The inclusion of methodological information in poll-based news: How do Spanish newspapers conform to professional recommendations and legal requirements?

Abstract

Opinion polls are widely reported in the media, and special attention is often paid to pre-election polls in order to protect voters’ interests, as well as to preserve the research industry’s credibility. In accordance with these two purposes, professional associations publish guidelines for journalists on how to report the underlying methodological details of opinion polls/surveys. In Spain, such details are even legally required for poll news reporting. But to what extent news outlets follow these professional recommendations and legal requirements? Based on a content analysis of poll news reporting in five Spanish newspapers before the 2012 Catalonian Parliament election, this article attempts to answer that question. The results show that basic methodological information was present in almost two-thirds of the sampled news stories. The newspapers tend to do better when they commission the reported polls. However, some vital information was still missing or mistaken, highlighting the need for collaboration between pollsters and journalists.

Keywords

Election coverage Esomar poll news reporting poll reporting standards Wapor Spanish media statistics in news

Introduction

During a pre-election period, the media publish many poll results estimating votes or on issues related to the campaigns (Medvic and Dulio, 2008; Strömbäck, 2009, 2012). There has even been an increase over time in media coverage of opinion polls (Brettschneider, 2008; Strömbäck, 2009; Traugott, 2005). This attention is due to the relevance that polls have for the general public and specific groups such as politicians.

The public is interested in opinion polls as opinion issuers and as voters. Citizens can express their views through surveys, so their voices can be heard (Moy and Rinke, 2012). On the other hand, polls can ‘help them make a rational voting choice’ (Esomar and Wapor, 2014b: 11). So, publishing polling results in the media may influence voters.¹ Politicians also use opinion polls in campaigns, criticizing and using them at the same time, for their own purposes (Brettschneider, 1997). As the media report how candidates and parties are doing in opinion polls, they play an important role in the relationship between politicians and public opinion research (Holtz-Bacha, 2012). For the media, polls help to provide information, pose critical questions to political actors or analyse the political horse race, for example (Strömbäck, 2009). They fulfil many functions for journalists: as information sources, as attention-catchers or as a source of journalistic power (Wichmann and Brettschneider, 2009). Media outlets can enjoy exclusive news when they sponsor and cover their own opinion polls (Frankovic, 2012; Strömbäck, 2012).

The important and relevant role of polls in the news encourages polling organizations to publish guidelines to help journalists to report polls (e.g. BBC, 2014; Commonwealth Broadcasting Association, 2004). Professional associations representing pollsters also try to train journalists on how to report and use survey results. The European Society for Opinion and Market Research (Esomar),² the World Association for Public Opinion Research (Wapor), the US National Council of Public Polls (NCPP) (n.d.) and the British Polling Council (n.d.), among others, have written specific codes and guidelines for journalists (Esomar and Wapor, 2014b; Gawiser and Witt (n.d.); Kellner, 2014).

Among the common recommendations of these bodies for journalists is the inclusion of methodological information in their poll reporting. Sometimes, as in the case of Spain, such inclusion is even required by law. In this article, we will first explain the relevance of this information and review research into such reporting of methodological information in several countries. It will then present a content analysis of news stories based on election poll results published before the 2012 Catalonian Parliament election in five Spanish newspapers. We analyse the extent to which, and how, these newspapers include methodological information about surveys in the coverage and whether this is compliant with requirements by Esomar and Wapor (2009: 17)³ and the Spanish Organic Law 5/1985 on the General Electoral System. Based on the results, we make recommendations for the reporting of polls in Spain and beyond.

The relevance of reporting methodological information about polls

Pollsters have a special responsibility to the media, politicians and voters, which is to deliver ‘to the public the information required to ensure transparent, unbiased reporting of the results supported by comprehensive documentation’ (Esomar and Wapor, 2014b: 4). Polls must use appropriate samples and methods, as their quality is conditioned by the application of a suitable methodology (Trujillo, 2002). Errors in the sample size, the geographical coverage of the sample or the questions asked can jeopardize the results’ validity (Maarek, 2009). In the case of pre-election polls, the accuracy of results is judged by their proximity to real election outcomes. When a variation occurs, surveys are often criticized by the media, politicians and voters, and pollsters’ reputation is damaged (Esomar and Wapor, 2014b: 5). However, pre-election polls are just a ‘reflection of opinion at the time the poll was conducted’ (Esomar and Wapor, 2014b: 11). New events and information may change voting preferences before the Election Day. So, differences between poll results and election outcomes can be explained by the time of fieldwork, in addition to other methodological aspects, such as the margin of error, which is related to the sample size.

