Australian Federal Election 18th of May 2019

[Smid]

I see according to polls, that Morrison is way more popular than Shorten in terms of popularity and PM preference. Could that hurt the ALP numbers in the next few weeks, or is the Coalition brand so damaged that Morrison popularity will not matter?

It's very uncommon for an incumbent to not be leading on the question of preferred PM. Of all typical poll questions, this one is perhaps the least useful.

[Smid]

I see according to polls, that Morrison is way more popular than Shorten in terms of popularity and PM preference. Could that hurt the ALP numbers in the next few weeks, or is the Coalition brand so damaged that Morrison popularity will not matter?

Preferred PM is basically meaningless in Australia, and it has been calculated by Kevin Bonham that it normally has around a 15% bias to the current prime minister when up against an opposition leader who has never been in that position.

I haven't seen that particular work of Dr Bonham, however I find that estimate entirely credible.

[Smid]

What policy differences exist between the United Australian Party, Katter's Australian Party, and One Nation? I know they broadly compete for a similar base of voters.

Well they are all wrapped up in the personae of their leaders, so might as well go over them.

Bob Katter is a longtime fixture of Queensland politics, having inherited his rural Northern seat of Kennedy from his father after serving in the rather infamous Queensland cabinet of Joh Bjelke Peterson (it's very hard to get Queensland conservatism without a look at this guy).

You might also wish to add that his son, Robbie Katter, is the Katter Party leader in the Queensland Parliament, representing a state electorate that is largely (entirely?) contained within his father's federal seat.

[Smid]

What time do polls close? I want to watch the ABC livestream, for whatever reason I find Australian newscaster accents very pleasant and soothing, and I love the swing graphics and music they use

Polls close at 6pm across the country, but nothing is reported until polls have closed in Western Australia - 6pm AWST is apparently 6am EST.

Two interesting things:
1) The fact that the time zones don't correspond precisely to state or electorate boundaries always causes a bit of chaos on election days.
2) Although no results are declared from the eastern states before 6pm AWST, the Electoral Commission does announce that they will be running the election night preference counts between particular candidates before polls close in WA. Usually this will be the Coalition and Labor candidates, but if it turns out that a third-party candidate has surprisingly reached the top two, the actual preference count will be redone and the initial one won't count. Clive Palmer took the AEC to court over this on the grounds that if WA voters knew that he wasn't going to win a seat in the lower house in Queensland, they might not vote for his party in WA - but this argument was rightly rejected.

Australian timezones match state boundaries.

They only block coverage of results in places where the polls are yet to close. Assuming there are no regional blocks on the ABC webcasting, you should be able to catch Antony Green from 6pm Australian Eastern Standard Time (just under 24 hours from the timestamp on this post).

[Smid]

However, Broken Hill (in NSW) does use Central Time.

Oh, really? I was not aware of that. Actually, while on the topic of timezones in Australia, I remember seeing something about a town on the Nullarbor that has its own time zone, although it isn't officially recognised. Here's the article.

[Smid]

Does anybody want to compile the Sportsbet seat odds (plus other interesting odds, like highest/lowest turnout, overall stay numbers, etc? I did last time to see how well the odds compared to the outcomes, but I'm tied up at the moment. If anyone is interested, the odds can be found here: https://www.sportsbet.com.au/betting/politics/australian-federal-politics

Interested in each seat, if we can compile them.

[Smid]

Have they counted second choices or just first.  I would think most Greens would have Labor as second so that should close the gap a bit but probably not enough to win.  At this point either L/NP majority or hung parliament, Labor majority seems next to impossible even if votes counted break their way.  So strong likelihood Morrison stays on as PM.

Yeah all the projections take into account the 2 party preferred

  :Have they actually counted second preferences or just projecting based on polls?  Agree Labor realistically has no path to victory, but just curious as polls might be closer once counted if they haven't already.

I've scrutineered almost every state and federal election where I've lived in the past twenty years (plus a few council ones, plus a few by-elections, plus a couple of interstate elections where I've travelled, minus a couple that I didn't do because I was required elsewhere), so I can probably provide some insight into this question.

First they empty the ballot boxes on the table and just unfold the ballot papers and sort of even them up in bundles (not bundles based on votes or anything, merely bundles of unfolded papers).

Then they put out the names of the candidates on folded sheets of paper (folded to stand upright, like a triangle), and they sort them into primary votes for each candidate, checking for formality/informality. Then they count those primary votes and report the results back to the Electoral Commission. The scrutineers call or text those results also back to party CHQ.

Then they take the votes of the minor candidates (or maybe a major candidate in there - the Commission decides in advance which two candidates they think will be the top two after the allocation of preferences, and this can change during the recount, but tonight's count is only a preliminary count). They allocate these votes to the two chosen candidates and provide a count "Greens to Labor - 220 votes", "Greens to Liberal - 90 votes", etc. This gives a TCP figure that gets phoned back to the Electoral Commission and the scrutineers also phone or text back the results to their CHQ.

