Alternate Siena Models

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Wednesday September 26, 2018 · 12:09 AM PDT

2018/09/26 · 00:09

I’m not unskewing anything!

Just to put that right up front. I’m also going to assume everyone is good with how polls are conducted (call tens of thousands of people, get a few hundred responses, weigh those responses to match the likely electorate, publish).

However! There are multiple ways to model a predicted electorate. And Siena, who are polling 100 house races this cycle and have completed 24, have been kind enough to provide us with a variety of those models in addition to their topline numbers. These models all reflect different assumptions about who is going to turn out for the election and Siena could have easily chosen any of them to be their toplines.

However, Siena have gone with a model that weighs their respondents against voter file information (age, party registration, gender, race, region, education) and to balance their likelyhood to vote between their self-reported intention to vote and their historical voting practice. Which I actually think is probably the best model to use. But plenty of very good pollsters with long track records of accurate results use other models, similar to the ones I’ll discuss below.

In the interests of not flooding you with spreadsheets of data I’ve put results for each seat into five bins; Likely Dem (D+10 or more), Lean Dem (D+3-D+9), Toss-Up (D+2-R+2), Lean R (R+3-R+9), Likely R (R+10 or more). I’ve also attached a picture of a summary spreadsheet at the end of the diary in case anyone want to look in a bit more detail.

Siena (Published toplines; uses age, party registration (or predicted party/primary vote), gender, race, region, education from L2 voter files and balances self-reported intention to vote and historical voting practice)

Likely D; CA-49, CO-6, IA-1 (3)

Lean D; MN-3, CA-45, KS-3, PA-7 (4)

Tossup; CA-48, KS-2, NM-2, MN-8, KY-6, IL-12, IL-6, NJ-7, TX-32, CA-25 (10)

Lean R; TX-7, FL-26, VA-7, ME-2, WI-1, TX-23, WV-3 (7)

Likely R; (0)

Ochre (Siena, but doesn’t weigh by party registration, predicted party, or primary vote)

Likely D; CA-48, MN-3, CO-6, IA-1, CA-49 (5)

Lean D; TX-32, TX-7, CA-45, IL-6, KS-2, KS-3, PA-7 (7)

Tossup; NJ-7, KY-6, MN-8, IL-12, CA-25, VA-7 (6)

Lean R; ME-2, NW-2, WI-1, TX-23 (4)

Likely R; FL-26, WV-3 (2)

Umber (Siena, but weighs using Census data instead of voter files)

Likely D; TX-32, MN-3, TX-7, CO-6, IA-1, CA-49 (6)

Lean D; KY-6, IL-6, KS-2, PA-7, CA-48, KS-3 (6)

Tossup; CA-45, MN-8, CA-45, NJ-7, IL-12, ME-2 (6)

Lean R; VA-7, NM-2, WI-1, FL-26 (4)

Likely R; WV-3, TX-23 (2)

Vermilion (Siena, but doesn’t weigh by education)

Likely D; CA-49, MN-3, CO-6, IA-1 (4)

Lean D; MN-8, NM-2, KS-2, CA-45, KS-3, PA-7 (6)

Tossup; CA-48, IL-6, KY-6, CA-25, IL-12, NJ-7, TX-32, FL-26 (8)

Lean R; TX-7, VA-7, ME-2, WV-3, TX-23, WI-1 (6)

Likely R;

Ultramarine (Siena, but only includes voters who self report as certain to vote)

Likely D; KS-3, MN-3, CA-45, CO-6, PA-7, IA-1 (6)

Lean D; CA-45, KY-6, IL-6, MN-8, NJ-7, CA-48, CA-49, TX-7 (8)

Tossup; KS-2, IL-12, TX-23 (3)

Lean R; WV-3, NM-2, WI-1, TX-32, FL-26 (5)

Likely R; VA-7, ME-2 (2)

As you’ve noticed the alternative models are all more bullish on Democratic chances than the Siena toplines. Which isn’t to say the are more likely to be correct. But it is to say say that they aren’t particularly more likely to be incorrect. If a equally respected polling firm other than Siena was running the exact same polls and got the exact same responses the published toplines could easily look exactly like the Ochre results, or the Umber, or the Ultramarine.

Take-home message 1; Polls aren’t particularly accurate in tight elections because weighing can easily shift a topline 5-10 points. So don’t sweat 5 point differences in a single poll.

Take-home message 2; The Siena house polls look really good for Democrats and the toplines are about as pessimistic as it is possible to be without doing something unreasonable like assuming 2014 turnout.

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday February 10, 2017 · 5:32 AM PST

2017/02/10 · 05:32

Western Australia is a physically vast (larger than Texas and Alaska combined) Australian state unsurprisingly located in the Western part of the Australian continent. Three-quarters of its 2.6 million people live in the capital city of Perth. Only 8% of the state’s population live outside the south-western corner.

On the 11th of March, the entirety of the Legislative Assembly (State House) and Legislative Council (State Senate) is up for election. Australian Governors are not elected and have only ceremonial power under most circumstances.

Currently, the Liberal Party and National Party of Western Australia control both houses of parliament after a comfortable reelection in 2013. The Liberal party is basically a conservative centre-right party representing the interests of business. The Nationals represent primarily rural electorates and have a focus on the interests of agriculture.

