A Position Statement on the Conduct and Publication of Public Opinion Polls

Transcription

1 April 29, 2013 A Position Statement on the Conduct and Publication of Public Opinion Polls Board of Directors of the University of the Philippines School of Statistics Alumni Association (UPSSAA) In the run-up to the May 13, 2013 elections, media is again rife with discussion on the pros and cons of publishing public opinion polls 1 such as those of the Social Weather Station and Pulse Asia Survey. The discussions range from plain reporting of rankings of candidates on one hand to gross speculation on the other. As a consequence, readers at best remain neutral about polls. But what is worrisome, the polls and its publication are again perceived as threats to the democratic process. To allay these fears and corresponding calls to restrict the publication of public opinion polls it is right to put into broader perspective what purpose public opinion polls serve and why its publication should not be restricted. We, the members of the University of the Philippines School of Statistics Alumni Association (UPSSAA), stand by the use and unrestricted publication of public opinion polls as a means for information exchange between constituents and decision-makers and as a means for enriching collective choice and critical thinking in a maturing democracy. We ascribe to the same professional codes endorsed by the Philippine Statistical Association (PSA) and global professional research organizations such as the European Society for Opinion and Market Research (ESOMAR) 2 and the World Association for Public Opinion Research (WAPOR), that Public opinion is a critical force in shaping and transforming society and that Survey research of opinions gives the general public an opportunity for its voice to be heard. We believe that the use of public opinion polls helps STIMULATE THE FILIPINO VOTERS APPETITE FOR INFORMED DECISION-MAKING and, in the process, for the Philippines to rightfully ascend the path of a mature democracy. Choices and actions are best made in a climate of free inquiry and access to information. 1 An opinion poll is generally held to include all surveys of public opinion which are published to the general public. 2 Now also referred to the WORLD Association for Social, Opinion and Market Research in 2011 ESOMAR statutes

2 We recognize that true public opinion polls serve as a communication point between the electorate and the candidate. That true poll surveys provide IMPARTIAL information decision makers need about what the public thinks, needs and wants. As such, we need truly representative research that is: 1) PROPERLY CONDUCTED and 2) WELL-COMMUNICATED. We recognize that without SCIENTIFIC PUBLIC OPINION POLLING the public is left with unscientific, incomplete, or inaccurate assertions and, worse, likely presented with partiality by self-serving individuals or organizations. Societies would have no efficient way to assess citizens collective sentiments. We equally recognize that WELL-COMMUNICATED poll surveys help push different stakeholders to take action and make more meaningful and informed decisions. separated. We believe that opinion polls and freedom are intimately related. They cannot be Scientifically-conducted and well-communicated poll surveys in a climate of trust and free expression stirs up action of various stakeholders as can only be enjoyed in true democracies. Specifically: Journalists and political science experts cannot only can trace ups and downs of electoral campaigns; they can investigate winning (or losing) issues for the public to know Candidates act on how they connect (or disconnect ) as a platform and as a person Citizens know if their voices were really heard We believe ALL stakeholders benefit from good and well-communicated poll surveys.

3 On one hand, poll survey users (i.e. the public, politicians, decision makers) and other stakeholders get access to accurate measures for public attitudes and intentions (e.g. attitudes toward candidates) via well conducted and properly communicated public opinion surveys 3. On the other hand, broadcast media reports poll results in the service of the public so that as many people as possible, and not just the elite, are aware of the general public s sentiments. Together, users, doers, and media, raise the relevance of public polls and improve the climate of inquiry and action that help make a democracy vibrant. They all reap the benefits of a trusted feedback tool 4. Trust in poll surveys can only thrive with honesty, freedom of expression, respect for objectivity, and a strict adherence to a professional code of conduct. As such, we believe that the publication of public opinion polls should be unrestricted, as stakeholders build a climate of trust. In short, scientifically-conducted and well-communicated poll surveys can only thrive if researchers, media, decision-makers, and the public are jointly aware of the difference between good and bad research; between truthfully presented surveys and misleadingly presented ones. Specifically, in terms of three areas as guided by international standards: 1. The nature of the research techniques used and the efficiency with which they are applied; 2. The honesty and objectivity of the research organization carrying out the study; 3. The way in which the findings are presented and the uses to which they are put 5. 3 Locally, there are many forms of public opinion polls in the service of society. Banks and investors depend on the Banko Sentral ng Pilipinas (BSP) for its Business Climate Survey. There are self-rated poverty surveys. Educators mine surveys on the state of happiness of a country. 4 According to ESOMAR, the public s need to participate and media s need to ask questions have intensified the practise of using media polling for decades. Most mature democracies depend on public opinion polls to keep citizens, broadcasters and policy-makers engaged.

