not good if people are not allowed to express themselves freely. If citizens are assured of their security, they will talk more freely and give their views. Following this, pollsters should educate people about the importance of opinion polls and why they should give their views.”
Another issue that impacts upon the credibility of the polls is that citizens often refuse to be interviewed by pollsters as they do not know why they are collecting data on political topics. As one participant said, “Pollsters ‘cook’ the results, they are for the people in the system and it’s a trick CCM use to win elections. We want a pollster which cannot be influenced by the people in the system, a pollster we can trust.”Makulilo (2012) says that pollsters have never accurately predicted the outcome of an election in Tanzania and tend to favor certain political parties or candidates. Another participant stated, “I do not know what ways pollsters use to give information and get results, I have never been interviewed.”
The findings of this article demonstrate that that there is a lot of variance in individuals’ level of trust in opinion polls. The fact that different pollsters tend to produce different results leads to confusion among voters, thus creating suspicion. As a result, voters simply do not know who to trust when faced with often conflicting information. One respondent noted, “This is just a means politician used to confuse voters, pollsters just mislead voters and they [the polls] should not be published. Firstly, who gives them the mandate to collect citizen’s views and how do I know that the people they spoke to share my views. The respondents can just be from the same area or region and support the same candidate or party and you expect me to believe the polls. No, I have my own views.”
Although it was established that Mwanza City voters are aware of political opinion polls, when it comes to whether they trust them, four distinct categories emerged. The first group trust and agree with what the polls say. They tend to be independent minded voters and are not interested in candidates or political parties but only in good policies. They are keen followers of current affairs and how they affect their daily lives. The second group agrees or disagrees with the results depending on whether they support their pre-existing positions. This group is also well informed but tends to argue along political lines and is generally made up of politicians, or party cadres. Therefore, if the polls do not favor their political position they disregard them.
A third group contains people who are not particularly politically engaged and have not made up their minds about which candidate or party to support. Other people’s views affect their own. They believe polls and look to them for direction. One participant stated, “Voters should make up their mind and not just wait for pollsters to publish polls and follow them blindly, don’t wait for the political opinion polls. They will just mix you up and vote for a candidate or party that isn’t your choice. They target voters who have not made up their mind who to vote for.” The fourth group does not believe the polls out of sheer cynicism. This group believes that polls are done deliberately to project one candidate or political party. “Pollsters are just confusing people, they are not sure of themselves. They just mislead voters. Make up your mind and do not wait for them to completely confuse you.” As stated by one participant. The findings indicate that all four groups believed that pollsters should remain impartial when conducting their studies. For instance, carrying out a study in only one region, where one political party or candidate is perceived to be strong makes poll results partisan due to the target population of study. Therefore, the participants all questioned whether the target population used in studies by pollsters are a true representative of the broader population.
The study noted that polls should be done across different regions of the country. “Opinion polls should be conducted in various regions of the country to give a true representation of the voters. If pollsters are partial, voters will not trust their results and they will be viewed to work for a particular party or candidate.” As one key informant observed. Polls results should be carefully examined to affirm that they represent public opinion and do not merely push a single point of view. One key informant cautions that private polls conducted for political campaigns are unsuitable for publication as they are mainly designed to help the candidate or party win the elections. This is because results can be swayed by question wording, the timing of the poll, and the groups interviewed.
Although there is no specific requirement for the timing of releasing poll results, Mwanza residents would prefer poll results to be disseminated immediately so as not to be over-taken by events. Regarding the regulation of public opinion polls, the findings showed polls should not be regulated. This is because they are a platform for citizens to express their opinion freely about political debates and polls have the potential to promote democracy. In this perspective, regulation of polls is seen as a restriction on the right to free speech. However, some participants want legislation that would control the release of political opinion polls in order to allow voters time to analyze the polls’ findings for a period of at least 7 days before the elections. This would protect electorates against abuses and manipulation in their vote preferences by giving them a quiet period in which they can reflect for a few days before casting their votes. Ndati et al. (2014), argue that the authorities should ensure that voters have access to credible, reliable, and representative polls at any time as an election policy.
Opinion Polls and Voters’ Behavior Among Residents of Mwanza City?
