--- Page 1 ---
Report of Dr Michael Turner
Acknowledgement, Instructions & Details About the Author
1. Name & Address
1.1.
1.2.
My name is Michael Thomas Eugeniusz Turner PhD MSc BSc CMRS
My place of work is C|T Group Research. Strategies. Results., Level 26,
Governor Macquarie Tower, 1 Farrer Place, Sydney, 2000, New South Wales,
Australia
2. Acknowledgement
2.1.
I confirm that I have read the following documents and agree to be bound by
them:
2.1.1.
The Federal Court of Australia: Harmonised Expert Witness Code
of Conduct
The AAT guidelines on Persons Giving Expert and Opinion
The Federal Court of Australia: Expert Evidence Practice Note
2.1.2.
2.1.3.
3. Instructions
3.1.
Attached is a letter of instruction that I received from Harpur Phillips dated
3rd March 2021 to undertake a voter research survey with appropriate
parameters and provide a report in relation to the results of that survey
(Appendix 7).
3.2. On 18th March 2021, I provided to Harpur Philips, the research report that is
contained in appendix 6.
3.3.
Attached is a letter of instruction from Harpur Phillips dated 29th June 2021,
to provide a further supplementary report in relation to the original research
report (Appendix 8).
4. Qualifications & Experience
4.1.
I am a trained social and political scientist, having worked in the fields of
social, political and market research for more than a decade. I am qualified in
the practice of research methods, having completed methods-based courses
at both Bachelors and Masters degree level. My full curriculum vitae is
provided in appendix 1.
4.2. My Masters degree in Social Research covered many areas of research
methods, including quantitative data analysis, survey design, sampling and
analysis of quantitative results. My Masters thesis focused on the
quantitative analysis of public opinion data.
1
--- Page 2 ---
4.3.
4.4.
4.5.
4.6.
4.7.
I continued methods training during my doctoral studies in Political Science
at The Elections Centre, Plymouth University. My PhD thesis focused
significantly on the quantitative analysis of elections results and survey data1.
Since my academic studies, I have continued training, attending many
renowned courses on survey design and analysis, across several institutions
and industry bodies. Most relevant to this research, are the methods courses
that I have attended at the University of Essex, the University of Vienna, and
the University of Oxford.
I have also conducted industry recognised short courses in experimental
survey designs run by the Market Research Society in the UK. All these
courses have some relevance for the conduct and analysis of this research
project.
I am a certified member of the Market Research Society (MRS), a member of
the Research Society (RS) and a Fellow of The Royal Statistical Society
(RSS).
Being members of these industry bodies binds me to higher standards of
professional practice of research and statistical analysis. I am obliged, in my
professional activities, to uphold the MRS Code of Conduct, to comply with
the RS Code and abide by the RSS Code of Conduct.
4.8. My experience also includes working to produce representative and accurate
surveys that gather official population estimates, for the Office for National
Statistics (ONS), the national statistics agency for the UK.
4.9.
4.10.
4.11.
During my time as an Assistant Methodologist in the methodology
department at the ONS, I specialised in sample design and estimation,
working on projects such as the Census Non-Response Linkage Study
(CNRLS), which sought to understand how to make opt-in surveys more
accurate by using data from the Census, which is compulsory.
I was also responsible for the sample design of several key surveys, to
ensure that they were representative of the population.
Since my time at the ONS, I have held senior positions at several leading
social, political and market research firms.
4.12. Though I am a methods generalist, I have built up substantial expertise in
quantitative survey methods, design, and analysis. I have produced robust
research and insight for a range of academic institutions, corporations,
media, charitable organisations, government, and official agencies, including
the ONS and the Electoral Commission in the UK.
1 Turner, M (2014). Incumbency Effects in English Local Elections 1974-2010: Assessing the Advantage of
Electoral Defence, University of Plymouth, URL: https://pearl.plymouth.ac.uk/handle/10026.1/2994
2
--- Page 3 ---
Research Design & Methodology
5. Purpose of the Project & Considerations
5.1.
The purpose of this research project was to provide robust evidence to
identify the impact of including a political party named “The New Liberals”
into the voting process in Australian Federal elections. The following
questions guided the research design and analysis of results:
5.1.1.
Does the inclusion of a party called ‘The New Liberals’ make it more
difficult for voters to correctly identify the political party that they
want to vote for?
Does the inclusion of a political party called ‘The New Liberals’ lead
to confusion, among voters, about the nature of the relationship
between ‘The New Liberals’ and other political parties, that are
already registered?
5.1.2.
5.3.
5.4.
5.2. When designing the research approach required to sufficiently answer these
research questions, several primary criteria were taken into account.
First, it is important to have the ability to simulate the voting process as
closely as practically possible in a private environment, free from interviewer
and social desirability biases.
The voting process is a text-based exercise for voters. They follow written
instructions, not oral ones. The ballot paper is not read aloud for all voters by
polling officials. As is outlined in the Australian Electoral Commission’s
(AEC’s) own official resources, it is important “to read and follow the
directions on the ballot paper to make sure your vote is counted”2. In other
words, the best way to simulate the voting process is for our sample to take a
text-based, self-completion approach.
Second, to firmly establish whether voters were confused, or truly able to
correctly identify their political party of choice, the ability to convey more
detailed information to voters, about each political party, after they have
given their vote intention, would be required.
To know if The New Liberals are seen by voters as a separate political party
from existing ones, or a related, or even the same political party, we must be
able to corroborate voters’ intentions. We can do this by presenting them
with non-political, factual information about each party, that will not appear
on the ballot paper, and therefore cannot be considered simply in the act of
voting.
Information such as the party leader, the year of their formation and the
nature of the relationship between the parties has to be presented to voters
5.6.
5.5.
5.7.
2 Voting in Australia, Australian Electoral Commission, pp14, URL: https://education.aec.gov.au/teacher-
resources/files/voting-in-australia.pdf
3
--- Page 4 ---
5.8.
5.9.
for them to make an informed decision about who the party is, and whether
they intended to vote for them.
Finally, the ability to run an experimental design is essential in order to
provide robust evidence that establishes the causal effect of inclusion of ‘The
New Liberals’ on the ballot, and the magnitude of any electoral impact.
For this reason, a sample size of more than n=2,000 would be required to
provide sufficient confidence in the estimates provided. This will mean that
each of the four different ballot papers required to be evaluated will have a
minimum sample of approximately n=500.
5.10. The ballot designs were as follows.
Ballot A: The New Liberals are not displayed (a control ballot A to compare
results against).
Ballot B: The New Liberals are displayed below the Liberal Party (test ballot).
Ballot C: The New Liberals are displayed above the Liberal Party (test ballot).
Ballot D: New Liberals are displayed but the Liberal Party is not displayed
(test ballot).
By simulating different levels of prominence for the Liberal Party and The
New Liberals, while keeping other factors constant, the experimental design
ballot evaluation can confirm whether inclusion of The New Liberals
disproportionately confuses Liberal voters.