However recent polling (Newspoll, Labor 54%-46%) suggests that the Labor Party, the main centre-left party, will likely win the popular vote comfortably next month. This diary introduces a model (EMWA2017) to estimate which seats are most likely to be won by each party in the legislative assembly. Western Australian state elections for the legislative assembly operate under a system of both compulsory voting and compulsory preferential voting (instant run-off).

The state has been redistributed (redistricted) since the 2013 election. As Australian redistributions are non-partisan this shouldn’t have much impact on the election’s outcome.

As malapportionment in the Legislative Council gives one-third of the seats to the 8% of the population living outside the south-west it is unlikely Labor will be able to win control of that house.

Polling also suggests that the resurgent One Nation (a very Trumpish party) is likely to win perhaps 10-15% of the popular vote. While this probably won’t be enough to win Assembly seats it could help or harm the governing parties enough to throw the model out in the rural seats to some degree. I can’t do much about this other than note the possibility as there is no historical data to work with. Finally, this model is calibrated for left-right match-ups. In seats where the Liberals and Nationals are competing with each other (Kalgoorlie, Pilbara, etc ) the model just awards the seat to the incumbent party.

The table in this diary is fairly straightforward; the first column contains the names of each seat colour coded to the party currently holding them (Red=Labor, Blue=Liberals, Green=Nationals). Hillarys is held by an independent but as he was elected as a Liberal I’ve left his seat blue pending an election win without the party label.The percentage in the middle column represents EMWA2017’s estimate of the likelihood of either the Liberal or National party winning the seat.

If the public polling is accurate Labor are almost certain to win Government with a median of 37 seats in the 59 seat parliament.

Rather than reinvent the wheel, if you are interested in detailed seat profiles I’d suggest heading over to Ben Raue’s Tallyroom.

Modelling the 2016 Australian Federal Election: Final Week

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Wednesday June 29, 2016 · 8:42 PM PDT

2016/06/29 · 20:42

Australia is now just hours away from the federal election taking place on the 2nd of July.

The election is to be a double dissolution election, with both the House of Representatives and the entire Senate up for election. Generally, in Australia, governments call elections for the House and half of the Senate, this has been the case for the last 30 years, though technically House-only and half-Senate-only elections can also be called. A double dissolution election can only be called if the Senate fails to pass Government bills. The current Senate failed to pass a bill allowing the Government to create a watchdog with special powers to investigate the construction industry. This issue has not been discussed during the campaign. Australian federal elections for the House operate under a system of both compulsory voting and compulsory preferential voting (instant run-off).

This diary and my current predictive model, Federal Australian Model 2016 Election (FAME2016), are only interested in the House election, though the Senate election is certainly interesting in its own right.

The Government is currently a coalition of the Liberal and National parties, elected in 2013, referred to as the Coalition. These parties have worked together when in federal government for generations, though at times, in certain states, their relationship has been far less amiable. The Liberal party is basically a conservative centre-right party representing the interests of business. The Nationals represent primarily rural electorates and have a focus on the interests of agriculture. Taken together, the Coalition currently hold 90 seats in a 150 member parliament.

The main opposition is the Labor party, their name is less confusing, they represent the interest of labour. Although in recent decades they have embraced open-markets and the deregulation of the economy it would still be fair to categorise them as centre-left. The Australian Labor Party currently hold 55 seats.

The Greens are the third main force in Australian politics. Firmly left of center, they are no longer focused on the environment specifically but have a full range of policies. Progressives. The Greens hold only one House seat, they have considerably more influence in the Senate, but are looking to win more House seats over the next decade.

A small number of House seats are also held by Independents. In the event of a close election, where neither main party win 76 seats in their own right, these representatives would choose which party will form Government.

FAME2016 currently suggests that the most likely outcome of an “election held tomorrow” would be a small Coalition majority government holding 80 seats. However, the Government wins a second term in only 90.3% of simulations. In 9.7% of simulations the parliament would have no single party with majority control. In no simulations would Labor win government outright.

Unfortunately, this all assumes that the Coalition won’t lose any seats to independents this cycle. Which they probably will; see the discussion for New South Wales and South Australia. Therefore a Coalition majority is somewhat less likely and a hung parliament somewhat more likely. For example; if we assume the Coalition lose a total of three seats to independents they will only win majority government 60% of the time and a hung parliament occurs in 40% of simulations.

So in summary; provided the national polls are as accurate are they have historically been, Labor don’ t currently have a chance of forming majority government. The Coalition will be reelected to either a slender majority, or a difficult hung parliament.

The tables in this diary are fairly straightforward; the first column contains the names of each seat colour coded to the party currently holding them. The percentage in the middle column represents FAME2016’s estimate of the likelihood of the Coalition parties winning the seat.

New South Wales has undergone a redistribution (redistricting) since the 2013 election. This has all but guaranteed Labor pick-ups in Barton and Paterson. There are also a large number of other opportunities for Labor if they can pick up just a little more of a swing in the final days of the campaign. Labor can go a long way to winning government with a strong result in this state, though a popular Liberal state government should help to keep the Coalition vote high.

The Seat of New England, while out of reach for Labor, is well worth watching. The former independent member, Tony Windsor, is challenging the colourful deputy Prime Minister, Barnaby Joyce. Windsor was at one point extremely popular, however, his support for the Labor minority Government of 2010-2013 has potentially eroded some of his support. He did not stand for reelection in 2013 and Joyce moved to New England. Joyce often finds himself in the news but he may be most notable in the States for once threatening to shoot Johnny Depp’s dogs.

Similarly, Rob Oakeshott, who supported the Labor minority government of 2010-13, is aiming to return to parliament in the seat of Cowper. Polls suggest it will be a close race.