From the finding of the study, it is evident that there are several possible ways in which polls can affect voters’ decisions. First, they can entice voters to follow the majority due to the desire to pick the winner. In one focus group it was stated, “Polls add more votes to leading candidate or party. People think my party or candidate is losing, why should I vote for a losing candidate or party? So, let me just vote for the leading party or candidate.” The above situation is applicable to the first group, which trust and agree with what the polls say, and to the third group, which can be influenced by other people’s views because they have not made up their mind. Therefore, a candidate or political party projected to be in the lead gets the extra support from people who would not have otherwise voted for them. However, some participants didn’t agree, and as one stated, “I cannot change my stand. After all, the poll results are not the real thing. The results just give people a direction to follow.” This stance is applicable to the second group which agrees or disagrees with the results depending on or not they support their pre-existing positions. That leaves the fourth group, which does not believe the polls out of sheer cynicism. Often these voters lend their support to the candidate who is behind just to prove the pollsters wrong. This because they believe that polls are deliberately manipulated to benefit one candidate or political party.
Polls may also have an effect on turnout. For example, they can discourage people from taking part in an election if they are made to believe it is already going their way, as they assume that their vote will not matter. Another focus group participant noted, “If my party is already winning the elections, why should I go and vote?” However, opponents of the apparently strong candidate are also likely to be discouraged from going to vote. As one participant noted, “They lose hope, why should I go and waste my vote, when my party or candidate is losing.” One participant lamented, “Most voters will come out and vote because poll results say their candidate is winning. This will entice voters to go out and vote for their preferred candidate or party.” However, one of the participants said, “Equally, more voters will come to vote because they want to save their candidate or party from losing the elections.” Also, in other case, polls can lead to a large voter turn-out as voters turn up in big numbers to try and help the party or candidate win the elections.
The impact of the polls on voters is also influenced by the level of awareness among citizens. The voters in urban areas mostly follow the polls and react to them as in the case of Mwanza residents. As one informant observed, “voters in urban areas are more informed due to social interaction. As a result, they are likely to be more influenced by the polls than the voters in the rural areas.”Makulilo (2013) points out that the majority of voters in Tanzania live in rural areas and most have low civic awareness. The participants also believed that polls could influence voters due to the fact that political actors react to them. There was a general agreement among key informants that major events have a dramatic impact on polls because they influence voters, candidates, or political parties to reinterpret the findings. Politicians use the poll results as a strategy in their campaigns, as the results, have a significant influence on the actual voting as well as turnout. However, participants observed that when poll results favor a candidate or party, they embrace the results but those who don’t do well in the polls disregard them. The participants suggested that candidates or parties should promote their political agenda to the voters and not rely on political opinion polls. Notably, they agreed that publishing polls persuades voters to vote for the candidate or party perceived to be in the lead by the opinion polls.
Conclusion
The article has observed that polling is a new phenomenon in Tanzania, and the level of awareness of polls is high amongst the public. Overall, an individual’s opinion on the accuracy of polls is mostly determined by which group of people the individual connects itself to. Further, independent and undecided voters are more easily swayed by polls than those who already hold strong political views. The results show that a candidate or party reported to be in the lead by the polls gets extra votes because voters are more likely to vote for a candidate they expect to win than one they expect to lose. This confirms Marsh’s bandwagon effect theory and runs against the underdog theory presented by Marsh (1985). Although both bandwagon and underdog theories suggest increased election turnout, the findings indicate that the opposite can occur as voters do not bother to go and cast their votes. It noted that pollsters use media to release their findings. Evidence of support for the Two-Step theory effect, which explains how voters through media can be influenced by their opinion leaders such as local politicians. The study stressed the importance of journalists understanding the technical details of the polls such as sample size and nature of questions asked. This will enable the journalist to do a well-researched analysis of the opinion polls that will allow their audience to better understand the meaning of the information contained in the polls. Therefore, the study recommends that media should ensure that journalists are well trained on reporting poll-related stories as they have become a major source of news and whether there is a need to have legislation which should regulate publishing of political opinion polls in Tanzania.
Footnotes
Declaration of Conflicting Interests
Funding
ORCID iD