If voters felt The New Liberals were an entirely separate party, unrelated to
the Liberal Party, then there is no reason to suggest that their inclusion and
position on the ballot would disproportionately impact on the Liberal party,
while leaving other political parties, across the political spectrum, relatively
unchanged.
5.11.
5.12.
5.13. An example ballot paper for Federal elections to the House of
Representatives (Figure 1, Appendix 4) shows clearly that a Liberal Party of
Australia candidate and a candidate for The New Liberals can be placed
some distance away from each other (due to random order of the
candidates/political parties). Therefore, there is a high chance that voters will
see and read ‘The New Liberals’ before knowing if there is a candidate
standing for the Liberal Party.
In this scenario, ballot order effects are very likely to lead to erroneous
results if voters conflate the two parties or if they feel that two parties are
related or connected in the same way that other the Coalition parties are, or
divisions within political parties.
5.14.
4
--- Page 5 ---
5.15. The ballot order effect is well-known phenomenon confirmed by the
academic publications for both experimental election design3 and for actual
election results in the UK4, and Australia5.
5.16. Given these primary considerations, it is clear that a computer-assisted, self-
completion, text-based approach would be required for this research.
It is worth noting that a telephone survey approach was ruled out on several
grounds, including, but not limited to:
5.17.1.
5.17.
5.17.2.
5.17.3.
5.17.4.
First, that the introduction of an interviewer in the process could
lead to erroneous effects, such as interviewer and social
desirability biases, which are well known in social and market
research6.
Second, the inability to simulate as closely as practically possible,
the voting process, such as reading instructions and observing a
mock ballot.
Third, in order to verify whether voters intended to vote for a
particular party, there is a requirement to convey significant and
detailed information about several political parties to voters. It is an
unreasonable expectation for voters to answer this question on an
informed basis, without it available to them to consider in written
form.
Fourth, telephone response rates have been in significant decline in
recent years7, and when probability-based survey designs have low
response rates, the tools that are used for remediation, are similar,
if not identical, to those used for non-probability samples8. This
means there is likely no benefit in taking a telephone or a paper-
based approach from a sampling perspective. The bias introduced
by high levels of non-response is unlikely to guarantee more
accurate results than an adequately weighted non-probability
design.
5.18. The only probability-based survey approach that would yield any material
benefit to the sample design, whilst meeting the modal criteria, would have
been a simple random sample, household, computer assisted personal
3 Johnson, J & Miles, C (2011). Order effects of ballot position without information-induced confirmatory bias,
British Politics, pp479-490, URL: https://link.springer.com/article/10.1057/bp.2011.26
4 Webber, R et al (2012). Ballot Order Positional Effects in British Local Elections, 1973-2011, Parliamentary
Affairs, Vol 67(1), pp119-136, URL: https://academic.oup.com/pa/article-abstract/67/1/119/1539298
5 King, A & Leigh, A (2009). Are Ballot Order Effects Heterogeneous?, Social Science Quarterly, Vol 90(1),
pp71-87, URL: https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1540-6237.2009.00603.x
6 Alazar, M (1990). Interviewer Bias: How it Affects Survey Research, AHAOHN Journal, Vol 38(12), pp568-
569, URL: https://journals.sagepub.com/doi/pdf/10.1177/216507999003801203
7 Kennedy, C & Hartig, H (2019). Response rates in telephone surveys have resumed their decline. Pew
Research Center. 27/02/2019, URL: https://www.pewresearch.org/fact-tank/2019/02/27/response-rates-
in-telephone-surveys-have-resumed-their-decline/
8 Kennedy, C et al (2016). Evaluating Online Nonprobability Surveys. Pew Research Center, 02/05/2016,
URL: https://www.pewresearch.org/methods/2016/05/02/evaluating-online-nonprobability-surveys/
5
--- Page 6 ---
interview (CAPI) approach. A computer-based, self-completion survey that is
delivered to households via trained interviewers.
5.19. This approach would have been both impractical, and prohibitively expensive
to conduct, for little additional benefit to a research project of this scope and
nature. Even for a very short survey, the fieldwork required would have taken
several weeks, and could have cost several hundred thousand dollars.
5.20. Trained and experienced researchers regularly make judgements about the
5.21.
size and scope of studies that are required to give adequate evidence to
answer research questions. Researchers identify the appropriate approach
and design, based on the timeframe, resources and level of precision
required to come to a set of conclusions that we can be confident about.
In conclusion, given the objectives of this research project, and key
considerations outlined above, it is my expert view that the optimal approach
and most appropriate research design for this project, was for the survey to
be conducted online, via a self-completion survey approach, of sufficient
sample size (n=2000+), using commercially available panel providers, taking
careful consideration of the sample design (i.e. stratification), and following
best practice guidance on questionnaire design and analysis of results.
6. Overview of Approach
6.1.
6.2.
6.3.
6.4.
The overall approach taken was typical of much of the high-quality public
opinion and market research that is produced for academic, commercial, and
public policy focused organisations. Though there are very slight differences
across the designs, the general approach is similar to polls such as ANUPoll9,
Newspoll10, the Global Consumer Confidence Index11, and polling by the Lowy
Institute12.
The research findings are based on results from survey responses of
n=2,036 voters aged 18 years and over, who are eligible to vote in Australian
Federal Elections.
The fieldwork was conducted between the 6th and 9th of March 2021 using an
online approach.
The online research panel comprised a number of traditional actively
managed market research panels that either adhere to the ESOMAR
standards or are ISO20252 certified. Multi-source sampling approach aimed
to mitigate any potentially inherent bias that might occur from using a single
source of online sample.
9 Methodology, ANUPoll, Australian National University,
https://csrm.cass.anu.edu.au/research/surveys/anupoll/methodology
10 Ramadila, N (2018). Australian Polling Council: Public Polling Methodology Statements, YouGov,
18/05/2021, https://au.yougov.com/news/2021/05/18/apc/
11 Jackson, C & Silverstein, K (2021). Global consumer confidence at highest reading since March 2020,
Ipsos, 17/06/2021, URL: https://www.ipsos.com/en/global-consumer-confidence-June-2021
12 The Lowey Institute Poll, 2021 Methodology, https://poll.lowyinstitute.org/report/2021/methodology
6
--- Page 7 ---
6.5.
6.6.
6.7.
6.8.
Participation was by invitation only. Invitations were sent at random (a
double-blinded selection process) to members of online panels, via a
stratified design with proportions set according to age-sex interlocking and
state/territory. These were based on Australian Bureau of Statistics (ABS)
Census results from 2016.
The questionnaire design, fieldwork, and analysis of survey results followed
code of conduct and best practice guidance from both The Research
Society13 and the Market Research Society14.
As per industry standards, all participants gave their consent for survey
responses to be used for market research purposes, by opting-in to the
study via a process of informed consent.
Calibration weights were used to ensure that the final sample was
representative of the Australian electorate, using weighting targets derived
from official population statistics, and Federal election results collected by
the Australian Electoral Commission (AEC).
7. Questionnaire Design
7.1.
7.2.
7.3.
7.4.
7.5.
The questionnaire (Appendix 1) was developed with the intention of collecting
quantitative evidence that is both representative and statistically
generalisable, to help answer the research questions detailed above.