Labor performed quite strongly in Victoria back in 2013 so there are only a limited number of seats that are realistically available to take from the Liberals.

Labor have to play defence, in some cases due to retiring incumbents, in others because of a state issue with the County Fire Authority, in Bruce, McEwen, Bendigo, and Chisholm but do have opportunities in Dunkley, Deakin, Corangamite, and LaTrobe.

Indi should be held by the incumbent independent Cathy McGowan who defeated the personally unpopular Sophie Mirrabella in 2013. The Liberals have made the questionable decision to renominate Mirrabella. Polls suggest McGowan will win easily.

Melbourne should be easily held by the popular Greens representative Adam Bandt. In 2013 he recorded a strong swing to him in the teeth of a fall in the Greens vote nationally. The Greens are also competing for the seats of Batman, Wills, Melbourne Ports, and Higgins; however, the Liberal party has chosen to recommend their supporters preference the Labor ahead of the Greens. This should protect sitting Labor members, though the Greens are still a strong chance in Batman, a candidate for the greatest electoral name anywhere in the world.

Labor have lost most of their Queensland seats over the past two elections, which leaves a lot of pick-up opportunities. The path to majority government will most likely need to include a good night in Queensland.

There are some suggestions that Labor looks stronger in the non-Brisbane seats of Capricornia, Leichhardt, Herbert, and Flynn than in the Brisbane seats of Petrie, Bonner, Brisbane, and Forde.

Eccentric billionaire(???) Clive Palmer will not be recontesting his Sunshine Coast seat of Fairfax, so that seat reverts to a likely Coalition gain.

Bob Katter had a surprisingly close race in his outback seat of Kennedy three years ago but will probably win again this year without evidence to the contrary. Particularly given the Labor party will be recommending its supporters give him their preference vote.

Western Australia is traditionally a very tough state for Labor to compete in. However, they are currently polling rather well there and could well pick up a number of seats from the Liberal party. Particularly targetted seats are Cowan, Hasluck, and the newly created seat of Burt.

The Western Australian Nationals are, confusingly, not part of the Coalition federally and will be pushing hard in the vast rural seats of Durack and O’Connor. This could add a wrinkle if the election is very close. That said, Labor are suggesting their supporters should offer preferences to the incumbent Liberals, which should help keep the WA Nationals out of federal parliament.

South Australia has been turned upside down by popular populist Senator Nick Xenophon who is running candidates in every South Australian lower house seat. Polling suggests these candidates are likely to win at least one seat, and very possibly more. Mayo is the most likely Nick Xenophon Team (NXT) pickup, probably followed by Barker, Grey, and Sturt. You’ll notice these are all Liberal held seats. It is entirely possible that a Coalition victory nationwide could be spoiled by NXT destroying the South Australian Liberal contingent.

Labor are making a much stronger push for the seat of Hindmarsh then they are Boothby, though right now it seems they will come up short in both seats.

Tasmanian poll numbers tend to be a little less reliable than the mainland states. Right now it looks like Tasmania isn’t experiencing the same swing to Labor as the rest of the country. This could well be true, and FAME2016 assumes it is, but it is a little surprising and doesn’t mesh with seat based polling. If FAME2016 is wrong anywhere I expect it to be in Tasmania.

Polling of the territories is notoriously bad and therefore I don’t necessarily trust FAME2016 for those four seats.

The two seats in the Australian Capital Territory (ACT), Canberra and the newly renamed seat of Fenner, are both perpetually completely safe for the Labor party.

The current territory-level government in the Northern Territory is breathtakingly unpopular due to an absolutely epic level of incompetence and in-fighting. There is the very real possibility that they could lose every single seat at this year’s territorial elections. Surely that will affect federal voting behaviours? If so FAME2016 is probably overestimating Country Liberal Party chances.

Finally, I’ve got a handy table and pie chart to summarise FAME2016’s current seat ratings.

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Thursday May 12, 2016 · 8:55 PM PDT

2016/05/12 · 20:55

The Prime Minister of Australia, Malcolm Turnbull, has announced an election will take place on the 2nd of July.

The election is to be a double dissolution election; both the House of Representatives and the entire Senate is dissolved and is up for election. Generally, in Australia governments call elections for the House and half of the Senate, this has been the case for the last 30 years, though technically House only and half-Senate only elections can also be called. A double dissolution election can only be called if the Senate fails to pass Government bills. The current Senate failed to pass a bill allowing the Government to create a watchdog with special powers to investigate the construction industry. Australian federal elections for the House operate under a system of both compulsory voting and compulsory preferential voting (instant run-off).

The Government is currently a coalition of the Liberal and National parties elected in 2013, referred to as the Coalition. These parties have worked together when in federal government for generations, though at times, in certain states, their relationship has been far less amiable. The Liberal party is basically a conservative centre-right party representing the interests of business. The Nationals represent primarily rural electorates and have a focus on the interests of agriculture. Taken together, the Coalition currently hold 90 seats in a 150 member parliament.

The Greens are the third main force in Australian politics. Firmly left of center they are no longer focused on the environment specifically but have a full range of policies. Progressives. The Greens hold only one House seat, they have considerably more influence in the Senate, but are looking to win more House seats over the next decade.

FAME2016 currently suggest that the most likely outcome of an “election held tomorrow” would be a small Coalition majority government holding 77 seats. However the Government win a second term in only 62.1% of simulations. In 37% of simulations the parliament would have no single party with majority control. In just 1.8% of simulations would Labor win government outright.