The methodological context of a poll is therefore necessary for users to judge its quality and the importance of its findings as well as to interpret them (Holtz-Bacha, 2012: 107; Mateos and Penadés, 2013; Traugott, 2008: 239). The media, politicians and voters should receive enough information to assess the reliability and validity of the reported results and/or understand the reasons for any differences from the real election results. According to Smith and Verral (1985), journalists probably omit methodological details in an effort to simplify the report, and they believe that most of the audience is not likely to understand this information. Even so, the media should still provide methodological information and, if it believes the audience will not understand it, should note at least whether the survey conforms to standards (Sonck and Loosveldt, 2008: 499; Strömbäck, 2009: 65). As such, the inclusion of methodological information is essential to differentiate between professional and unprofessional polls (Esomar and Wapor, 2014a: 1). The size of the sample, the places where interviewers were found or the researcher in charge of the study are crucial issues to judge the quality of a poll. The dates of data collection or whether the survey was done by phone or in person can affect results, so journalists should provide all this information.

Although disclosure does not guarantee quality, some polling associations have already required transparency (e.g. British Polling Council, n.d.; NCPP, n.d.). These professional bodies encourage survey firms to disclose all relevant methodological information – such as the sampling method, the size of the sample, the survey mode, the dates of the fieldwork and the sponsorship of the survey – so that clients and the public can assess the polls at stake for themselves. In Spain, the inclusion of such information is legally required, as established under the Law on General Electoral System (Organic Law 5/1985; Organic Law 2/2011), and is endorsed by the National Association of Researchers, Asociación Española de Estudios de Mercado, Marketing y Opinión (AEDEMO),⁴ whose members observe Esomar’s codes and guidelines. Disclosure is considered so essential because poor pollsters would probably be more reluctant to publish all of this information: ‘Reputable polling organizations do not have anything to hide’ (Welch, 2002: 112). Such information in news coverage can also affect the perception of the accuracy and quality of a poll (De Vreese and Semetko, 2002: 373).

Methodological information in poll news reporting around the world

The inclusion of methodological information in news coverage of polls has been analysed by authors in a number of countries. Brettschneider (1997) content-analysed 443 pre-election poll articles published in the four most relevant German newspapers and found that even those that focused on poll results contained little information about the employed methods. Brettschneider suggested that a pre-requisite for journalists’ improvement in dealing methodological questions was ‘the efforts of pollsters to make their work more transparent and comprehensible’ (p. 263). In a study for the federal elections from 1980 to 2002, again in the four most relevant German newspapers, Brettschneider (2008) discovered that the most frequently mentioned information in reports were polling organizations, commissioning institutions and fieldwork times. In another study of the same newspapers, plus a tabloid and two weekly magazines, Holtz-Bacha (2012) found that the name of the polling institute was the most frequently mentioned information but even this was not found in 30 per cent of the articles.

Studying a sample of Swiss newspapers that had the potential to conduct or to buy polls, Hardmeier (1999) observed a tendency to report only some simple, straight information, with complicated details, such as the methods of data collection, the sample selection procedures, the sampling errors or the response rates ‘virtually always ignored’ (p. 262). Hardmeier also observed that the presentation format of methodological information can influence its quality: for example, when a small box is used to present such information, the quality of technical details was better (p. 263). In Singapore, Chang (1999) examined the form and content of public opinion reporting in the country’s two most important newspapers. Coding the inclusion of information recommended by the American Association for the Public Opinion Research (AAPOR), Chang (1999) concluded that most of the stories did not provide such vital information as response rates, sampling methods, sampling errors or interview methods.

These results are similar to those obtained in Canada by Andersen (2000), who found an ‘alarming absence of basic technical information’ (p. 292) in television and newspaper reporting of pre-election polls during the 1997 Canadian election. Again, even the sample size was seldom mentioned. Andersen (2000) also found that newspapers did better at reporting methodological information than television,⁵ a finding similar to that of De Vreese and Semetko (2002) in Denmark. In a study of articles from 18 Canadian newspapers reporting numerous poll results concerning a Canada election, Ferguson and De Clercy (2005) found that the organizations that sponsored and conducted the survey and the dates when the survey was conducted were the three most reported type of content included.

Minimal disclosure of essential information for people to determine a poll’s reliability and validity was also discovered in the United States, where according to Welch (2002), the sponsor of the poll and the sample size were the most common information given by national and more local newspapers (pp. 107–110). Additionally, US newspapers tend to provide more information about the polls when they are the sponsor than when it was an outside poll (p. 109).⁶ In Belgium, Sonck and Loosveldt (2008) observed that 45 per cent of 1359 poll-based Flemish newspaper articles were about election polls, with numbers growing in pre-electoral months (p. 495). However, the growth in poll news did not go ‘hand in hand with a similar increase in the quantity of methodological poll information disclosed’ (p. 490). According to the results of the content analysis, the elements most frequently disclosed were pollsters and sample sizes (p. 496). In Sweden, Strömbäck (2009) studied articles about political actors and institutions from the main public and commercial television news channels, broadsheets and tabloids. When a news story presented results from at least one poll, methodological information recommended by Esomar and Wapor was coded. According to his results, ‘neither newspaper nor television news did a good job at providing the kind of methodological information that should be presented when a poll is published’ (p. 64), and, again, television did worse than newspapers. In Brazil, newspapers published a growing number of articles with poll results with at least some information on the method, although the information was rather brief (Biroli et al., 2012: 148). Similar poor information was reported in South African media (Mattes, 2012) and Taiwan newspapers and television (Willnat et al., 2012).