Careful consideration was given to the questionnaire design. Specifically,
care was taken to minimise any bias that may arise from question ordering,
wording, or the order of response options.
To ensure that the survey conformed to the established understanding of
survey response processes, such as ensuring the comprehension of
questions, that they have reasonable retrieval and judgement requirements,
and have adequate response frames15, the survey was drafted entirely by me,
drawing on my more than 10 years of experience as a social researcher and
a highly qualified market research professional.
As is standard practice in market research, the survey was then piloted on a
small sample of respondents (n=30). An optional, text response, feedback
box was given at the end of the survey, which returned no queries or
concerns about the design of questions, or the questionnaire overall.
The feedback box was then removed, before the survey was full launched for
the remaining n=2,006 responses.
7.6. Where practical, the order of questions, and question response options were
randomised. These have been detailed in the questionnaire (Appendix 1).
13 Best practice for interviews and questionnaires, The Research Society, April 2021, URL:
https://researchsociety.com.au/documents/item/2800
14 MRS Guidelines for Online Research (2014), Market Research Society, URL:
https://www.mrs.org.uk/pdf/2014-09-01%20Online%20Research%20Guidelines.pdf
15 Tourangeau, R et al (2000). The Psychology of Survey Response, Cambridge University Press, pp7-16
7
--- Page 8 ---
7.7.
Key considerations that were taken into account in the questionnaire design
include:
7.7.1.
Ensuring that key prevalence questions, such as 9a, 9b, 9c, 9d, 11,
12, 13, and 14, were placed earlier in the survey to ensure that
statements containing both the name and details of ‘The New
Liberals’ did not have an influence on responses to these questions.
Prevalence questions are designed to accurately measure the
proportion of people with a particular view, or voting intention, at a
particular timepoint. To get accurate results it is important to
minimise the placement of subjective content that may raise their
awareness of an issue, or in this case a political party, in advance of
asking them.
Questions that were less likely to be impacted by the presentation
of new information to respondents, such as attitudinal statements
requiring judgment, or clarification of previous responses, were
asked towards the end of the survey in order to ensure that the
new information did not impact on questions that sought to identify
what voters understood about the specific nature of the
relationship between the Liberal Party of Australia, and ‘The New
Liberals’.
7.7.2.
7.7.3. Where practically possible, question answer options were either
randomly rotated, or the direction of agree/disagree Likert scales16
was randomly assigned (i.e. randomly ‘flipped’) in order to ensure
that the order of response options did not impact or bias results in
a particular way (Appendix 2).
7.8.
7.9.
Given the requirement for quantitative results that are both robust and
generalisable, the use of clear and plain language closed questions was
favoured over ‘loosely framed’ open questions.
Survey researchers tend to favour closed-ended questions as they are
simpler to understand and help clarify the meaning of the question for
respondents (see Tourangeau et al. 200017).
7.10. Open-ended questions can suffer from issues of generalisability,
categorisation, and bias, including but not limited to:
7.10.1.
Respondents to open-ended questions will often give different
degrees of detail when answering, which can be associated with
different levels of engagement. If more engaged respondents give
clearer, more thorough, and more detailed responses, then this will
significantly bias the coding and classification of their responses in
favour of a more highly engaged set of respondents.
16 A Likert scale is a closed-ended scale question, that enables people to self-report the extent to which
they agree or disagree with a given proposition.
17 Tourangeau, R et al (2000). The Psychology of Survey Response, Cambridge University Press, pp231
8
--- Page 9 ---
7.10.2. Using open-ended questions can lead to some answers becoming
unconnected or lost in detail that is either insignificant or irrelevant
to the topic we are interested in. An average voter cannot be
expected to understand intricate detail of the Commonwealth
Electoral Act. So, in circumstances where the subject matter may
be difficult to understand, or requires significant expertise, or
background knowledge, it is better for research professionals to
craft clear and easy to understand closed questions.
7.10.3. Analysis can become too focused on answers from respondents of
advantaged groups, such as the highly literate, older respondents,
those who regularly write or use computers, which will bias results
in favour of the experienced and technically skilled18, and could
result in greater survey drop-out rates19.
7.11.
These concerns are just some that led the report author to conclude that
open-ended questions would provide little, if any, additional benefit for this
research project. Consequently, the questionnaire design favoured an
approach that gave respondents as much freedom of response as was
reasonable, within a closed-end response environment.
7.12. The questionnaire design followed code of conduct and best practice
guidance from both The Research Society and the Market Research Society.
Market research industry guidance includes a requirement that participants
of any market research process should give their consent to participate in
the study on an informed basis about how their response data will be used.
This is the case, no matter the mode through which data is collected, online,
telephone, or face-to-face.
7.13. All respondents of this study were informed that their responses would be
anonymised and used for market research purposes (Appendix 1). The final
sample of respondents opted-in to the research process by giving their
consent for use of their data for research purposes.
8. Survey Sample
8.1.
The process taken to achieve a representative sample of Australian voters
for this research project, is similar in nature, to other online market research
projects that I have been commissioned to undertake in the past. It is now a
standard industry approach used regularly by social and market research
firms, for commercial and government clients, in Australia and globally.
18 Reja, U et al (2003). Open-ended vs. Close-ended Questions in Web Questionnaires, Advances in
Methodology and Statistics, Vol 19, pp173-174, URL: https://1library.net/document/zp7dm34z-open-ended-
vs-close-ended-questions-web-questionnaires.html
19 Desai, S & Reimers, (2019). Comparing the use of open and closed questions for Web-based measures of
the continued-influence effect, Behavior Research Methods, Vol 51, pp1426-1440, URL:
https://link.springer.com/article/10.3758/s13428-018-1066-z
9
--- Page 10 ---
8.2.
8.3.
8.4.
8.5.
8.6.
8.7.
The survey was conducted online, with invitations sent to a sample of online
panellists via a random invitation stratified design that was representative of
adult voters in Australia.
In order to send invitations to a sample of residents, I used an intermediary
exchange, who gave access to a number of specialist panel companies,
sometimes referred to as ‘panel providers’. These panel providers are
responsible for engaging, recruiting, and managing approximately 1.57 million
residents living in Australia.
Participants are typically paid a small and standardised fee to take part in
market research surveys, based on the length of each survey that they
complete.
A wide range of social and market research studies, for commercial and
government clients, is conducted in this way.
Although I was responsible for the sample design (i.e. setting the strata and
final sample targets), and for designing and hosting the online questionnaire, I
was not responsible for inviting Australian residents to participate and had no
knowledge or means to invite people to take part. This was the responsibility
of the panel provider.
Respondents were invited through a randomised, double-blinded process.
Neither me, nor the online panels, nor respondents, influenced which specific
individuals could take part. The process for selecting panellists to invite, is an
automated and randomised one.
8.8. Once the stratification targets were set, panellists were selected at random
within key strata. Sample stratification targets were age-sex interlocking and
state/territory. These were based on Australian Bureau of Statistics
population estimates and data from the 2016 Census. The target sample
structure can be found in table 1, below.
