In this blog post I’ll be analysing how much people fill out their ballot paper, and how it differs between parties and based on how big the ballot paper is.
For the next three blog posts, I will be using the full preference data published by the NSW Electoral Commission. Starting in 2015, the NSWEC now data enters every lower house ballot. This allows them to not only publish the official two-candidate-preferred count, distribution of preferences and a two-party-preferred count between Labor and the Coalition, but also publish 2CP counts between every pair of candidates. Wanna know the 2CP between the Greens and One Nation? Now you can. They also publish the raw ballot data for each electorate – every preference marking for each ballot.
While it is not uncommon for this data to be published for upper house elections like the Senate or the Legislative Council, this is the only time this has ever been done for a single-member election, so there’s a lot to be unpacked.
Today’s post is simple: how completely do people fill out their ballot? In NSW you are not required to number every box. A single one is sufficient as long as it’s clear who it’s for.
Firstly, the overall completion rate is very low. For all formal ballots, 64.3% of votes were just a ‘1’ only, with only 23.7% completing their ballot. But there is substantial variation by party.
One thing that is clear is that progressive voters are more likely to mark preferences, as we saw in the Legislative Council. The Greens are a clear standout, but Keep Sydney Open, Sustainable Australia and Keep Sydney Open have stronger preference flows than comparable right wing parties, although the Christian Democrats violate that pattern.
Generally the major parties have very low rates of completion but Labor still outperforms the Liberals and Nationals significantly.
Secondly, I wanted to examine whether these trends differ based on the size of the ballot paper. The big surge in exhausted votes (from around 30% of potential preferences to around 50%) occured in 1999, when the size of ballot papers doubled. That record has never been beaten, but ballot papers have not returned to pre-1999 sizes.
While there is a trend, it mostly seems to affect the rates of partial and complete ballot papers. The rate of '1 only' votes is still well over 60% for even the smallest ballot papers. It should be noted that over three quarters of electorates at the 2015 and 2019 elections had 5-7 candidates, so the other categories cover a relatively small number of cases. But the rate of voters casting a complete ballot drops from over 30% for a ballot of 3 candidates to barely 10% for a ballot of 10 candidates.
Finally I wanted to consider trends over time. Unfortunately I only have two elections worth of data. I did look at the rate of voters casting a '1 only' vote at upper house elections and did find a drop in 2019.
| Election | 1 only | Partial | Complete |
| 2015 | 65.14 | 9.71 | 25.15 |
| 2019 | 64.25 | 12.00 | 23.74 |
There is a little bit of evidence of voters being less likely to cast a 1 only ballot in 2019 compared to 2015, but there's also fewer voters casting a complete ballot, with the 'partial' category growing substantially.
That's it for today - tomorrow I'll look at two-party-preferred preference flows in 2015 and 2019 based on this richer data, and later this week I will be analysing Labor-Greens reciprocal preference flows.
The greens are very well trained with preference flow.
Interesting, Ben, but really the number of those who do a “complete” ballot isn’t terribly relevant. One can have fun ordering all the no-hopers in reverse order of ratbagginess, but that’s only for the small emotional release that it gives. Preferences to all, or all but one, of those with a realistic chance of winning is all that matters, so that the vote will stay in the count and eventually express a preference for one of the last two candidates left standing. .
This is interesting stuff
I have done the same sorts of analyses for Manly (2017 by-election and 2019 election) and Manly&Curl Curl for the NBC Council elections in 2017 and 2021.
The Greens and ALP seem to be the only groups that offer a full HTV, with the ALP voters being more likely to follow it.
Just Vote 1 was the predominant pattern, averaging 56% of all ballot papers in the six contests – even GRN ballot papers averaged 44% JV1
“Following the ticket” figures were:
GRN: 11.2% of 16.1%;
ALP: 26.2% of 12.4%;
Your Northern Beaches: 10.7% of 27.3% and;
LIB: 68.8% of 37.6%.
I did a quick analysis on the previous NSW election last week:
The candidate with the highest primaries wins 97% of the time;
The second-place candidate wins 3% of the time
Third-placed candidates never win.
I have told this to some of the “Tealish” candidates in my neck of the woods “Northern Beaches”. They didn’t like it.
No they aren’t.
ALP is better
while they dont win from 3rd yet it is not impossible. esoecially when a vote is split lets say labor polls 30% primary and liberal polls 25% and nationals 24% in a conservative electoarate the other 20% can theorectically push the nationals into second and then they would gain the liberal votes. this would happen with any split vote. im guessing this may well happen in port macquarie at the nsw state election.