In Spain, until recent years, the media are no different when it comes to including methodological information in poll reporting. In a study in the 1980s, Monzón and Dader (1990) found that the Spanish press seldom included question wording, the number of non-respondents, or the address of the author of the survey institute, although all of these were required by the General Electoral System Law (p. 476). Similar problems were found during the Basque Parliament Election in 1998, according to Vara (1999). However, when a poll was sponsored by public institutions, the provided information was more complete, including dates of fieldwork, sample size and expected abstention (Vara, 1999: 146). Díaz de Rada (2002) examined methodological details published in six newspapers for six polls sponsored by themselves, all related to the Spanish General Election in 2000. In all cases, they included the sample size, but the sampling method was not correctly explained, and there were mistakes in the reporting of the margin of error (pp. 240–242). Most recently, Mateos and Penadés (2013) analysed reporting about electoral polls sponsored by the media (but conducted by private organizations) and those from public institutions. In all cases, although key methodological information was included, it was not complete. All these studies about poll reporting in Spanish media suggest that information is still scarcely given about methodologies. However, they consider the survey as the unit of analysis, instead of news stories, as seen in studies from other countries. In our research, we will consider the issue at story level, asking the following questions:

RQ1: How frequently, and in what formats, do Spanish poll-based news stories include statements about methodological information?

RQ2: Do such statements include all the standard information required by Spanish law and by professional standards required by Esomar and Wapor (2009)?⁷

RQ3: When media outlets sponsor the polls, is there more and better methodological information included in their reporting?

Methodology

In this research, we applied a quantitative content analysis of news stories published in five Spanish dailies in the weeks before the 2012 election to Parliament of Catalonia.⁸ Pre-election reporting is often more intense (Mateos and Penadés, 2013; Sonck and Loosveldt, 2008), and the review of newspapers began with publications dated 1 October 2012, the official date of the call for election, and finished with publications on 25 November 2012, the day of the election. As in the research of Strömbäck (2009), news stories needed to include data from a survey, a poll or a barometer about an election issue. The five general daily newspapers – El País, El Mundo, La Vanguardia, ABC and El Periódico – were selected according to their readership in Spain and more specifically in Catalonia (Table 1). These also reflect different political lines of news outlets in Spain.⁹

Table 1.

Audience and ranking positions of the five selected newspapers.

Newspapers	Spain – readership ranking		Catalonia – readership ranking
Newspapers	Audience (thousands)/position		Audience (thousands)/position
El País	1928.8	1st	169.1	3rd
El Mundo	1180.8	2nd	50.6	8th
La Vanguardia	811.4	3rd	78.8	1st
ABC	647.8	4th	10	12th
El Periódico	612.9	6th	553.4	2nd

Source: ‘Estudio General de Medios’ (EGM), AIMC (www.aimc.es). February–November 2012. The newspaper with fifth highest readership in Spain, ‘La Voz de Galicia’, was not considered as it is relevant in the Spanish region of Galicia but not in Catalonia.

The sampling process started with reading the print versions and searching for relevant news through the search engine ‘My News Online’ (http://mynews.es/) which provides access to more than 80 million articles from 1057 sources. All news stories with data from surveys, polls and barometers about issues directly related to the election to the Parliament of Catalonia were included in the initial sample. We then excluded small news stories (less than 1/8 of the page) where their topic was related with the content of a bigger article on the same page. The inclusion of these small news stories would have artificially made the number of items in some newspapers bigger than others that did not present news using this structure. Finally, letters from readers or news about web polls conducted by the newspapers themselves on their websites were excluded. The final sample has 142 news stories.