These selected panellists were then invited to take part in a study, the
content of which they did not know.
In total, 71 panels were used, via an intermediary exchange.
The panel provider sends all randomly selected panellists a notification,
usually via email, but also via other means, that contains the length of the
survey, the expected reward for completion, and a web address link to the
survey. No information about the survey’s subject or content is provided in
advance.
If they have received the invitation, and would like to take part, then they can
click the link in the invitation where they will be taken to the start of the online
hosted survey. The survey was hosted and run by me.
8.9.
8.10.
8.11.
8.12.
8.13. When they arrive at the web address, they are given information about how
their data may be used and the length of the survey. They will be asked to
consent to their data being used for market research purposes, which forms
the basis of question 1 of the questionnaire (Appendix 2).
10
--- Page 11 ---
8.14.
‘Soft’ quota survey completion caps were set for age-sex groups (10%), and
state/territories (25%) in order to mitigate the effects of differential non-
response. This is a typical approach used widely in market and social
research20.
8.15. Approximately 11,110 respondents were invited to take part in the survey.
8.16. The total number of eligible residents who started the survey was n=2,265,
this puts the estimated response rate at approximately 20.4%, which is
considered typical for online research21.
8.17. After screening, the survey had a completion rate of 90%, meaning that of
the 2,265 people who started the survey, 2,036 completed it.
8.18. The final completed sample size was n=2,036 respondents, who were all
8.19.
Australian residents aged 18 years and over, and eligible to vote in Australian
Federal elections. This forms the final dataset used for analysis.
It is now standard in the social and market research industry to assign
confidence intervals to similarly designed online survey research. Examples
include but are not limited to; YouGov’s Newspoll22 which takes a classical
approach to calculating the margin of error, and online polling conducted by
market research firm Ipsos MORI23 which uses Bayesian Credibility
Intervals24.
8.20. Online, random invitation, stratified design approaches have also been shown
to be as accurate, if not more so, than modern-day random digit dial (RDD)
telephone surveys, with the correct weighting approach25.
8.21. The ‘panel providers’ used were actively managed market research suppliers,
and all are ISO2025226 compliant and all comply with the standards that
ESOMAR27 provides.
8.22. These suppliers recruit panellists using a variety of approaches, including but
not limited to, random probability designed surveys, face-to-face and
telephone studies, referral programmes, direct marketing, and online adverts.
In order to mitigate any inherent bias that might occur from using a single
source of panel provider, based on their recruitment or management style, a
8.23.
20 What are opinion polls? Guidance on how to read opinion polls, The Research Society, January 2021,
https://researchsociety.com.au/documents/item/3063
21 Nulty, D (2008). The adequacy of response rates to online and paper surveys: what can be done?,
Assessment & Evaluation in Higher Education, Vol 33(3), pp301-314, URL:
https://www.tandfonline.com/doi/abs/10.1080/02602930701293231
22 Ramadila, N (2018). Australian Polling Council: Public Polling Methodology Statements, YouGov,
18/05/2021, URL: https://au.yougov.com/news/2021/05/18/apc/
23 Polling Methodology Disclosure Statements, Ipsos, URL: https://www.ipsos.com/en-au/polling-
methodology-disclosure-statements
24 Credibility Intervals for Online Polling (2012). Ipsos Public Affairs, URL:
https://www.ipsos.com/sites/default/files/2017-03/IpsosPA_CredibilityIntervals.pdf
25 Kennedy, C et al (2016). Evaluating Online Nonprobability Surveys. Pew Research Center, 02/05/2016,
URL: https://www.pewresearch.org/methods/2016/05/02/evaluating-online-nonprobability-surveys/
26 ISO20252 is the international standard for organisations that produce market, opinion, and social
research, including insights and data analytics.
27 European Society for Opinion and Marketing Research (ESOMAR) is a membership organization for
market, social, and opinion researchers.
11
--- Page 12 ---
--- Page 13 ---
8.28. The results show that final, pre-weighting sample is reasonably close to
expected counts based on official population statistics. This suggests that
the sample is, generally, of a good quality.
9. Post-Fieldwork Weighting
9.1.
To ensure the results were both accurate, and generalisable, calibration
weights were applied to the final collected data. Calibration weighting is an
approach for reducing the error of population estimates that are produced
from a survey sample by, applying mild adjustments to the sample.
Calibration weighting does this by adjusting certain ‘calibration’ variables, to
equal known, or better-estimated population totals.
9.2. Weighting targets included age-sex interlocking, state/territory, level of
9.3.
9.4.
9.5.
9.6.
9.7.
9.8.
9.9.
education and 2019 Federal election primary vote.
Targets were derived from population statistics collected by the Australian
Bureau of Statistics (ABS), as well as the results of the 2019 Federal election
collected by the Australian Electoral Commission (AEC).
A raking technique was used for calculating the weights, which is a typical
approach in market and social research28.
The smallest weight assigned was 0.46, and the largest weight was 2.32,
suggesting that the weighting efficiency was good, and the final collected
sample was reasonably close to national population statistics for weighting
targets.
The effective sample size is a measure of the precision of a survey sample.
Even if a survey research study has a sample of 1,000 respondents, an
effective sample size of 500 would suggest that the final weighted results
are no more robust than a well-executed, but un-weighted simple random
sample of 500 respondents. So, it is important to take the weighting design
into consideration when assigning confidence to survey research.
Taking the weighting design into account, the final efficiency for this study
was 91.5%. This means that across the study as a whole, results for the poll
had an approximate effective sample size of n=1,863.
Please note that efficiency estimates will vary slightly across the question
results as the impact of the weight differs slightly for different statistics
calculated.
An approximate margin of error breakdown has been calculated for the study
toplines at the 95% confidence level (Table 2 below).
28 Mercer, A et al (2018). For Weighting Online Opt-In Samples, What Matters Most? Pew Research Center,
26/01/2018. https://www.pewresearch.org/methods/2018/01/26/how-different-weighting-methods-work/
13
--- Page 14 ---
--- Page 15 ---
--- Page 16 ---
--- Page 17 ---
Appendix 1 – Michael Turner PhD MSc BSc CMRS –
Curriculum Vitae
Academic Qualifications
PhD, Political Science, The Elections Centre, Plymouth University, 2009-2013
My doctoral study included the detailed quantitative analysis of survey and elections
data to examine whether electoral advantage can be assigned to incumbency, and if so,
to estimate its magnitude.
Thesis: Incumbency Effects in English Local Elections 1974-2010: Assessing the
Advantage of Electoral Defence.
Synopsis: The thesis serves as a major quantitative examination of the effects of
incumbency and campaigns in English local elections. The research presents evidence in
support of the hypothesis, that councillors have a mild, yet significant advantage and
constructs models to estimate its nature and electoral effect. The study utilises both
elections results and survey data.
Supervisors: Professor Colin Rallings & Professor Michael Thrasher
MSc, Social Research, Plymouth University, 2008-2009 – Pass with Merit
My postgraduate study focused extensively on social research fundamental design,
methods, analysis, and reporting of results, with modules teaching both quantitative and
qualitative approaches to research practice. Survey design, sampling, analysis, and
reporting were included as part of the programme.
