Thank you to everyone who participated in the first wave of the Voter Choice Project. For each wave we will be releasing selected results of the more interesting questions. The following are selected highlights from Wave 1 of the main (federal) study; some highlights from the by-election study will be released in a few days.
First: a little about the sample. 1390 people from across the country have registered to take part in the Voter Choice Project so far, 801 completed the Federal questionnaire. Every electorate with the exception of Watson (the Lakemba area in Sydney) has someone in the panel. Currently Greens and Independent voters are significantly over-represented and Liberal/National Coalition voters are significantly under-represented. We also need more women under 40. Recruitment continues throughout July – sign up here – and we hope that with increased numbers the panel will become more balanced, although we are weighting all data by age, gender and reported 2016 vote. All voters are encouraged to sign up as we would like to more than double the sample for the July wave – new participants will get a catch up questionnaire which includes the baseline questions from June as well as the July questions.
Wave 1 was a baseline questionnaire which asked a lot of questions about who you are and your interest in politics. We also asked who you voted for in 2016. The vast majority of respondents were happy with how they voted in 2016, and would not have changed their vote if they had known how close the result would be.
Do not read too much in to 5.48% percent of respondents wishing they had voted differently. Yes, if we voted nationally that would change the result of the election. However, we do not vote nationally on anything (except marriage equality in a non-compulsory survey). There is also movement both ways, something we shall explore more later in the Voter Choice Project.
One of the features of the Voter Choice Project is a lot of open text boxes that allow respondents to say whatever they want, rather than forcing them into options. So when we asked what issues or other factors were important to you in deciding your vote, we got a list of over 2000 issues and other factors like who the candidate is, honesty, and vision for the nation. Here’s a word cloud of the top issues raised in those responses.
Issues word cloud | Wave 1
You’ll note there are no words there indicating sentiment. We’ll do that a lot, you may as well get used to it. For example, if someone is pro- or anti- refugees, it will just be reported as refugees was an important issue in determining their vote. After the election the full reports will reveal some more nuance; before the election we’ll be as neutral as possible so as not to influence anyone’s votes, nor any candidate or party’s campaigning.
But what about the 2PP?
For the record, the whole concept of two party preferred polling is the very opposite of what the Voter Choice Project is about. 2PP conceals change and movement in voters’ intention to make it simple for people to understand (if you’re a political science student, look up Murray Goot, Joan Rydon and particularly Henry Mayer’s criticism of the Mackerras Pendulum and two party preferred methodology for some excellent examples of how to write colourfully in academic prose while really taking something to task – refs at the end of the post.) The Voter Choice Project is all about revealing change in voters’ choices. However, as the panel is not large enough to do what I’d like to do, which is individual tracking polls on all 151 electorates and then report a seat win out of 151 seats (we’d need 1200ish voters from each electorate, or 181,200 participants to do that), here you go, beginning with first preference vote intent.
Yes, we are aware our numbers are slightly different. Our entire methodology is different. The question asked is not the typical ‘if the election were held today’ but ‘who do you think you will vote for in the upcoming election’. Participants can put in any party they like, or none, and we don’t remove the don’t know’s and deliberately voting informals.
Independent voters is likely high because it is both offered as an option, there is a significant number or rural voters in the panel, and the concept of the study is appealing to independent, swinging and undecided voters. Also a note: a number of voters wrote Pauline Hanson’s One Nation/Katter’s Australia Party/Australian Conservatives (some switching Katter for Australian Liberty Alliance or Liberal Democrats) as though it was a single entity, and some commented that they wanted a coalition of these parties. These respondents were distributed evenly among the parties for the purposes of creating these results.
For the preference allocation we are also using a slightly different methodology to others, using neither a direct preferences question, nor last election preferences. We do have a direct preferences question that we will ask closer to the election, but in testing (and reinforced by the comments received in this wave) we found that a significant number of people do not understand how preferencing works. Some think you’re asking who did they put last, some think you’re asking who did they put second, and others believe that each preference gets a proportion of the vote.
So we asked ‘who do you want to win’ and when necessary combined that with other qualitative analysis, particularly ‘what do you think will change if Labor wins Government’, to infer a likely preference. For example, if they put they wanted the Greens to win government, and answered that if Labor wins the country will be doomed, we inferred a likely Liberal National Coalition preference. Similarly if they said they wanted no one to win because they all suck, and identified that they wanted the corporate tax cuts to be stopped, more funding for hospitals, and penalty rates to be restored, then we inferred a likely Labor Party preference. This should be probably be called an inferred two party preferred vote. It’s a lot more work; we’ll find out if it’s worth it come election day.
Wave 2 of the Voter Choice Project will be sent out to the main study participants on July 15, and by-election study participants on July 18. You can still sign up to join the study (until July 20 to be included in the July round). Those in the by-election seats can no longer join the by-election study, but you can join the main study. The July wave includes questions about what kind of candidates you want to vote for, and an optional set of questions on Pauline Hanson and One Nation.
Goot, M. 1973, ‘Growth and Decay: Party Support between Elections, 1966-70’, Australian Journal of Political Science, vol. 8, no. 1, pp. 265-271. https://doi.org/10.1080/00323267308401368
Goot, M. 2016, ‘The Transformation of Australian Electoral Analysis: The Two‐Party Preferred Vote—Origins, Impacts, and Critics’, Australian Journal of Politics & History, vol. 62, no. 1, pp. 59-86. https://onlinelibrary.wiley.com/doi/abs/10.1111/ajph.12208
Mayer, H. 1980, ‘Big Party Chauvinism and Minor Party Romanticism’, in H. Mayer (ed.), Australian Politics: A Fifth Reader, Longman Cheshire, Melbourne, pp. 345-360 (No link? Oh noes! YOU’LL HAVE TO GO TO A LIBRARY! It’s worth it.)
Rydon, J. 1973, ‘Swings and Predictions: The Analysis of Australian Electoral Statistics’, Australian Journal of Political Science, vol. 8, no. 1, pp. 259-264. https://doi.org/10.1080/00323267308401367