All sample units were first registered with information relevant to the identity of the story such as the name of the newspaper, the page where it is published and the headline. Then, we detected the inclusion of a statement of methodological information, classifying these statements according to three formats: a paragraph in the article, a part of a source text blurb at the foot of tables or graphs or a lift-out box (usually in a corner of the page). If there was any kind of statement with methodological information, their content was coded according to the information required by the Spanish electoral law and the points listed in the guide of Esomar and Wapor. A total of 15 elements were coded: sample size, name of commissioning organization (only explicitly required by the law), name of the research organization (pollster), dates of the fieldwork, universe (only explicitly required in Esomar and Wapor’s guide), survey mode (face-to-face, telephone interview, Internet panel, etc.), margin of error (only required by the law), sampling method, use of weighting, how it is applied (only explicitly required in Esomar and Wapor’s guide), response rate, geographical coverage of the sample (only explicitly required in Esomar and Wapor’s guide), full text of main questions on the statement, number of cases with no answer and addresses of organizations (only explicitly required by the law). As question wording and no answers could be included when presenting the data, these two elements were also coded (see table 3). In 8 of the 142 news stories in the sample, there was comparison of results from two surveys. In those cases, we considered the presence of any of the required information for the two polls, as in the research of Welch (2002: 106).

Results

In relation to RQ1, 48 of the 142 sampled units (33.8%) included a statement of methodological information within the story (Table 2). In another 45 cases, this information was included within a related news story on the same or another page of the same newspaper edition. So 65.5 per cent of the poll-based news stories during the pre-election period in question (93 out of 142 stories) included methodological information in the same version as the newspaper.

Table 2.

Frequency of news stories with methodological information (n = 142).

Methodological information	Number of stories	Percentage
Within the news story	48	33.8
In another news story on the same page	5	3.5
In another news story in the same copy (but not same page)	40	28.2
None	49	34.5
Total	142	100

The format of the methodological statement was mainly text inside the story. Of the 48 news stories that included a methodological statement (Table 2), 23 (or 48%) did so as part of the story (i.e. a paragraph in the news item). In 16 stories (33%), the format was a box at the corner of the news story, and in 9 cases (19%), methodological information was included as part of the source text blurb in the tables or graphs that accompanied the story.

So, the answer to RQ1 is that most news stories based on poll data about a relevant issue during Catalonian Parliament election were reported with some methodological information, either as part of the story or in a related one in the same copy of the newspaper. Such frequent inclusion of methodological information shows that journalists have seen this information as relevant as other aspects of the content (Almazán and Villarejo, 1998: 106). However, there is space for improvement as such information was missing in about one-third of the sample of news stories.

In regard to RQ2, it should be first noted that the surveys were properly identified – that is, with the names of both the research and the commissioning organizations so that readers know which other sources to consult when they need to obtain further methodological information or to contrast the analysis – in 79 per cent of the 142 news stories in this sample. Nineteen of the stories (13%) only included the names of the research organizations or the clients. Presenting poll results with only the name of the sponsor or of the researcher is misleading. One researcher can work for multiple clients, or one organization can pay for multiple surveys to different pollsters. Thus, a name alone is not enough for readers to identify a survey. Even worse, 11 articles indicated ‘surveys’ and ‘polls’, without any indication of the institutions conducting or sponsoring the polls. This is similar to what has been found in the United States (Frankovic, 2005).

Looking specifically at the 48 in-story methodological statements, we found that sample sizes, names of commissioning organizations, names of research organizations and the dates of the fieldwork were the most frequently mentioned (Table 3). This is quite in line with previous research into technical information about survey data in the Spanish media (Díaz de Rada, 2002; Monzón and Dader, 1990; Vara, 1999).

Table 3.

Information included in the 48 in-story methodological statements.

Types of legally required or professionally recommended methodological information	Number of stories (percentage based on n = 48)
Sample size	41 (85%)
Name of the commissioning organization (only explicitly required by the law)	41 (85%)
Name of the research organization (pollster)	40 (83%)
Dates of the fieldwork	37 (77%)
Universe (only required in Esomar and Wapor’s guide)	27 (56%)
Survey administration mode (face-to-face, telephone interview, Internet panel, etc.)	25 (52%)
Margin of error (only required by the law)	22 (46%)
Sampling method	18 (38%)
Use of weightings (only required in Esomar and Wapor’s guide)	14 (29%)
How weightings applied (only required in Esomar and Wapor’s guide)	13 (27%)
Response rate	5 (10%)
Geographical coverage of the sample (only required in Esomar and Wapor’s guide)	3 (6%)
Question wording in methodological statements	2 (4%)
Question wording when presenting the data	29 (60%)
Number of ‘No Answer/Don’t Know’ responses in methodological statements	0
Number of ‘No Answer/Don’t Know’ responses when presenting the data	48 (100%)
Addresses of polling and commissioning organizations (only required by the law)	0

On the names of those responsible for the surveys (sponsors and pollsters), the polls were properly identified only in 34 of the 48 statements (71%). In 13 cases, only one name was given. And in one case, the story included the sample size, the method of interviewing and the fieldwork dates, but quite unexpectedly, there was no name to identify the poll.