Thesis: Do Londoners View the Implementation of a Directly Elected Mayor as a
Success of Failure?
Synopsis: The thesis assessed the introduction of a directly elected mayor in the capital
and whether it was viewed as a success through the eyes of Londoners. Primarily a
quantitative analysis, the research utilised elections, aggregate polling, and survey data.
The results showed that Londoner’s were relatively unfamiliar with the nuances of the
Supplementary Vote electoral system, that they treated the mayoral post in a
representative democracy style and were willing to break with their traditional party ties
to vote for candidates they believed to hold the best personal qualities to run the city.
Supervisor: Professor Colin Rallings
BSc, International Relations with Politics, University of Plymouth, 2004-2008 –
Upper Second Class with Honours
My undergraduate study included several modules that taught social research methods,
including primary data collection, data analysis, and reporting practice.
17
--- Page 18 ---
Thesis: An Analysis of the European Union’s Energy Security
Supervisor: Professor Karl Cordell
Relevant Professional Courses
Quantitative Data Analysis – Pass with Distinction - July 2012, Dr Sona Golder, 45th
Essex Summer School in Social Science Data Analysis – University of Essex
Course details: The Essex Summer School in Social Science Data Analysis provides
intensive social science methods training with instruction from an internationally
renowned faculty. The course covered a wide range of statistical methods for analysing
survey data, including core statistical methods, such as crosstabulation, t-tests, analysis
of variance, correlation, and various forms of regression. These methods are widely used
by academic and professional researchers in a range of fields such as political, social
and market research.
Comparative Cross-National Survey Design – Pass with Distinction – February 2012, Dr
Ineke Stoop, European Consortium of Political Research Winter School in Methods and
Techniques, University of Vienna
Course details: The European Consortium of Political Research (ECPR) is the leading
scholarly society for political scientists in Europe. The Winter School is designed to
equip participants with valuable new knowledge for their research. The course covered
key information about how to robustly design survey research, but also how to
systematically collect good-quality data for statistical analysis, how to compile and
manage these data.
Conjoint Analysis Masterclass - November 2017, Ray Poynter – Market Research
Society
Course details: The Market Research Society runs a number of short courses for
research professionals. This course provided detail on how to manage, design, and
interpret a wide range of experimental design projects, including conjoint analysis,
MaxDiff scaling and other variants of Discrete Choice Modelling (DCM) This course
covered material directly relevant to the experimental design ballot evaluation
conducted in this research.
Other professional courses
Mixed Methods Research – Unassessed - August 2010, Dr Jason Seawright, 43rd Essex
Summer School in Social Science Data Analysis – University of Essex
Multilevel Analysis - April 2012, Prof Tom Snijders & Daniel Stegmueller, Oxford Spring
School in Quantitative Methods for Social Research – University of Oxford
18
--- Page 19 ---
A Practical Application of Behavioural Economics in Research – Understanding Consumer
Behaviour – December 2017, Laura Parker - MRS
Introduction to Behavioural Economics – November 2017, Anthony Tasgal – MRS
Professional Experience
Head of Research (Director) – C|T RSR (Jan 2019 – Present)
In my current role, I specialise in producing public opinion, communications, social, public
policy and political insight for major corporations and organisations, in Australia and
across the globe.
The research tends to involve the collection of primary data, via quantitative and
qualitative approaches. This includes online surveys, which then feed into the production
of strategic research reports through statistical analysis of the data.
At C|T, I work across the full range of subject areas, and tailor research approaches and
methods to the project requirements, often utilising online panel-invitation survey
designs, when necessary, to produce accurate statistics and robust insight for our
partners and clients.
I am responsible for ensuring that C|T RSR’s research outputs are produced to the
highest standards of practice and oversee the design and analysis of all major research
projects, including online survey research projects.
I continue to produce research projects end-to-end, from inception and design to
analysis and reporting.
Director of Research & Head of Polling - BMG Research (May 2015 – Jan 2019)
I specialised in the conduct of public opinion, public policy, elections, and social
research. I led the research approach and methodology design for major projects for the
business and worked across the full range of quantitative survey research methods and
statistical analyses.
I was the company representative to the British Polling Council (BPC) and principal
academic liaison. I was programme leader for the BMG Research Graduate Training
Programme & the summer placements scheme, principally because of my academic
background and methods training.
I designed, managed, analysed, and reported on more than two hundred survey research
projects that required a detailed understanding of sample design, survey design,
statistical analysis, and reporting to the highest industry standards, across a range of
approaches, including online survey research.
I produced research for several key institutions including universities, government
departments and official agencies. These include but are not limited to; the Office for
National Statistics (ONS), the UK Electoral Commission, the Department for the
19
--- Page 20 ---
Environment, Farming and Agriculture (Defra), Department for Transport, Department
for Business, Energy and Industrial Strategy (BEIS), the Home Office, the Department
for International Trade (DIT), the University of Essex, and the University of Sheffield.
Director of Political & Social Research, Survation (Oct 2014 – May 2015)
During my time at Survation, I specialised in the conduct of elections, political and social
research. I worked across the full range of research methods, including online and
telephone survey research approaches.
I was responsible for designing, managing, analysing, and reporting on survey research
projects which required a detailed understanding of sample design, survey design,
statistical analysis, and reporting to industry standards.
Assistant Methodologist, Office for National Statistics (ONS) (Sep 2013 – Oct 2014)
During my time as an Assistant Methodologist at the ONS, I worked in the methodology
department, and was responsible for the sample design and calculation of corrective
weights for several major surveys.
My work on sample design was to ensure that major national surveys were
representative of the adult population, and my work calculating weights was to help
produce more accurate population estimates and official statistics which the ONS was
responsible for collecting.
While I was at the ONS, I worked on several major projects, including the Census Non-
Response Linkage Study (CNRLS), which was a study that sought to identify and
correct for those residents who took part in the compulsory Census, but tended not to
opt-in to other types of survey conducted by the ONS. This is a set of skills and
experience that I have taken forward, into the rest of my professional career.
My time at the ONS gave me valuable skills, experience, and expertise in the designing
of representative and accurate surveys, across a variety of modes of fieldwork and
interview completion.
Associate Lecturer, School of Management, Plymouth University (Sep 2009 – Jul
2013)
Research Assistant, The Elections Centre, Plymouth University (Mar 2012 – Apr
2013)
Membership of Professional Bodies
Certified Member of the Market Research Society – No. 2201375
20
--- Page 21 ---
The Market Research Society is the professional body and regulator of the market
research industry in the UK. They promote the highest professional standards
throughout the sector via the MRS Code of Conduct35.
Certified membership is the higher professional grade of membership and recognises
academic achievements and career experience of individuals in market research. It is
awarded to senior market research practitioners who have progressed to a leadership
level in research, insight, or analytics. Michael’s membership binds him to uphold the
MRS Code of Conduct.
Member of The Research Society
The Research Society is the highest level of organisation in Australia for those engaged
in market, social and opinion research. The Research Society regulates research
professionals by compelling its members to adhere to the Code of Professional
Behaviour36 that aligns with Australian Privacy requirements.