The tables in this diary are fairly straightforward but the first column contains the names of each seat colour coded to the party currently holding them. The percentage in the middle column represents FAME2016’s estimate of the likelihood of of of the Coalition parties winning the seat.

New South Wales has undergone a redistribution (redistricting) since the 2013 election. This has all but guaranteed Labor pick-ups in Barton and Paterson. There are also a large number of other opportunities for Labor if they can pick up just a little more of a swing. Labor can go a long way to winning government with a strong result in this state. A popular Liberal state government should help to keep the Coalition vote high.

Labor performed quite strongly in Victoria back in 2013 so there are only a limited number of seats that are realistically available to take from the Liberals.

Labor have to play defence, in some cases due to retiring incumbents, in Bruce, McEwen, and Chisholm but do have opportunities in Dunkley, Deakin, Corangamite, and LaTrobe.

Melbourne should be easily held by the popular Greens representative Adam Bandt. In 2013 he recorded a strong swing to him in the teeth of a fall in the Greens vote nationally. The Greens are making noises about competing for the seats of Batman, Wills, Melbourne Ports, and Higgins. If the Liberal party recommend their supporters preference the Greens ahead of Labor, as there are some suggestions might happen, the Greens could very well win Batman and Wills.

Labor have lost most of their Queensland seats over the past two elections, which leaves a lot of pick-up opportunities. The path to majority government most likely will need to include a good night in Queensland.

Eccentric billionaire(???) Clive Palmer will not be recontesting his Sunshine Coast seat of Fairfax, so that seat reverts to a likely Coalition gain.

Bob Katter had a surprisingly close race in his outback seat of Kennedy three years ago but will probably win again this year without evidence to the contrary.

Western Australia is traditionally a very tough state for Labor to compete in. However, they are currently polling very strongly there and could well pick up a number of seats, even a large number of seats, from the Liberal party. Particularly targetted will be Cowan, Hasluck, and the newly created seat of Burt. An unpopular state Liberal government appear to be dragging down the federal vote.

The current Liberal incumbent in Tangney, Dennis Jensen, has been dropped by his party and will be contesting the seat as an independent. If Tangney is close this could make the difference at the margins. Should he drop his skittles entirely and unexpectedly direct preferences to Labor this may cost the Liberals the seat. I don’t believe Jensen is popular enough to win in his own right absent polling.

South Australia isn’t a state where a lot of seats change hands at any given election. Labor will make a play for Boothby and Hindmarsh and the Liberals will take a look at Wakefield and maybe Adelaide.

This would be all there is to say except that popular Senator Nick Xenophon is going to be running candidates in every South Australian lower house seat. Polling suggests these candidates might be capable of winning a seat or seats. Mayo being the most likely, probably followed by Sturt. Kevin Bonham has written an extensive analysis. I’ll be keeping a close eye on the likelihood of Xenophon endorsed seats because it is very hard to say which party Xenophon would support if it came down to that.

Tasmanian poll numbers tend to be a little less reliable than the mainland states. Right now it looks like Tasmania isn’t experiencing the same swing to Labor as the rest of the country. This could well be true, and FAME2016 assumes it is, but it is a little surprising. Labor’s shock victory in the Tasmanian upper house seat of Elwick suggests polling might become more favourable to Labor before the election. Labor are certainly spending resources trying to reclaim seats in this state.

Polling of the territories is notoriously bad for a few reasons and therefore I don’t trust FAME2016 for those four seats.

The two seats in the Australian Capital Territory (ACT), Canberra and the newly renamed seat of Fenner, are both perpetually completely safe for the Labor party.

The current territory-level government in the Northern Territory is breathtakingly unpopular due to an absolutely epic level of incompetence and in-fighting. There is the very real possibility that they could lose every single seat at this year’s territorial elections. That will surely affect federal voting behaviours. I expect the Coalition will win neither Lingiari or Solomon in a close election. Total seat numbers assume as much, so if you think that I’m wrong you could mentally adjust the Coalition’s final tally up a seat or two.

Finally, I’ve got a handy table to summarise FAME2016’s current seat ratings.

Final Model of Today's New South Wales Election (Australia)

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Saturday March 14, 2015 · 5:12 AM PDT

2015/03/14 · 05:12

New South Wales (Australia) will be holding a state election today, the 28th of March, to decide the make-up of its parliament for the next 4 years. Currently New South Wales is governed by the centre-right Liberal party and the rural National Party. The opposition is the centre-left Labor party.

The government currently commands a 61-23 majority in the 93 seat Legislative Assembly. At the previous state election in 2011 the Liberals and Nationals had returned 69 representatives but a lengthy series of corruption scandals caused 8 MLAs (Members of the Legislative Assembly) to resign from the Liberal party. Six of these have remained in parliament as independent MLAs but most are not standing at the election and none are regarded any chance to win reelection. Two, in Newcastle and Charlestown, resigned altogether and their seats are now held by Labor after by-election (special election) wins. In Miranda a Liberal MLA resigned for non scandalous reasons and Miranda was also won by Labor in a by-election. Finally the Nationals won a by-election in Northern Tablelands after the incumbent Independent resigned over... corruption.

NESWEM (NEw South Wales Electoral Model) is largely based off of and builds on Venom (Victorian Electoral NOminative Model) and my Federal model. Changes are mostly small tweaks to adjust for variance in variance between rural and urban divisions (districts). Like previous models NEWSEM takes into consideration the prior voting history of each electorate, incumbent strength (where applicable), and public polling.