The population (universe) of the study was reported in 27 of the 48 statements (56%). This is probably due to the obvious need to indicate that only Catalonian voters were interviewed: as three of the five selected dailies were national newspapers, the reference to Catalonia was necessary. However, the geographical distribution of the sample inside Catalonia was rarely included: only three statements included information regarding the geographical region-by-region composition of the sample.

The methods used to collect the data were reported in more than 50 per cent of the 48 methodological statements. This information is easy to digest, so journalists included it frequently in the text of the article. The margin of error was reported in 22 statements (46%). This information should go along with the sampling method, but the latter was reported only in 18 cases.

Esomar and Wapor (2009) require news outlets to indicate whether the results were adjusted by weighting procedures. These weights can modify raw data substantially so they need to be reported. In our study, 14 of the 48 methodological statements had this information, and all but one of these explained how the weights were applied.

The response rate was included only in five cases. These cases were related to a poll conducted among a small universe of Catalonian companies about an electoral issue. It was not included for voters. Interestingly, the number of ‘Don’t Know/No Answer’ responses was included with the data in all the 48 in-story methodological statements. The question wording was included in the reporting of 29 stories (60%). Finally, the addresses required by law were never present. Esomar and Wapor (2009: 19) recommend the inclusion of other information, such as the webpage. Web addresses would be more relevant than the postal addresses required by Spanish law, but these were never included in these news stories.

So, in relation to the second research question, it can be concluded that much of the methodological information required by Spanish law and professional standards is still missing. Of the 14 details required by the Spanish law and by Esomar and Wapor, a maximum of 11 were included, and this was only in one statement. Twenty-seven (56%) only included five or fewer of the required 14 methodological aspects.

Turning to RQ3, we took a close look at the 14 polls sponsored by the five newspapers selected in this research (Table 4). According to previous studies, the media do better in providing methodological information about their surveys (Ferguson and De Clercy, 2005; Miller and Hurd, 1982; Welch, 2002). This also seems to be the case in our study.

Table 4.

Polls about an election issue sponsored by the selected Spanish newspapers prior to the 2012 Catalonian Parliament elections.

Newspaper	Pollster	Number of sponsored polls	Number of news stories on sponsored polls	Total number of poll-based news stories	Percentage of stories about sponsored polls over the total
La Vanguardia	Feedback	5	35	51	69%
El Periódico	GESOP	3	18	39	46%
El País	Metroscopia	3	9	21	43%
El Mundo	Sigma Dos	2	6	19	32%
ABC	DYM	1	3	12	25%
Total		14	71	142	50%

First, there was a heavy focus on these polls: half of the 142 sampled news stories in the five newspapers were related to them (Table 4). In the case of La Vanguardia, the most read newspaper in Catalonia, 69 per cent of the poll-based news stories were related to their own polls. So, the value of these surveys for the content of the newspapers is clear.

Second, we observed that each newspaper worked with only one pollster. As such, the way those polls were reported followed a constant format across different stories in each newspaper. The format of the methodological statement, for example, was a box at the corner of the page for La Vanguardia, El Periódico and ABC, but a source text under the tables and graphs in the case of El País and El Mundo.

Third, in 61 of the 71 news stories (86%), some methodological information was included. This suggests that the media seem to do better in reporting their own surveys because in the entire sample, such information was missing in more than one-third of stories (Table 2). Besides, all the 61 news stories included the universe, the geographic scope, the sample size, the date of fieldwork and the weighting (if any). So, they provided information on at least five technical elements, in addition to the name of the pollster. The quality of the content was clearly better than it was in their reporting of other surveys.

However, a closer look to the information given for these surveys showed some mistakes related to the margin of error. In some of the stories on sponsored surveys, the margin of error was given for non-random samples, which is not correct as this error can only be applied to random sampling (Hair et al., 2003). Also, supplementary information for the margin of error such as the confidence level was missing in a good number of statements. These mistakes are highlighted here because they are often within the responsibility of the pollster during their preparation of such information to their client, the media in this case. In other words, these errors proceed from the researcher, not the journalist, and should be dealt with by researchers. For example, they can approximate a margin of error when non-random sampling is used by providing an estimation. This practice was followed by one of the pollsters of our research (Metroscopia for El País).

To summarize, the answer to the last research question is that when a media outlet is the client, more methodological information is included, with better quality. It is clear from the data that the Spanish newspapers pay more attention to their sponsored than other surveys. However, there is room for improvement on both sides of the media and the pollsters as some important technical information was still missing or reported with mistakes.

Conclusion

Polls are a source of news, and the number of articles reporting their data increases in pre-electoral times (Sonck and Loosveldt, 2008). This content analysis, conducted with 142 Spanish news stories that used data from at least one poll about an issue related to the 2012 Catalonian parliamentary election, confirms the importance of pre-election polls for the media. The five selected Spanish newspapers showed a strong willingness to pay for election-related polls, sponsoring a total of 14 surveys that generated 50 per cent of the election poll-based news stories during the analysed period.