The Code is intended to foster public confidence and to demonstrate practitioners'
recognition of their ethical and professional responsibilities when carrying out research,
and Michael’s membership binds him to comply with the Code in his professional
activities.
Fellow of the Royal Statistical Society (RSS)
The Royal Statistical Society is a professional and learned society which, through its
members, has an obligation in the public interest to provide the best possible statistical
service and advice.
Professional membership of the Society is an assurance of ability and integrity. Fellows
are required to abide by the Code of Conduct37. In particular, they must:
1) Act in the public interest
2) Fulfil their obligations to employers and clients
3) Fulfil their obligations to the profession and the Society
4) Show professional competence and integrity at all times
The Code is mandatory for all professionally qualified fellows, and it is recommended to
all levels of fellow of the Society.
35 https://www.mrs.org.uk/pdf/MRS-Code-of-Conduct-2019.pdf
36 https://researchsociety.com.au/standards/code-of-professional-behaviour
37 https://rss.org.uk/RSS/media/File-library/About/2019/RSS-Code-of-Conduct-2014.pdf
21
--- Page 22 ---
Appendix 2 – Questionnaire
ASK ALL
1) Thank you for taking part in this survey. We hope you will find it interesting.
The survey takes, on average, around 10 minutes to complete and we have allowed
plenty of time for you to consider your responses carefully.
Please note, your responses will be treated in the strictest confidence. We abide by
the Market Research Society Code of Conduct, and your personal data will not be
passed on to any third parties.
If you agree for your survey response data to be used for market research purposes,
please click the 'Yes' below to begin the survey. If you would rather that your
responses are not used, then please select ‘No’.
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) Yes
2) No [SCREEN OUT]
ASK ALL
2) Which state or territory are you from?
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) New South Wales
2) Victoria
3) Queensland
4) South Australia
5) Western Australia
6) Tasmania
7) Australian Capital Territory
8) Northern Territory
9) Other Commonwealth Territory [SCREEN OUT]
10) Elsewhere [SCREEN OUT]
ASK ALL
3) And to ensure we have a representative sample; can you please enter your
postcode in the box below?
[OPEN TEXT RESPONSE]
ASK ALL
4) Are you eligible and enrolled to vote in Australian Federal and State Elections?
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) Yes - I am enrolled to vote
2) No - I am not enrolled to vote [SCREEN OUT]
3) Unsure [SCREEN OUT]
4) Refused [SCREEN OUT]
ASK ALL
5) Please indicate your gender
22
--- Page 23 ---
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) Male
2) Female
3) Prefer not to say
4) Other (please specify): __________
ASK ALL
6) Which age group do you fall into?
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) 17 or under [SCREEN OUT]
2) 18 – 19
3) 20 – 24
4) 25 – 29
5) 30 – 34
6) 35 – 39
7) 40 – 44
8) 45 – 49
9) 50 – 54
10) 55 – 59
11) 60 – 64
12) 65 – 69
13) 70 – 74
14) 75+
15) Prefer not to say
ASK ALL
7) Which one of the following best describes your current employment situation?
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) Employed full-time (more than 35 hours per week)
2) Employed part-time / casual (less than 35 hours per week)
3) Self Employed
4) Retired / Pensioner
5) Unemployed
6) Not in the Labour force
7) Home duties
8) Student
9) Other (please specify): _________
ASK ALL
8) What is the highest level of education you have received?
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) None - I have no qualifications
2) Certificate 1 or 2
3) Year 9 or below
4) Diploma
5) Certificate 3 or 4
6) Year 10 – 12
7) Bachelor’s degree
8) Master’s degree or higher
23
--- Page 24 ---
9) Other (please specify): __________
RANDOMLY ASSIGNED EITHER 9A, 9B, 9C OR 9D
9a) If a Federal Election were held today, which party would you give your first
preference vote to in the House of Representatives?
[SINGLE CODE, RESPONSE OPTIONS 1-15 IN RANDOM POSITION, ELSE FIXED]
1) Liberal Party [SHOW IN ALL OTHER THAN QLD + NT]
2) Liberal National Party [SHOW ONLY IN QLD]
3) Country Liberals [SHOW ONLY IN NT]
4) Australian Labor Party
5) The Greens
6) The Nationals [SHOW IN ALL OTHER THAN QLD + NT]
7) United Australia Party
8) An Independent Candidate
9) Pauline Hanson’s One Nation
10) Animal Justice Party
11) Christian Democratic Party
12) Katter's Australian Party
13) Anning Conservative National
14) Centre Alliance
15) Shooters, Fishers and Farmers
16) Other (please specify): __________
17) Unsure
18) Refuse
9b) If a Federal Election were held today, which party would you give your first
preference vote to in the House of Representatives?
[SINGLE CODE, RESPONSE OPTIONS 1-15 IN RANDOM POSITION, ELSE FIXED]
1) Liberal Party [SHOW IN ALL OTHER THAN QLD + NT]
2) Liberal National Party [SHOW ONLY IN QLD]
3) Country Liberals [SHOW ONLY IN NT]
4) Australian Labor Party
5) The Greens
6) The Nationals [SHOW IN ALL OTHER THAN QLD + NT]
7) United Australia Party
8) An Independent Candidate
9) Pauline Hanson’s One Nation
10) Animal Justice Party
11) Christian Democratic Party
12) Katter's Australian Party
13) Anning Conservative National
14) Centre Alliance
15) Shooters, Fishers and Farmers
16) The New Liberals
17) Other (please specify): __________
18) Unsure
19) Refuse
24
--- Page 25 ---
9c) If a Federal Election were held TODAY, which party would you give your first
preference vote to in the House of Representatives?
[SINGLE CODE, RESPONSE OPTIONS 2-16 IN RANDOM POSITION, ELSE FIXED]
1) The New Liberals
2) Australian Labor Party
3) The Greens
4) The Nationals [SHOW IN ALL OTHER THAN QLD + NT]
5) United Australia Party
6) An Independent Candidate
7) Pauline Hanson’s One Nation
8) Animal Justice Party
9) Christian Democratic Party
10) Katter's Australian Party
11) Anning Conservative National
12) Centre Alliance
13) Shooters, Fishers and Farmers
14) Liberal Party [SHOW IN ALL OTHER THAN QLD + NT]
15) Liberal National Party [SHOW ONLY IN QLD]
16) Country Liberals [SHOW ONLY IN NT]
17) Other (please specify): __________
18) Unsure
19) Refuse
9d) If a Federal Election were held TODAY, which party would you give your first
preference vote to in the House of Representatives?
[SINGLE CODE, RESPONSE OPTIONS 1-13 IN RANDOM POSITION, ELSE FIXED]
1) The New Liberals
2) Australian Labor Party
3) The Greens
4) The Nationals [SHOW IN ALL OTHER THAN QLD + NT]
5) United Australia Party
6) An Independent Candidate
7) Pauline Hanson’s One Nation
8) Animal Justice Party
9) Christian Democratic Party
10) Katter's Australian Party
11) Anning Conservative National
12) Centre Alliance
13) Shooters, Fishers and Farmers
14) Other (please specify): __________
15) Unsure
16) Refuse
ASK ALL
10) Under the preferential voting system, you must assign a preference to either the
Liberal Party/LNP or the Australian Labor Party. Regardless of WHERE you would
place them in your preference list, would you put the Liberal Party/LNP ahead of the
Labor Party or the Labor Party ahead of the Liberal Party/LNP?