Voting in New South Wales is compulsory and uses a optional preferential ballot (this means that in certain seats the way parties direct their supporters to allocate their preferences will be crucial) in single-member seats for the Legislative Assembly.

Tables in this diary are colour coded. Shades of red reflect Labor predictions, shades of blue represent Liberal predictions, Nationals predictions are dark green, Greens predictions are light green, Independent predictions are grey. The percentages indicate the overall percentage of simulated outcomes where either the Liberal or National party wins the seat. The second column is the median two-party preferred result with the Liberal/National percentage first, that is Liberal/National-Labor.

While there are currently no rural Independents in the New South Wales parliament, despite a long tradition, there is a strong possibility of the Nationals losing rural seats due to perceived National support of coal seam gas extraction particularly in the northern parts of the state. The seat in particular danger from an Independent is Tamworth. Lismore, Ballina, and Tweed are in more danger than the model suggests from Labor and/or the Greens. The Independents in the seats of Sydney and Lake Macquarie are both likely to win reelection and could support either a Liberal or Labor government.

The Greens could quite possibly, polling is non-existent, win Balmain, which they currently hold, and Newtown, which they nominally hold (that is to say if every voter in the division of Newtown votes the same way as they did in 2011 the Greens would win Newtown), Summer Hill is an outside chance if everything had gone right. The Greens are also very strong in Lismore and Ballina and have a chance of winning their first state seat outside of a capital city. The Greens would almost certainly support a Labor government if the later didn't reach 45 seats themselves.

NESWEM currently projects the Liberals and Nationals to combine to win 57 seats. However it is more likely than not that at least one of Tamworth, Lismore, and Ballina will be lost and 55 is more realistic target for the Government. With only 47 needed to form government this should be sufficient to see the current Government returned for a second term.

Next Weekend's Victorian State Election Modeled with Venom

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday November 21, 2014 · 7:08 PM PST

2014/11/21 · 19:08

Victoria (Australia) is currently governed by the Coalition, a permanent coalition of the centre-right Liberal party and the rural National Party. The opposition is the centre-left Labor party. The Coalition currently command just a 44-43 majority. There is also a single scandal-slicked independent, Geoff Shaw, who resigned from the Liberal party ahead of a vote to expel him earlier in the year.

The next Victorian state election will be held today, the 29th of November, and public polling has become very frequent, ReachTel, Galaxy, Essential, Morgan, Ipsos, and Newspoll have all released polling in the last few days, so it's time to make one last pre-election modelling post.

Venom (Victorian Electoral NOminative Model) is largely based off of my Federal model but has a few additional features to account for a recent redistribution (redistricting) and for systemic polling errors across the entire industry, this refers to an additional variable for accidental systemic bias not an attempt to unskew. Like the Federal model Venom takes into consideration the prior voting history of each electorate, incumbent strength (where applicable), and public polling. Venom calculates a predicted two-party preferred vote for each seat 30,000 times. Venom predicts a Labor majority government 90.4%, a hung parliament 3.8%, and a Coalition victory 5.8% of the time based on final polling.

Voting in Victoria is compulsory and uses a preferential ballot (this means that in certain seats the way parties direct their supporters to allocate their preferences will be crucial) in single-member seats for the Legislative Assembly.

Tables in this diary are colour coded. Shades of red reflect Labor predictions, shades of blue represent Liberal predictions and Nationals predictions are dark green. The percentages indicate the overall percentage of simulated outcomes where one of the Coalition parties wins the seat. The final column is the median two-party preferred result with the Labor percentage first, that is Labor-Coalition.

The column graph shows the number of simulations in which the governing Coalition wins a certain number of seats in the 88 member Victorian Legislative Assembly. Red columns represent a Labor government and blue columns a Coalition government. As Victoria, inconveniently, has chosen to have 88 seats in their Legislative Assembly it is possible for the voters to elect a deadlocked parliament with 44 members from each major party. As Victoria has fixed term elections this could prove to be rather problematic. This column representing this hung parliament situation is grey.

It is possible that not all seats will be won by either Labor or one of the Coalition parties. Current the Coalition are predicted to hold (both mean and median) 40 seats after the election, by implication this would give Labor 48 seats and a 8 seat majority. However...

The Greens have come very close to winning a number of seats in the past and on current polling would be considered non-zero chances of winning one or more of Brunswick, Melbourne, Northcote, Prahran, and Richmond. The Greens would almost certainly support a Labor government if the later didn't reach 45 seats themselves.

Whilst there are no rural independents at present in the Victorian parliament, despite a long tradition, there is always the possibility of unexpected victories. National held Morwell is perhaps the most likely seat to elect an independent as there is significant local resentment at the government response to a mine fire that burned for 45 days and dumped large amounts of toxic smoke on the town of Morwell.

Finally I'll note that there is a chance that Ivanhoe will underperform for Labor as former Labor member Craig Langdon is running as an independent and directing his preferences to the Liberals.

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Monday October 27, 2014 · 4:34 AM PDT

2014/10/27 · 04:34

The next Victorian state election will be held on the 29th of November and public polling has begun to become more frequent, ReachTel, Galaxy, Essential, Morgan, Ipsos, and Newspoll have all released polling in the last few weeks, so it's time to update my latest model. Venom (Victorian Electoral NOminative Model) is largely based off of my Federal model but has a few additional features to account for a recent redistribution (redistricting) and for polling errors. Like the Federal model Venom takes into consideration the prior voting history of each electorate, incumbent strength (where applicable), and public polling. Venom calculates a predicted two-party preferred vote for each seat 10,000 times. Currently Venom predicts a Labor majority government more than 99.9% of the time based on current polling.