Before calling in the primary votes, they probably do a tally to ensure that the number of votes counted matches the number of votes issued - if not, they might look in the bins or accidentally put in the Senate box, but if it's just one or two missing, everyone probably agrees that it's close enough.

When you see the election night coverage, when Antony Green gives 2CP figures, that's based on preference flows at the booths that have reported 2CP figures, not based on polling. If you look at the results page on the AEC website, you can see what time each booth in an electorate reported its primary vote and 2CP vote: https://tallyroom.aec.gov.au/HouseDivisionPage-24310-302.htm (scroll down). A couple are about a minute apart - I suspect in those booths they completed both the primary and 2CP counts and then phoned both results in simultaneously (or maybe they do it by laptop, I don't know, I'm normally on the phone back to CHQ while the Commission staff are doing that thing).

[Smid]

If you think that it is a simple case of pollsters always over/underpolling one side, deliberately or not, then it is because you have no idea of how polling actually works (in part because it is hardly in the pollsters own interests to always get it wrong - their own financial viability relies on actually being vaguely credible).

It's not just one poll. It's *every* poll for the last five years. There's a difference. Systematic bias on the part of pollsters is the best explanation given the amount of error and the consistent error of +1, +2 for Labor.

One problem Australia has and maybe this is the reason is for preferential votes, you can rank them yourselves or go by party rankings.  UAP put L/NP ahead of Labor in rankings and it seems biggest errors were areas they were strongest so that is perhaps one possible explanation whereas pollsters go on the assumption every voter will rank individually never mind parties only publish their rankings close to e-day so would only work in final polls if you ask people will you use party or individual ranking, not further out.

I think the bigger problem in the polling was in the primary vote. Antony Green mentioned a few times that the primary vote difference between the Coalition and Labor was about 8 percentage points, rather than the two or three that had been referenced in the polls.

Generally speaking, preference flows from minor parties tends to be relatively consistent from election to election. I have that data lying around somewhere; if you're interested, I'll try to dig it up. In other words, despite a proportion not following the HTV cards, they're fairly similar each election and can therefore be modelled accordingly (Greens is about 70% - 80% to Labor, minor conservative parties tend to be about 60% to the Coalition, etc) - these are just rules of thumb, but tend to serve well when scrutineering or looking at primary-only vote counts.

[Smid]

As always, bigger maps in the gallery, etc. Bigger versions still available upon request. Plotted using R, and relying specifically the magrittr, here, glue, tidyverse, ggvoronoi, rgeos, maptools, ggmap, broom, rgdal, sf and ggspatial packages.

These booth results only show booth on-the-day results. The preliminary results used were downloaded on Sunday morning AEST and in some cases may have been superseded by recounts and rechecks. The maps will be updated following the declaration of results.

There is no requirement to vote at a particular polling place, however many Australians vote at their nearest polling place. These maps use Voronoi polygons to make shades of colour more obvious, however this may give the false appearance of a specified booth catchment. Some booths might be small but appear large on the map. Others may spill across electorate boundaries despite not being joint booths for both electorates.

Joint booths have 2CP votes summed for all electorates voting at the booth. In cases where different parties are included in the 2CP count, this can distort the results and potentially result in a 2CP figure of less than 50%. Take, for example, a joint booth where an equal number of votes are cast from both electorates, and where one electorate voted 60-40 Liberal vs Labor and the other electorate voted 60-40 Nationals vs Labor. In this hypothetical example, the booth would appear to be 40-30-30 Labor vs Liberal vs Nationals and thus be shaded a pale red. Please be aware of this while considering the maps. Joint booths including Independents, Greens or other minor parties may be affected in this manner. One clear example is the "Melbourne" booth, which was a joint polling booth for all(?) Victorian electorates on election day. It's possible I'll update how some joint booths are handled before uploading final results, but for now, that's how it is.

2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth




2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth - Greater Brisbane


The extremely strong Labor booth in NW Brisbane is Mount Nebo.



2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth - Wider South-East Queensland


This map ranges from about Noosa in the North, to the Tweed and Queensland border in the South, and out to Toowoomba in the West (the cluster of LNP booths near the left edge of the map).



2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth - Greater Sydney


Take note of the interesting North-South divide in Wentworth, between the Liberal-voting North/harbourside, and the independent-voting South.



2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth - Coastal New South Wales


This map ranges from Newcastle in the North, to Woolloongong in the South. I believe the town on the Western edge of the map is Bathurst.



2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth - Greater Melbourne


This map includes Geelong and Phillip Island, as well as areas considered Metropolitan Melbourne. Interesting to note that the Greens don't appear to have won a single booth in Cooper (the old Batman) for the first time in... I think a decade...



2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth - Hobart and Surrounds


Compare Wilkie's results North and South of the Derwent.




2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth - Greater Adelaide





2019 Australian Federal Election Two Candidate Preferred Results (Preliminary Counts) by Booth - Greater Perth