[SINGLE CODE, RESPONSE OPTIONS 1-4 IN RANDOM POSITION, ELSE FIXED]
25
--- Page 26 ---
1) Liberal Party [SHOW IN ALL OTHER THAN QLD + NT]
2) Liberal National Party [SHOW ONLY IN QLD]
3) Country Liberal [SHOW ONLY IN NT]
4) Australian Labor Party
5) Unsure
6) Refused
ASK ALL
11) The New Liberals may stand a candidate in your area at the next Federal election.
From the list below, who do you think the current leader of the New Liberals is?
[SINGLE CODE, RESPONSE OPTIONS 1-7 IN RANDOM POSITION, ELSE FIXED]
1) Victor Kline
2) Scott Morrison
3) Anthony Albanese
4) Adam Bandt
5) Michael McCormack
6) Pauline Hanson
7) Clive Palmer
8) Other (please specify): __________
9) Unsure
10) Refused
ASK ALL
12) From the list of options below, please select the parties that you think form part
of the Liberal-National Coalition, commonly known as the Coalition?
Please select all those options that you feel apply
[MULTI CODE, RESPONSE OPTIONS 1-11 IN RANDOM POSITION, ELSE FIXED]
1) The Liberals
2) The LNP
3) The Nationals
4) The Country Liberals
5) The New Liberals
6) Labor
7) The Greens
8) Pauline Hanson’s One Nation
9) United Australia party
10) Katter's Australian party
11) Shooters Fishers and Farmers
12) Unsure
13) Refused
ASK ALL
13) Below is a list of political parties.
From the list below, please select those that you feel are connected, or have an
existing relationship with the Liberal Party?
Please select all that apply
26
--- Page 27 ---
[MULTI CODE, RESPONSE OPTIONS 1-11 IN RANDOM POSITION, ELSE FIXED]
1) The Liberal National Party
2) The Country Liberal Party
3) The National Party
4) The New Liberals
5) The Australian Labor Party
6) The Green Party
7) Pauline Hanson’s One Nation
8) The New Progressives
9) The Green Progressives
10) None
11) Unsure
12) Refused
ASK ALL
14) Do you think the Liberal Party and The New Liberals are both part of the same
party or different parties?
[SINGLE CODE, RESPONSE OPTIONS 1-2 IN RANDOM POSITION, ELSE FIXED]
1) Different parties
2) The same party
3) Unsure
4) Refused
ASK ALL
15) As far as you are aware, does a relationship exist between the Liberal Party and
the New Liberals?
[SINGLE CODE, RESPONSE OPTIONS 1-2 IN RANDOM POSITION, ELSE FIXED]
1) They have no existing relationship
2) They have an existing relationship
3) Something else
4) Unsure
5) Refused
ASK ALL
16) If a party standing at a Federal election, had the word ‘New’ in front of an existing
name in party politics (for example, New Labor, New Liberals, New Greens); in your
view, which of the following options best represents who the party is?
[SINGLE CODE, RESPONSE OPTIONS 1-5 IN RANDOM POSITION, ELSE FIXED]
1) A new political party entirely separate to existing parties with no relationship to
them
2) A separate party but has a connection or relationship with them
3) A national rebrand of an existing political party
4) Localised branding of an existing party
5) A separate division within an existing party
6) Other (please specify): __________
7) Unsure
8) Refused
27
--- Page 28 ---
ASK ALL
17) The New Liberals may stand a candidate in your area at the next Federal election.
From what you know, which of the following options best represents who you think
the New Liberals are?
[SINGLE CODE, RESPONSE OPTIONS 1-5 IN RANDOM POSITION, ELSE FIXED]
1) A new party entirely separate to the Liberal Party with no relationship to them
2) A separate party from the Liberal Party but has a connection or relationship with
them
3) A national rebrand of the Liberal Party
4) Localised branding of the Liberal Party
5) A separate division within the Liberal Party
6) Other (please specify): __________
7) Unsure
8) Refused
ASK ALL
18) Which of the following is closest to your view?
[SINGLE CODE, RESPONSE OPTIONS 1-2 IN RANDOM POSITION, ELSE FIXED]
1) The name ‘The New Liberals’ sounds like it is a completely separate organisation
from the 'Liberal Party' with no connection or relationship
2) The name ‘The New Liberals’ sounds like it is a rebrand of the 'Liberal Party', with
slightly changed priorities
3) Something else: ___________
4) Unsure
5) Refused
ASK ALL
19) Which of the following is closest to your view?
[SINGLE CODE, RESPONSE OPTIONS 1-2 IN RANDOM POSITION, ELSE FIXED]
1) The name ‘The New Liberals’ sounds like it is a completely separate organisation
from the ‘Liberal Party’ with no connection or relationship
2) The name ‘The New Liberals’ sounds like it is a division within the ‘Liberal Party’
3) Something else: ___________
4) Unsure
5) Refused
ASK ALL
20) Which of the following is closest to your view?
[SINGLE CODE, RESPONSE OPTIONS 1-2 IN RANDOM POSITION, ELSE FIXED]
1) The name ‘The New Liberals’ sounds like it is a completely separate organisation
from the ‘Liberal Party’ with no connection or relationship
2) The name ‘The New Liberals’ sounds like it is a new and fresh party that is
separate from the ‘Liberal Party’ but is connected and works closely with them
3) Something else: ___________
4) Unsure
5) Refused
ASK ALL
28
--- Page 29 ---
21) Which of the following is closest to your view?
[SINGLE CODE, RESPONSE OPTIONS 1-2 IN RANDOM POSITION, ELSE FIXED]
1) The name ‘The New Liberals’ sounds like it is completely separate from the
‘Liberal Party’
2) The name ‘The New Liberals’ sounds like it is connected in some way to the
‘Liberal Party’
3) Something else: ___________
4) Unsure
5) Refused
ASK ALL
22) Do you think that placing the term ‘New’ in front of an existing party name is
sufficient for voters to fully understand that this party has no relationship with an
existing party?
[SINGLE CODE, RESPONSE OPTIONS 1-2 IN RANDOM POSITION, ELSE FIXED]
1) Yes – it is sufficient
2) No – it is not sufficient
3) Unsure
4) Refused
ASK ALL
23) Do you think that placing the term ‘New’ in front of an existing party name is
sufficient for voters to fully understand that this party is a completely separate
entity from an existing political party?
[SINGLE CODE, RESPONSE OPTIONS 1-2 IN RANDOM POSITION, ELSE FIXED]
1) Yes – it is sufficient
2) No – it is not sufficient
3) Unsure
4) Refused
ASK THOSE WHO SELECTED ‘THE NEW LIBERALS’ AT 9A, 9B, 9C OR 9D.