Whilst there are no rural independents at present in the Victorian parliament there is always the possibility of unexpected victories. National held Morwell is perhaps the most likely seat to elect an independent as there is significant local resentment at the government response to a mine fire that burned for 45 days and dumped large amounts of toxic smoke on the town of Morwell.

I'll run a fresh simulation on Venom whenever fresh polling data is available.

Modelling the Victorian State Election - With Venom

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Saturday October 04, 2014 · 1:31 AM PDT

2014/10/04 · 01:31

Victoria is currently governed by the Coalition, a permanent coalition of the centre-right Liberal party and the rural National Party. The opposition is the centre-left Labor party. The Coalition currently command just a 44-43 majority. There is also a single scandal-ridden independent, Geoff Shaw, who resigned from the Liberal party ahead of a vote to expel him six months ago.

The next Victorian state election will be held on the 29th of November and public polling has begun to become more frequent so it's time to unveil my latest model. Venom (Victorian Electoral NOminative Model) is largely based off of my Federal model but has a few additional features to account for a recent redistribution (redistricting) and for polling errors. Like the Federal model Venom takes into consideration the prior voting history of each electorate, incumbent strength (where applicable), and public polling. Venom calculates a predicted two-party preferred vote for each seat 10,000 times.

Tables in this diary are colour coded. Shades of red reflect Labor held seats and predictions, shades of blue represent Liberal held seats and predictions and Nationals are dark green. I've elected to colour Geoff Shaw's seat of Frankston Liberal blue despite his current independence as there is a zero chance of his re-election.

The column graph show the number of simulations in which the governing Coalition wins a certain number of seats in the 88 member Victorian Legislative Assembly. Red columns represent and Labor government and blue columns a Coalition government. As Victoria, inconveniently, has chosen to have 88 seats in their Legislative Assembly it is possible for the voters to elect a deadlocked parliament with 44 members from each major party. As Victoria has fixed term elections this could prove to be rather problematic. This column representing this hung parliament situation is grey.

It is possible that not all seats will be won by either Labor or one of the Coalition parties.

I'll run a fresh simulation on Venom in the week immediate prior to the election in late November using the final polling data.

Australian Federal Election (2016?) Modelling

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Saturday August 30, 2014 · 3:15 AM PDT

2014/08/30 · 03:15

It is now almost a year since the last Australian Federal election, which saw the centre-right Liberal/National Coalition defeat the governing centre-left Labor party and take a 90-55 seat advantage into the 44th Parliament of Australia.

Since then things have not gone especially well for the Coalition. By the start of December they had slipped behind in the opinion polls and haven't been back in front since. Polling looked particularly dire in the months following the handing down of a budget that was widely seen as deeply unfair for retirees, students, and low-income earners. Whilst they were successful in repealing the Carbon Tax successive Senates have proven hostile to most other reforms.

In recent weeks a number of overseas incidents have dominated the news and the Government has narrowed Labor's persistent lead in the polls. Nevertheless my modelling gives Labor a 24 seat gain and a 79-66 majority in the next parliament. The model assumes that all five independents/minor party Representatives will hold their seats. The timing of Federal elections in Australia is not fixed so the Government is able to call an early election if they so chose, however a late 2016 election is most probable.

Voting in Australia is compulsory and uses a preferential ballot (this means that in certain seats the way parties direct their supporters to allocate their preferences will be crucial) in single-member seats for the House of Representatives.

Tables in this diary are colour coded. Shades of red reflect Labor held seats and predictions, shades of blue represent Liberal held seats and predictions (the occasionally different party names and abbreviations are courtesy of local party branches having inconsistent names), independents are grey, Greens are light green (shocker), and Nationals are dark green. The percentage is the odds of the Coalition winning the seat. I've dropped the predicted two-party preferred vote for now, it's just a bit cluttery this far out from an election.

New South Wales will be undergoing a redistribution before 2016 where it will lose a seat. This is on balance better news for Labor than the Coalition as currently the Liberals hold most of the state's marginal seats and incumbents are stronger if they don't have to represent new areas.

In non-federal news my home city of Newcastle has had both its State MP and its Lord Mayor resign over corruption issues (envelopes full of $100 bills) in the last couple of weeks. So that's something.

If Labor are to regain government at the next election they will need to make gains in Queensland. Eccentric billionaire Clive Palmer and his eponymous Party poll by far the highest in Queensland which presents a modelling problem because it's somewhat hard to say whether Palmer's United Party (PUP) is drawing votes from the Coalition or Labor and how these voters will direct their preferences. PUP's policies are broadly populist but also profoundly unpredictable. The party seems to draw mainly those voters who feel alienated from the traditional parties.

Labor are unlikely to make significant gains in Victoria but polling doesn't give the Coalition much hope of winning back anything to counter-balance probably losses elsewhere. Freshman independent MP Cathy McGowan seems popular enough and I'd consider her likely to hold Indi.

Western Australia is the other state that will be redistributed before 2016, this time gaining a seat. This is probably good for Labor as the new seat is likely to be competitive. But you never can tell with redistributions.