When poll results are reported in the news, methodological information is needed for the public and other stakeholders to assess their reliability and validity. This study shows that there has been an improvement in the inclusion of such information in the Spanish media. Almost two-thirds of the poll-based news stories in our sample included methodological information within the news story or in a related piece in the same copy of the newspaper. Sample sizes and fieldwork dates are generally reported. However, some vital information is still missing in a substantial number of stories. For instance, the names of sponsors and pollsters were not mentioned in one-fifth of the sampled news stories. Without this information, the audience cannot identify the poll at stake and, if needed, look for more details to judge its quality from other sources. Esomar and Wapor (2014b) recommend providing a link to web address of the pollster or to full details of the poll at stake on the media’s own website. The link can be included in any news item, written or spoken. In Spain, some institutions have already begun to include complete information and even raw data on their websites (Mateos and Penadés, 2013). In this study, a private research firm (GESOP) did so for their surveys commissioned by the newspaper El Periódico. This practice is highly recommended for any country. As said above, disclosure does not guarantee quality, but the inclusion of all the information permits a complete assessment of the validity of the research and its results.

Journalists are not the only ones that need to improve their reporting. Polling firms are responsible for the information published by their clients (Esomar, 2007), and some of the mistakes discovered in the case study can be blamed on pollsters, not journalists. Thus, researchers should adhere to the professional guidelines to improve their reports to clients. If they do not provide the proper information to the media, the reporting will always be deficient. For example, pollsters could indicate the estimated value of the margin of error in certain conditions, as Metroscopia did for El País. Professional standards committees representing polling associations have an important role in helping journalists to follow guidelines and comply with codes (Cea, 2013). Besides, some disclosure policy can be agreed between the media and the pollsters. The reputation of pollsters is damaged when surveys fail to predict electoral results, so they are the most interested in disclosing their procedures if they are properly done. So, if it is not possible to give all the information about the poll in the media website, researchers could offer space in their own websites. As Smith and Verral (1985) noted, ‘whether polls are accurate is the responsibility of the pollsters: whether they are perceived to be accurate is largely a function of their coverage in the news media’ (p. 58). As such, research and media organizations could and should work more closely with each other to put forth the best news reporting of poll data.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author biography

Idoia Portilla holds a PhD in Economics and Business Administration from Universidad de Navarra, where she currently is a Lecturer in the Department of Media Management at the School of Communication, and serves on the executive board of Master de Investigación en Comunicación (Research in Communication). Her areas of research are television and new media audience research, ethics and methodology in new marketing research and the effects of new media on traditional media. She is Chair of the Professional Standards Committee of Asociación Española de Estudios de Mercado, Marketing y Opinión (AEDEMO) and is a member of ESOMAR (The World Association of Research Professionals for Market, Social and Opinion Research).

References

Aalberg

Van Aelst

(2014) Who is afraid of preelection polls? How perceptions of polls influence support for polling regulations among elites. International Journal of Public Opinion Research. Available at: https://dx-doi-org.web.bisu.edu.cn/10.1093/ijpor/edt040 (accessed 20 May 2014).

Almazán

Villarejo

(1998) Análisis del discurso de la prensa sobre las encuestas electorales en las elecciones generales de 1996. Empiria: Revista de metodología de ciencias sociales 1: 97–120.

Andersen

(2000) Reporting public opinion polls: The media and the 1997 Canadian election. International Journal of Public Opinion Research 12: 285–298.

Barrera

Zugasti

(2006) The role of press in times of transition. The building of the Spanish democracy (1975–78). In: Voltmer

(ed.) Mass Media and Political Communication in New Democracies. Oxon: Routledge, pp. 23–41.

BBC (2014) Section 10: Politics, public policy and polls. Available at: http://www.bbc.co.uk/editorialguidelines/page/guidelines-politics-practices-opinion/ (accessed 25 August 2014).

Biroli

Miguel

Mota

(2012) Opinion polls and the media in Brazil. In: Holtz-Bacha

Strömbäck

(eds) Opinion Polls and the Media: Reflecting and Shaping Public Opinion. Basingstoke: Palgrave Macmillan, pp. 135–154.

Brettschneider

(1997) The press and the polls in Germany, 1980-1994 poll coverage as an essential part of election campaign reporting. International Journal of Public Opinion Research 9(3): 248–265.

Brettschneider

(2008) The news media’s use of opinion polls. In: Donsbach

Traugott

(eds) The SAGE Handbook of Public Opinion Research. London: SAGE, pp. 479–486.

British Polling Council (n.d.) Statement of disclosure. Available at: http://www.britishpollingcouncil.org/statement-of-disclosure/ (accessed 27 May 2014).