24) Earlier you said that you would most likely vote for The New Liberals.
The New Liberals are a new political party led by Victor Kline, which is seeking
registration, with no connection or relationship to the Liberal Party currently in
government.
Having been given this information, can you confirm that you intended to vote for
The New Liberals earlier, or did you intend to vote for another party?
I meant to vote for...
[SINGLE CODE, RESPONSE OPTIONS 2-8 IN RANDOM POSITION, ELSE FIXED]
1) The New Liberals, led by Victor Kline, currently seeking registration
2) The Liberal Party, currently led by Scott Morrison, currently in government and
established in 1944
3) The Liberal National Party of Queensland, established in 2008
4) The Country Liberal Party, established in 1974
29
--- Page 30 ---
5) The Australian Labor Party, currently led by Anthony Albanese, currently in
opposition and established in 1901
6) The Australian Greens, currently led by Adam Bandt, established in 1992
7) Pauline Hanson’s One Nation Party, led by Pauline Hanson, established in 1997
8) The National Party, led by Michael McCormack, established in 1920
9) Other (please specify): __________
10) Unsure
11) Refused
ASK ALL
25) If they were on the same ballot paper, would you find it clear or confusing, to
correctly identify the party you wish to vote for when choosing between ‘The New
Liberals’ or the ‘Liberals’?
[SINGLE CODE, RESPONSE OPTIONS 1-4 RANDOMLY FLIP, ELSE FIXED]
1) Very confusing
2) Fairly confusing
3) Fairly clear
4) Very clear
5) Unsure
6) Refused
ASK ALL
26) Can you say whether you agree, or disagree with the following statements?
[SINGLE CODE, RANDOM POSITION OF STATEMENTS A-F]
Statement Text
A
B
C
D
E
F
The names ‘Liberal’ and ‘The New Liberals’ very clearly tell me that they
are entirely separate organisations from one another
The Liberal Party and The New Liberals have names that are too similar
for me to be sure who I am voting for
It sounds like there is no difference between the Liberal Party and The
New Liberals
By voting for The New Liberals, I am helping to elect a Liberal
government
A vote for The New Liberals goes to the Liberal Party
‘The New Liberals’ and ‘Liberal Party’ sound like they have an existing
relationship and are connected to one another in some way
[RESPONSE OPTIONS, SCALE RANDOMLY FLIP 1-5]
1) Strongly agree
2) Somewhat agree
30
--- Page 31 ---
3) Neither agree, nor disagree
4) Somewhat disagree
5) Strongly disagree
6) Unsure
7) Refused
ASK ALL
27) Can you say whether you agree, or disagree with the following statements?
[SINGLE CODE, RANDOM POSITION OF STATEMENTS A-E]
Statement Text
A
B
C
D
E
If both a 'New Liberals' candidate, and a 'Liberal' candidate stood in my
electorate, I might feel confused about which party to preference above
the other
The Liberals and The New Liberals sound very similar to me
If candidates from both the Liberals and The New Liberals stood in my
electorate at the next election, I would find the voting experience much
more difficult
The New Liberals is just a marketing exercise from the Liberals to get
more votes
If a Party uses the name The New Liberals, it will confuse voters into
voting for a party they do not intend to vote for
[RESPONSE OPTIONS, SCALE RANDOMLY FLIP 1-5]
1) Strongly agree
2) Somewhat agree
3) Neither agree, nor disagree
4) Somewhat disagree
5) Strongly disagree
6) Unsure
7) Refused
ASK ALL
28) Thinking back to the 2019 Federal Election, who did you vote for in the house of
representatives?
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) Liberal Party or LNP or Country Liberals
2) Labor Party
3) Greens Party
4) National Party [SHOW IN ALL OTHER THAN QLD + NT]
5) Pauline Hanson's One Nation Party
6) United Australia Party
7) An independent
31
--- Page 32 ---
8) Other (please specify): __________
9) Unsure
10) Refused
11) I did not vote
ASK ALL
29) Which of the following best describes your living situation?
[SINGLE CODE, RESPONSE OPTIONS FIXED POSITION]
1) Own home with a mortgage
2) Own home outright
3) Private Renting
4) Social Housing
5) Other (please specify): __________
6) Refused
32
--- Page 33 ---
--- Page 34 ---
Appendix 4 – Sample House of Representatives Ballot
Paper
Figure 1: Sample House of Representatives Ballot Paper
This is a sample ballot paper taken from the Australian Electoral Commission’s
website38. Please note that on the ballot paper, candidate and therefore party names,
are placed in a column, with positions being determined by random lot39. It demonstrates
that there can be substantial distance between candidate and party names.
38 https://www.aec.gov.au/Voting/How_to_vote/practice/practice-house-of-reps.htm
39
https://www.aph.gov.au/About_Parliament/Parliamentary_Departments/Parliamentary_Library/pu
bs/rp/RP0708/08rp05#preferential
34
--- Page 35 ---
--- Page 36 ---
--- Page 37 ---
--- Page 38 ---
--- Page 39 ---
--- Page 40 ---
--- Page 41 ---
--- Page 42 ---
--- Page 43 ---
--- Page 44 ---
--- Page 45 ---
--- Page 46 ---
--- Page 47 ---
--- Page 48 ---
--- Page 49 ---
--- Page 50 ---
--- Page 51 ---
Appendix 6 – Research Report
51
--- Page 52 ---
52
--- Page 53 ---
53
--- Page 54 ---
54
--- Page 55 ---
55
--- Page 56 ---
56
--- Page 57 ---
57
--- Page 58 ---
58
--- Page 59 ---
59
--- Page 60 ---
60
--- Page 61 ---
61
--- Page 62 ---
62
--- Page 63 ---
63
--- Page 64 ---
64
--- Page 65 ---
65
--- Page 66 ---
66
--- Page 67 ---
67
--- Page 68 ---
68
--- Page 69 ---
69
--- Page 70 ---
70
--- Page 71 ---
71
--- Page 72 ---
72
--- Page 73 ---
73
--- Page 74 ---
74
--- Page 75 ---
75
--- Page 76 ---
76
--- Page 77 ---
77
--- Page 78 ---
78
--- Page 79 ---
79
--- Page 80 ---
80
--- Page 81 ---
81
--- Page 82 ---
82
--- Page 83 ---
83
--- Page 84 ---
84
--- Page 85 ---
85
--- Page 86 ---
--- Page 87 ---
--- Page 88 ---
88
--- Page 89 ---
89
--- Page 90 ---
--- Page 91 ---
91
--- Page 92 ---
--- Page 93 ---
93
--- Page 94 ---
94
--- Page 95 ---
95
--- Page 96 ---
96
--- Page 97 ---
97
--- Page 98 ---
98
--- Page 99 ---
99
--- Page 100 ---
100
--- Page 101 ---
101
--- Page 102 ---
102
--- Page 103 ---
103
--- Page 104 ---
104
--- Page 105 ---
105
--- Page 106 ---
106
--- Page 107 ---
107
--- Page 108 ---
108
--- Page 109 ---
109
--- Page 110 ---
110