Code of Conduct

Introduction

A1. The Australian Polling Council is an organisation whose aims are to advance the quality and understanding of public opinion polling in Australia and in furtherance of these purposes:

(a) to ensure standards of disclosure that provide an adequate basis for judging the reliability and validity of results of surveys that are published in the public arena;

(b) to encourage the highest professional standards in public opinion polling and advance the understanding of and the confidence in how polls are conducted and how to interpret poll results;

(d) to advance the Australian public’s understanding of how polls are conducted and how to interpret poll results; and

(e) to inform media, government and the public about best practice in the conduct and reporting of polls.

About the code

B1. This code of conduct applies to members of the Australian Polling Council. The code is divided into two sections, the first section relates to codes of conduct and standards that member organisations should follow at all times. The second relates only to individual research projects that bear the Australian Polling Council Quality Mark.

B2. Quality Mark refers to the tool used by the Australian Polling Council to indicate that a poll is compliant with this code. The tool may take the form of a graphic and/or text, and the details are independent of this Code

Standards which are to be followed at all times

The Code of Professional Behaviour

C1. All individuals are required to abide by all clauses of The Code of Professional Behaviour as maintained by The Research Society. For clarity, membership of The Research Society is not a requirement of membership of the Australian Polling Council.

No Disrepute

C2. All individuals and organisations are required to behave in a manner which does not bring the Australian Polling Council or its members into disrepute

C3. Individuals may only speak, or communicate via any method, on behalf of the Australian Polling Council if:

a. They have been elected to the office of Chairperson or Secretary; or

b. The Council has voted to allow them to be a spokesperson on a particular issue

C4. All individuals and organisations are forbidden from stating or implying that any work is compliant with the Australian Polling Council standards when it is not

C5. The Australian Polling Council will have a complaint handling mechanism which can be used by members, non-members, or by the public. Until such time as this mechanism is set up, complaints can be made by email directly to the elected Chairperson or Secretary. Any member of the Australian Polling Council who believes another member is in breach of the code has the right to remain anonymous when bringing the breach to the attention of the Australian Polling Council.

Use of the Australian Polling Council Quality Mark

D1. The Australian Polling Council Quality Mark can only be used by members of the Australian Polling Council

D2. The Australian Polling Council Quality Mark can be applied to Reports, Media Releases or other deliverables which have been conducted in compliance with this Code

D3. There is no requirement for member organisations to apply the Australian Polling Council Quality Mark to any work that they do – however any work which is political in nature and destined for publication should be compliant with the Code, or it should explicitly state that it is not compliant to comply with Paragraph C4.

D4. The Australian Polling Council Quality Mark can be used for any research conducted by members that complies with the requirements of this code regardless of the topic of the research.

D5. Members are obligated to comply with the code where the publicly released polling is related to political issues, or expectations and performance of public policy or governments. This is not limited to surveys which measure voting intention or political leader and party performance. It also encompasses studies evaluating government performance; political leaning; election, plebiscite or referendum outcomes; and those which focus on public policy and societal issues with political implications.

D6. Research conducted by Members which reaches the public domain intentionally with the knowledge and consent of the commissioning client or their authorised representative is expected to comply with the code. Research which is not intended for the public domain is not required to comply. Research released as the result of a Freedom of Information request, or jurisdictional equivalent, is not required to comply.

Standards which are to be followed for all projects which bear the Australian Polling Quality Mark

E1. Members are obligated to tell clients that for any work intended for the public domain, we are obligated by our membership to make public the information in both the Short and the Long Methodology Disclosure Statements

E2. A member should take all reasonable steps to ensure the Short Methodology Disclosure

Statement is published within the text of the content, or on associated charts or tables. Specifically:

a. Where the member is directly publishing the content, they are to ensure this information is published within the text of the content, or on associated charts or tables

b. Where the member is working with a journalist or writer, they are to ensure both the short and long methodology statements are provided to the journalist or writer, and to encourage publication within the text of the content, or on associated charts or tables

c. Where the member is not directly publishing content, nor working directly with a journalist or writer, a member must:

a. Ensure both the short and long methodology statements are provided to their clients

b. Encourage their clients to ensure an unabridged and complete version of the Short

Methodology Statement is published, or provided to the journalist or writer

In the event b or c above apply AND the Long Methodology Disclosure Statement is not provided by the entity publishing the content, Members will publish the relevant Long Methodology Disclosure Statement on their website within two business days of the researcher becoming aware the research has been published. The statement will remain accessible for a minimum of two years after being published on the Member’s website.

Short Methodology Disclosure Statement

F1. All content intended for publication, including but not limited to reports and media releases, is required to have a short disclosure statement which includes the following information:

a. The name of the research company conducting the research

b. The name of the client commissioning the research

c. If the research is commissioned by an intermediary, the name of the end client for whom the research is being conducted

d. The dates of the fieldwork when the research was conducted

e. The methodology/ies by which the research was conducted, which clearly delineates both the mode of data collection, and where there is any doubt, whether it was self-completion or interviewer administered. For clarity, automated telephone polls (robo-polls) must be clearly distinguished from operator administered telephone polls

f. If multiple methodologies are used, the proportion of completed surveys undertaken by each methodology

g. The target population for the research

h. The sample size of the research

i. That the research is compliant with the Australian Polling Council Quality Mark standards, including the URL where these are published

j. A URL where the Long Methodology Disclosure Statement is published

k. For surveys including voting intention, additional information to comply with section G2 below must be included.

Long Methodology Disclosure Statement

F2. The Long Methodology Disclosure Statement must include all the information in the Short Methodology Disclosure Statement, as well as the following:

a. Either the weighting efficiency, or the effective sample size of the research

b. The Margin of Error associated with the effective sample size for the research

c. The variables to which the research was weighted, including the source of the data used to calculate the target weights

d. The weighting methodology used (e.g. cell weighting, raking etc.)

e. The full question text for questions which are published, including any preamble, and whether response categories were visible/read to participants and whether the response categories were presented in random order or not

f. If any questions were asked prior to the questions which are published which may have materially influenced results, these prior questions should also be published

g. If telephone sample is used, the proportion of completed interviews on mobile / landline / other

h. If online sample is used, whether respondents were selected from an online panel, river sampling, or whether participants were able to opt in to the survey.

Research Principle Adherence

G1. The following principles should be adhered to for all published research:

a. The research must be free from intentional bias

b. The research should be fit for publication

c. The research must be reported in a manner which is not intentionally misleading

d. Research results must not be manipulated with the intention to make them similar to research published by others.

e. Order effects are taken into consideration, both within and across questions.

G2. The following principles apply specifically to polls publishing voting intention

a. Voting intention figures should only be published as representative of the population where the questions are asked before any other questions or preamble which could influence results. This includes questions about issues or attitudes to policies. This does not apply where voting intention responses are used solely for the purpose of sub-group analysis of other questions in the poll.

b. Reporting of polls measuring voting intention should identify whether participants were allowed to state that they are undecided, or whether an answer was forced.

c. If published voting intention figures are published with the undecided participants excluded, the proportion who were thus excluded should be published

d. If the undecided participants are not excluded in the published voting intention figures, the proportion who are undecided must be explicitly published and separated from “other” voters. NB: if undecided voters are not excluded, it is not meaningful to refer to swing or compare with historical voting figures from previous elections or other polls where undecided voters were excluded.

e. Where a Two-Party Preferred figure is presented, it must be stated whether this was collected by stated preference or historic preference flows. If some other method (for example modelling) is used, this must be disclosed.