4 We believe that all public opinion survey firms should be transparent in their methodologies to assure the public that surveys have been conducted with strict compliance to the professional standards of survey research. For practical purposes, we endorse the guidelines espoused in the ESOMAR/WAPOR GUIDE TO OPINION POLLS AND PUBLISHED SURVEYS. ( uploads/ public/ knowledge-and-standards/ codes-and-guidelines/ WAPOR-ESOMAR_Guidelines.pdf). Research clients are encouraged to ask if poll researchers observe professional guidelines as endorsed by any professional research organization such as the PSA 6, MORES 7, ESOMAR and WAPOR. For instance, the WAPOR/ESOMAR Guidelines require that published research results show certain basic information to help establish the quality of information. [See ANNEX 1] We urge both research organization AND the client to each have a responsibility in the public interest to ensure that the published report on a public opinion poll DOES NOT misrepresent or distort the survey data. [See ANNEX 2]. Finally, we believe that the publication of survey results enriches the democratic process and is aligned with the University of the Philippines s advocacy of upholding the freedom of speech. Therefore, we implore: 1. The continued conduct and unrestricted publication of opinion polls provided these follow professionally-accepted standards. 2. Poll survey organizations to continue observing the highest professional standards in conducting scientific public opinion poll surveys. 3. Media to follow the internationally-accepted guidelines of reporting survey results. 5 ESOMAR International Code of Practise for the Publication of Public Opinion Poll Results 6 Philippine Statistical Association (PSA) 7 Marketing and Opinion Research Society of the Philippines (MORES)

5 4. That the collective appreciation of good and well-reported surveys be guided by internationally accepted codes of conduct for all stages of poll production, reporting, and communication. 5. That through a collective effort across stakeholders, an appropriate mechanism will be explored and selected that will ensure the quality standards of opinion surveys be adhered to and sustained. By respecting the value of scientifically-conducted and well-communicated poll surveys, we hope this, and all the above, will spur a dynamic and democratic social climate that is trusted, free, objective, and responsive to the needs of the Filipino people. Signed : Officers and Members of the Board of Directors University of the Philippines Alumni Association (UPSSAA) Nicco de Jesus Germaine Reyes Francisco de los Reyes Genelyn Ma. Sarte Lilia Catris-Guillermo Lorelie Santos Lans Lansangan Dr. Erniel Barrios Eric Bergara Fe Ferriols Nimfa Ogena Vivien Supanco Ofie Templo Rose Bautista

6

7 ANNEX 1. How To Tell Whether A Poll Report Should Be Taken Seriously or Not Per ESOMAR, research organizations publishing opinion polls should consistently provide certain basic information to enable audiences the opportunity of judging for itself the results and deciding whether or not it agrees with any conclusions drawn from the research. a) When any public opinion poll findings are published in PRINT MEDIA these should always be accompanied by a clear statement of : i. The name of the research organization carrying out the survey; ii. iii. iv. The universe effectively represented (i.e. who was interviewed); The achieved sample size and its geographical coverage; The dates of fieldwork; v. The sampling method used (and in the case of random samples, the success rate achieved); vi. vii. The method by which the information was collected (personal, telephone, etc.) The manner by which the questions were asked. b) In broadcast media, items (i) to (v) should at least be reported. c) The percentage of respondents who give don t know answers (and in the case of voting intention studies, OF THOSE WHO SAY THEY WILL NOT VOTE) must always be present when comparing the findings from different surveys, any changes (other than minor ones) in these percentages must also be indicated. d) In the case of voting intention surveys, it must be made clear if voting-intention percentages quoted include any of these respondents who answered don t know or may not/will not vote in reply to the voting questions asked. Full guidelines are available in the link below. Source:

8 ANNEX 2. Role of Research Client and Researcher in Reporting Research Results The research organization and the client each have a responsibility in the public interest to ensure that the published report on a public opinion poll does not misrepresent or distort the survey data. Misleading comments based on non-significant differences must be avoided. Special care must be observed to ensure charts and graphs do not convey misleading impression of trends over time or current survey results. Where any of the findings of a research project are published by the client the latter has the responsibility to ensure that these are not misleading. The Researcher/Research organization must be consulted and agree in advance of the form and content of publication, and must take action to correct any misleading statements about the research findings. The research organization must reserve the right to publish the total study and not just the technical specifications if : a shortened version of the publication distorts the analysis of the results an unforeseen and abridged version of the publication had to be used a publication which does not conform to the prior agreements The research organisation cannot normally be held responsible for any subsequent use made of public opinion poll results by people other than the original client. It should however be ready to immediately issue such comments or information as may be necessary to correct any cases of misreporting or misuse of results when these are brought to its attention. Source:

9 ANNEX 3. WAPOR/ESOMAR Guidelines on Pre-Election Polls. Pre-election polls are just one type of opinion polls. We believe that guidelines should be observed to achieve two things when doing pre-election polls: 1) Protect the interests of the voter in a democracy. Polling organizations should take ALL possible technical steps to ensure that polls published close to the vital decision point for voters are an objective guide to the state of public opinion and voting intentions. People do change their minds some even just a second before marking ballot slips. 2) Protecting the reputation of marketing (and poll) research. While it is true that opinion polls are a SNAPSHOT of intentions AT A SPECIFIC POINT IN TIME, the publication of this snapshot is treated by some as prediction. Pre-election polls are a very public test of sampling theory and survey research in action. Polls have a good track record for accuracy. But the occasional polls which appears to be wrong gets extensive media coverage. We believe that for the Filipino appetite for democracy and informed decision-making to grow, the following practical guidelines should be observed: 1. Sample Size. Pre-election polls should not have a sample of less than 1,000 respondents. In circumstances where the gap between leading parties is expected to be small, the sample size should be larger than 1,000. 1,500 to 2,000 or more should be used. 2. Weighting. The demographic profile of pre-election polls should be checked for representativeness and if needed, weighting should be applied to represent correctly the electorate. Polling companies should provide information about the variables used in the weighting as part of the full details of the survey published (or make available for people who enquire). 3. Survey Content. Wherever possible, pre-election polls should measure reasons for party choice or attitudes on issues or other aspects of the campaign. This point is especially of interest to audiences who are issues-, rather than, personality-based.

10 4. Time Series. The validity of the methods used by an organization can be judged better if they produce a series of voting intention estimates during the campaign. Any obvious bias will become apparent by comparison with the published polls of other organizations. 5. Consistent Design. Polling organizations should attempt to keep key elements of methodology consistent throughout the election campaign. 6. Rogue Polls. It is unacceptable to suppress a pre-election poll that looks out of line with earlier polls unless a valid technical reason why the poll is wrong has been established. It is also unacceptable to make a gut-feeling adjustment. The poll should be published with appropriate warning about the unusual poll result. The odds are strongly in favour of the poll being correct and voters do change their minds even in the polling booth. 7. Timing of Fieldwork. Polling companies should try to reduce the time elapsed between fieldwork and publication. Fieldwork period should also be such that people in full time employment are available for interview. Sources: ESOMAR/WAPOR Guide to Opinion Polls including the ESOMAR International Code of Practice for the Publication of Public Opinion Poll Results