10.

Cea (2013) ¿Qué datos deben incluirse en la ficha técnica para publicar una encuesta? Available at: http://www.aedemo.es/aedemo/images/stories/varios/Datos_ficha_tecnica_publicar_encuestas.pdf (accessed 20 May 2014).

11.

Chang

(1999) Reporting public opinion in Singapore: Journalistic practices and policy implications. Harvard International Journal of Press/Politics 4(11): 11–28.

12.

Commonwealth Broadcasting Association (2004) CBA Editorial Guidelines. Available at: http://unesco.org.pk/ci/documents/publications/CBA%20Editorial%20Guidelines.pdf (accessed 25 August 2014).

13.

De Vreese

Semetko

(2002) Public perception of polls and support for restrictions on publication of polls: Denmark’s 2000 Euro Referendum. International Journal of Public Opinion Research 14(4): 367–390.

14.

Díaz de Rada

(2002) La ficha técnica de la investigación información imprescindible pero. ¿suficiente? Metodología de Encuestas 4(2): 235–242. Available at: http://casus.usal.es/pkp/index.php/MdE/article/view/926/867 (accessed 16 March 2015).

15.

Esomar (2007) The ICC/ESOMAR Code on Market and Social Research. The Netherlands. Available at: http://www.esomar.org/uploads/public/knowledge-and-standards/codes-and-guidelines/ICCESOMAR_Code_English_.pdf (accessed 1 August 2014).

16.

Esomar and Wapor (2009) Esomar/Wapor guide to opinion polls and published surveys. The Netherlands: Available at: http://www.esomar.org/knowledge-and-standards/codes-and-guidelines.php (accessed 20 May 2014).

17.

Esomar and Wapor (2014a) Esomar and Wapor key requirements for opinion polls and published surveys. The Netherlands: Available at: http://www.esomar.org/uploads/public/knowledge-and-standards/codes-and-guidelines/ESOMAR_WAPOR-Key-Requirements-for-Opinion-Polls-and-Published-Surveys.pdf (accessed 16 March 2015).

18.

Esomar and Wapor (2014b) Esomar/Wapor Guideline on Opinion Polls and Published Surveys. The Netherlands: Available at: http://www.esomar.org/uploads/public/knowledge-and-standards/codes-and-guidelines/ESOMAR-WAPOR-Guideline-on-Opinion-Polls-and-Published-Surveys-August-2014.docx (accessed 20 January 2015).

19.

Ferguson

De Clercy

(2005) Regulatory compliance in opinion poll reporting during the 2004 Canadian Election. Canadian Public Policy/Analyse de Politiques 31(3): 243–257.

20.

Frankovic

(2005) Reporting ‘the polls’ in 2004. Public Opinion Quarterly 69(5): 682–697.

21.

Frankovic

(2012) Opinion polls and the media in the United States. In: Holtz-Bacha

Strömbäck

(eds) Opinion Polls and the Media: Reflecting and Shaping Public Opinion. Basingstoke: Palgrave Macmillan, pp. 113–134.

22.

Gawiser

Witt

(n.d.) 20 Questions a Journalist Should Ask about Poll Results, 3rd edn. National Council for Public Polls (NCCP). Available at: http://www.ncpp.org/?q=node/4 (accessed 27 May 2014).

23.

Hair

Bush

Ortinau

(2003) Marketing Research: Within a Changing Information Environment, 2nd edn. Boston, MA: McGraw-Hill Irwin.

24.

Hardmeier

(1999) Political reporting in Swiss print media: Analysis and suggestions for quality improvement. International Journal of Public Opinion Research 2: 257–274.

25.

Holtz-Bacha

(2012) Opinion polls and the media in Germany: A productive but critical relationship. In: Holtz-Bacha

Strömbäck

(eds) Opinion Polls and the Media: Reflecting and Shaping Public Opinion. Basingstoke: Palgrave Macmillan, pp. 93–112.

26.

Kellner

(2014) A Journalists guide to opinion polls. Available at: http://www.britishpollingcouncil.org/a-journalists-guide-to-opinion-polls/ (accessed 27 May 2014).

27.

Maarek

(2009) Marketing político y comunicación: Claves para una buena información política. Barcelona: Paidós.

28.

Mateos

Penadés

(2013) Las encuestas electorales en la prensa escrita (2008–2011). Errores, sesgos y transparencia [Electoral polls in newspapers (2008–2011). Errors, bias and transparency]. Metodología de Encuestas 15: 99–119.

29.

Mattes

(2012) Opinion polls and the media in South Africa. In: Holtz-Bacha

Strömbäck

(eds) Opinion Polls and the Media: Reflecting and Shaping Public Opinion. Basingstoke: Palgrave Macmillan, pp. 175–197.

30.