South Australia is not a state of marginal seats. Labor will hope to regain Hindmarsh and maybe win Boothby but South Australia is not likely to be where a Government will be won or lost.

Labor won the two party preferred vote in Tasmania in 2013 and came home with one seat for their troubles. Gains are probable in 2016 but it isn't a big state and Andrew Wilkie has Denison locked down.

I'm, as always, completely lacking in confidence that the model has the Northern territory seats correct but right now it seems plausible enough that Labor would win both Solomon and Lingiari. Labor is perpetually safe in the ACT seats of Fraser and Canberra.

Analysis of Modelling the Australian Federal Election

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Tuesday November 12, 2013 · 2:48 AM PST

2013/11/12 · 02:48

This diary is an analysis of the model I developed to predict the seat-by-seat outcome of the Australian federal election held in September.

Unfortunately late polling released the morning of the election was somewhat worse for Labor than the consensus prior to that. My last full update used only the data available the day before the election. I mentioned in a last minute update that the model's final prediction was 91 seats to the Coalition and 56 to Labor but wasn't able to provide final seat by seat predictions due to time constraints (I work at Australian elections and the election day polling was released literally minutes before I was legally obliged to abjure posting on the internet). So this analysis post refers to the model's final prediction rather than my final posted prediction.

In reality the Coalition has won 90 seats and Labor won 55. However the Coalition also lost Indi to a conservative local independent and Fairfax to the conservative eccentric billionaire Clive Palmer (of Titanic II and world's-largest-dinosaur-park fame) so as far is the model is concerned the Coalition won 92 seats (the model can only actually predict if a seat will be won by the left or the right sides of politics). Therefore the model was, overall, off by only a single seat with Labor losing one more seat to the right then was predicted.

However the model purported to be able to predict the results and final two-party preferred percentages in each individual seat rather than merely an overall seat prediction. Therefore here are tables detailing the predicted Coalition winning percentage in each seat, the actual Coalition vote, and the model's predicted Coalition vote.

You might notice that eleven seats are missing from the tables. These seats had a "non-classic" two party preferred vote so Labor/Greens or Liberal/National or Coalition/Right-wing minor party were the top two vote-getters. Therefore I don't know the actual left-right two-party preferred breakdown it these seats, yet.

Labor had a tough time in New South Wales and lost eight seats including three of the four tossups and one lean Labor seat. Labor did manage to hold onto Greenway easily even though the model didn't care much for their chances. This one was down to local factors, the Liberal candidate was incredibly terrible. It was embarrassing for them.

Labor lost three seats in Victoria (including La Trobe against the prediction of the model) and had a very narrow escape in McEwen, which the model had rated as safe. Adam Bandt held Melbourne very easily for the Greens, in the end, to widespread surprise. The unpopular fire-breathing Liberal incumbent in Indi was edged out by a local independent in a display of the merits of preferential voting.

The media was talking about a wipeout for Labor in Queensland but the model always maintained they'd lose no more than one or two seats. In the end they lost two. Clive Palmer won in Fairfax and Bob Katter clung on in a shockingly close race in his personal fiefdom of Kennedy.

Not much happened in WA with no seats changing hands. The model did very badly in the Coalition seat of Canning with by far it's worst miss of the election (9.5 point miss, no other non NT misses by over 6).

Labor lost Hindmarsh but held the rest contrary to media speculation. It was the only seat the model believed to be in major danger.

Things did not go well for with Labor or the model here. I blame poor polling data. Garbage In Garbage Out. I mentioned frequently my lack of faith in polling for the territories (mostly because there basically is none) and doubt over the model's prediction there.

Labor had some bad luck in Tasmania. They won a majority of the two-party preferred vote but hold just one seat.

Overall the model was pretty accurate. Every seat ranked at least "likely" to be won by either party was (except Lingiari in the Northern Territory but I said not to trust that one) even though only 18 of 150 seats were regarded as a "lean" or "tossup". 30% of all seats were predicted within 1 percent of the actual result and a further 32% were within 2 percent (16% 2-3 percent miss, 14% 3-4 percent miss, 6.5% 4-6 percent miss, only Canning and Lingiari outside that).

Over the next six months or so I'll update the model for the next election cycle. There are data availability problems at present due to the probable fresh election for the Western Australian Senate.

Australian Federal Election - Final Predictions

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday September 06, 2013 · 5:53 AM PDT

2013/09/06 · 05:53

With polls opening in just a few hours there is universal consensus that the Coalition will defeat the Labor government and wield a sizable majority in the new house. My model is predicting a final result of 85 seat to the Coalition and 62 for Labor.

This prediction assumes Labor will not reclaim Melbourne from the Greens, the independents Wilkie and Katter will both hold their seats, and that the Coalition will hold O'Conner, Durack, and Indi from non-coalition conservative challengers.

The next few paragraphs are basic Australian background and housekeeping related information. If you've read my previous Australian Election modelling diaries you might like to skip straight to the state breakdowns where you'll find the usual tables now include percentage chances for the Coalition to win each seat.

Currently Australia is governed by the Labor party (centre-left, loves unions) with the support of a number of independents (3 conservative, 1 Laborish, 1 idiosyncratic but left-leaning) and the Greens (progressives). The opposition is known as the Coalition, with an uppercase C as their coalition is permanent, and is composed of the Liberals (centre-right, hates unions, loves free markets) and Nationals (conservative agrarian socialists, loves protectionism). The Labor party need the independent and Green support as they do not have enough seats to govern in their own right (and in fact have less seats (71) than the Coalition (72)).