The parties best able to direct preferences are Labor and the Greens. On port Macquarie the liberals and nats will obviously preference each other. In Rough figures I expect an alp primary vote to be round 35% the coalition maybe 33 to 35% and a green vote of maybe 10% that means roughly 20% with everyone else most of this 20% will largely exhaust. Under opv where you need to lead in the primary vote to win.. this complicates matters. The exceptions are where in cases with both high alp and green votes. Now the liberals are at a big disadvantage where they get no help from opv. Labor and the Greens will obviously be trying to ensure more 1 to votes or in seats like Pittwater will encourage 1 Labor or green and 2 teal or independent
Interesting topic. You will recall Killen retained Moreton (therefore Menzie’s government) in 1961 on getting a leakage of CPA preferences. But of course the CPA candidate was on top of the ticket and the leakage was the donkeys. Which is just a suggestion to check donkeys in lower house seats where it matters in your analysis. Donkeys can matter and you always want them if you can get them. Labor won the Eden-Monaro by-election in 2021 on donkeys and well the rest (as Albo would say) is history. Another habit of labor voters (not greens) is to tactically vote and not bother to go the long way. Preferences also can make working for minor parties handing out worthwhile but I have noticed each year more voters use Mr google in the booth to find preferences even when no how to vote is available.
@assistan minister thats because they know their vote wont stay with the greens if they only put a 1. 9/10 liberal labor and national will finish in the top 2 hence they dont need to bother
At the risk of stating the obvious, this shows there are 3 types of voter in NSW. The majority are rusted on major party voters and vote for their party and see no need or desire to give any preferences. Then there are those who fill every box, whether because they are making a silent and unheard statement or just out of sheer contrariness. The rest would pick their preferred candidate and then maybe 1 or 2 preferences, ending with one of the two major parties. The last being those who vote for independents or minor parties but realise their candidate won’t win and end their preferences with the least objectionable of the other candidates. I suspect most politically aware voters (including readers of this blog) would fall into the third group.
It would be interesting to see a comparison, over the past elections, of the % of voters who only vote for one candidate (ie no preferences) vs. the % who include at least one preference, both statewide and also by electorate.
David, the final table in this post answers your last question. It has the proportion of voters who just marked a single box (1 only) as well as those who filled it out partially or completely. Add those two together and you get the proportion who gave some preferences. Unfortunately we only have this data since 2015.
@mick are you serious? labor will be lucky to et that on 2pp vote. so your saying sustainable australia and informed medical options will poll 20% of the vote? . the combined coalition vote will be at least 65%. minimum. depending on how it splits it may be lib vs nats on 2pp. labor will not get a 20% swing. dont take this the wrong way but where do you get these fanciful ideas ?
Ben you missed something somewhere. I was talking about nsw as a whole not port Macquarie. Port Macquarie will have a tight flow of preferences between libs and nats. In the Abscene of a credible independent it will be coalition held. The 20% remaining will be everyone else onp uap ksa sff etc not alp coalition or greens. Pls read what I wrote again.
I agree with Mick
UAP & KSO no longer registered fyi
Actually now reading it again the other % is far too inflated for a state election. But agree with majors in the mid to high 30s and Greens at 10%
I was trying to point out where there were possible tight flows of Preferences
Labor to green and visa versa
Labor and green to a ind eg teal
Libs and nats to each other in the Port Macquarie electorate
@ mick you should probly state that clearly. Yes it will be about 35% a piece with 10% to greens and about 7% to onp. The rest will be others. The low primary vote is why it will be a hung parliament. Labor requires a swing of 10% and they’re looking at half that.there is a small pathway for liberal majority but everything will need to go right and it won’t. They need to hold all there current seats and take Murray,barwon or bega. Although there are a few seats that could swing to them outside of those. Time will tell
@daniel 7 of that 20 is onp so 20 is correct
One would think that One Nation voters would much prefer a Coalition government than a Labor one. Obviously they are sufficiently disgruntled with both majors that they won’t give either a preference.
My reasoning is that most votes for Greens have preferences because they mainly draw their support from traditional Labor heartland as well as affluent, small l-liberal electorates and their voters most likely still have a favourite major party. I could say the same about AJP an KSO voters. Christian Democrat voters will preference the LNP as their prefered major party.
One Nation tends to present themselves to traditionally LNP and Labor seats as alternatives for voters “fed up with the major parties”. This attracts an apolitical/apathetic crowd. Voters may have swung from Labor or Liberal and can’t bring themselves to vote for the major party that they used to vote for.
Comments are closed.