Medvic

Dulio

(2008) The media and the public opinion. In: Rozell

Mayer

(eds) Media Power, Media Politics. Lanham, MD: Rowman & Littlefield Publishers Inc., pp. 233–256.

31.

Miller

Hurd

(1982) Conformity to AAPOR standards in newspaper reporting of public opinion polls. Public Opinion Quarterly 46(2): 243–249.

32.

Monzón

Dader

(1990) Las encuestas y su tratamiento periodístico. In: Muñoz

Monzón

Rospir

Dader

(eds) Opinión pública y comunicación política. Madrid: Eudema, pp. 465–487.

33.

Moy

Rinke

(2012) Attitudinal and behavioural consequences of published opinion polls. In: Holtz-Bacha

Strömbäck

(eds) Opinion Polls and the Media: Reflecting and Shaping Public Opinion. Basingstoke: Palgrave Macmillan, pp. 225–245.

34.

National Council on Public Polls (NCPP) (n.d.) Principles of disclosure. Available at: http://www.ncpp.org/?q=node/19 (accessed 27 May 2014).

35.

Organic Law 2/2011, of 28 of January, that modifies the Organic Law 5/1985, of 19 June, General Electoral System. Published in BOE, 25, 29 January, 9504–9523. Available at: http://www.boe.es/boe/dias/2011/01/29/pdfs/BOE-A-2011-1639.pdf (accessed 20 May 2014).

36.

Organic Law 5/1985, of 19 June, General Electoral System. Published in the BOE, 147, 20 June: 19110–19134. Available at: http://www.boe.es/buscar/doc.php?id=BOE-A-1985-11672 (accessed 20 May 2014).

37.

Paletz

Short

Baker

. (1980) Polls in the media: Content, credibility, and consequences. Public Opinion Quarterly 4: 495–513.

38.

Riffe

Lacy

Fico

(2005) Analyzing Media Messages: Using Quantitative Content Analysis in Research. Mahwah, NJ: Lea Communication Series.

39.

Sánchez-Aranda

(2005) Análisis de contenido cuantitativo de medios. In: Berganza

Ruiz

(eds) Investigar en Comunicación. Guía práctica de métodos y técnicas de investigación social en Comunicación. Madrid: McGraw-Hill, pp. 207–227.

40.

Smith

Verral

(1985) A critical analysis of Australian television coverage of election opinion polls. Public Opinion Quarterly 49: 58–79.

41.

Sonck

Loosveldt

(2008) Research note: Making news based on public opinion polls: The Flemish Case. European Journal of Communication 23(4): 490–500.

42.

Strömbäck

(2009) Vox Populi or Vox Media? Opinion polls and the Swedish media, 1998–2006. Javnost: The Public 16(3): 55–70.

43.

Strömbäck

(2012) The media and their use of opinion polls: Reflecting and shaping public opinion. In: Holtz-Bacha

Strömbäck

(eds) Opinion Polls and the Media: Reflecting and Shaping Public Opinion. Basingstoke: Palgrave Macmillan, pp. 1–22.

44.

Traugott

(2005) The accuracy of the national preelection polls in the 2004 presidential election. Public Opinion Quarterly 69: 642–654.

45.

Traugott

(2008) The uses and misuses of polls. In: Donsbach

Traugott

(ed.) The SAGE Handbook of Public Opinion Research. London: SAGE, pp. 232–239.

46.

Trujillo

(2002) Criterios de calidad de las encuestas. Metodología de Encuestas 4(2): 229–231. Available at: http://casus.usal.es/pkp/index.php/MdE/article/view/926/867 (accessed 16 March 2015).

47.

Vara

(1999) Conflictos políticos, conflictos electorales y sondeos. In: Rodríguez

Sádaba

(eds) Periodistas ante conflictos. Pamplona: Eunsa, pp. 141–147.

48.

Welch

(2002) Polls, polls, and more polls: An evaluation of how public opinion polls are reported in newspapers. The International Journal of Press/Politics 7(1): 102–114.

49.

Wichmann

Brettschneider

(2009) American and German elite journalists attitudes toward election polls. International Journal of Public Opinion Research 21(4): 506–524.

50.

Willnat

(2012) The good, the bad and the ugly: Public opinion polling in Taiwan. In: Holtz-Bacha

Strömbäck

(eds) Opinion Polls and the Media: Reflecting and Shaping Public Opinion. Basingstoke: Palgrave Macmillan, pp. 198–222.

51.

Wimmer

Dominick

(2013) Mass Media Research: An Introduction. Belmont, CA: Wadsworth Publishing.

52.

Wojcieszak

Palange

(2012) Media and politics in Spain. The Political Communication Report Available at: http://www.politicalcommunication.org/newsletter_22_2_spain.html (accessed 8 January 2015).