Finally before we get to the numbers I'll just mention a few details in regards to the model I am using. The model takes into consideration the prior voting history of the electorate, incumbent strength (where applicable), and public polling. It doesn't account for potential asymmetric swings within states (not enough polling data available). I'm interested to see how much that ends up mattering.

Tables in this diary are colour coded. Shades of red reflect Labor held seats and predictions, shades of blue represent Liberal held seats and predictions (the occasionally different party names and abbreviations are courtesy of local party branches having inconsistent names), independents are grey, Greens are light green (shocker), and Nationals are dark green. The numbers either side of the dash are the two party preferred percentages (Labor-Coalition). The percentage is the odds of the Coalition winning the seat.

New South Wales

New South Wales is the state where Labor is in the biggest trouble and will almost certainly lose the most seats. Seat based polling suggests that the Western Sydney seats are more likely to elect a Coalition representative than the model predicts. Dobell is also widely considered to be certain to switch to the Coalition for seat specific reasons, I'm ok with the model's numbers there though.

Victoria

The model currently predicts only modest gains for the Coalition in Victoria.

Things are going horribly wrong for the Coalition in Indi where the controversial Liberal incumbent is receiving a strong challenge from a local independent. I suspect the Liberals are going to lose Indi at this point, fortunately for them it is unlikely to effect their chances of forming government.

Both Labor and the Greens are claiming they will win Melbourne. It's a tossup.

The Liberals may win Mallee from the Nationals. A Coalition seat either way.

Queensland

Queensland has big question marks over it given recent polling showing significant strength for Palmer's United Party (mining billionaire Clive Palmer has founded his own party and is running as a anti-politician, proudly ignorant of economics). The rise of Katter's Australian Party (an eccentric representative from the outback formed his own socially conservative/protectionist party) also makes things hard to predict.

Given Australia's compulsory preferential voting system it's going to matter a lot how these conservative party voters allocate their preferences. I'm going to guess everything is going to cancel itself out and the model will be pretty close but it is possible Labor's chances in Northern Queensland are understated and in Brisbane are overstated.

Bob Katter will retain his seat.

South Australia

The model predicts a status quo result in South Australia as the most likely result.

Western Australia

I suspect the model is overstating Labor's chances in Canning and understating them in Perth. The Western Australian Nationals are not technically part of the Coalition and might win either or both of Durack and O'Connor.

Tasmania

Labor seem unexpectedly strong in this iteration of the model. Andrew Wilkie is likely to retain his seat.

Territories

I've little faith in the model's ability to call the Northern Territory seats of Solomon and Lingiari accurately, the polling is too weak.

Summary

Thanks for following my modeling of Australia's Federal election! I'm working for the Electoral Commission again at the election and won't be around tomorrow. I'll post a full review of the model's accuracy in a few weeks when the election results are finalized.

Alizarin Indigo

Community (This content is not subject to review by Daily Kos staff prior to publication.)

Friday August 30, 2013 · 6:33 AM PDT

2013/08/30 · 06:33

Just one more week until election day in Australia and it appears clear that Tony Abbott and the Coalition will be winning government and it probably won't be particularly close. In the interests of not burying the results the model currently predicts the Coalition a 100% chance of winning government with a median seat result of 87-60 (3 independents). Coincidentally half of the Senate is also up for election and the Coalition will be hoping to, if not win control in their own right, at least help ensure the election of sufficient Senators that are at least somewhat friendly to their overall objectives.

The method of electing Senators is ridiculously complicated. There isn't, in my opinion, any way to realistically model likely outcomes. The subjectiveness of inputs, given there is no polling, combined with the byzantine nature of party assigned preferences, makes anything more than vague predictions unrealistic. So I'm going to suggest a potential range of outcomes in each state without assigning exact probabilities.

Each state is represented by 12 Senators. Every three years six of these Senate positions are up for election. All six Senators are elected out of the same pool of votes. Any parties or candidates receiving more than 14.3% of the total formal votes cast is elected to one of the available seats. In the event that all six vacancies are not filled at this point then we get into preferences (like the house, but way more complicated).

Preference allocation occurs in rounds. First the party with the lowest vote total is eliminated from contention. Then the votes gathered by this party are allocated to another party according to a list lodged by all parties with the Australian Electoral Commission a few weeks ago (or according to the voters wishes, if they are prepared to fill out as many as 110 boxes (NSW) in sequential order, not many people do that). If this process means a party has cleared the 14.3% threshold they are elected. If not then another round of preference allocation occurs.

There are a lot of parties running for Senate. It varies by state but often dozens run, most parties get almost no votes but due to the preference allocation system works it is quite possible to be elected to a Senate seat on way less than 14.3% of the primary vote. In New South Wales this year senate papers are printed in 7 point font and voters are provided with magnifying glasses so they can read them. It's not the best system.

The territories also elect two Senators each (at the same time as the House, don't wonder why) giving a total of 76 Senators. 39 Senators, one more than half, are required to pass legislation.

Currently the Left-leaning parties control the Senate (Labor, 31 seats; Greens, 9 seats). The Coalition have 34 Seats and there are two other Senators. The independent Nick Xenophon in South Australia and John Madigan of the DLP in Victoria. The Coalition would like to get as close to 39 seats as they can, but at the very least they would prefer Labor and the Greens are reduced to 37 seats or less so that they are unable to block legislation.

I'll talk about each State's Senate possibilities in each State's own section.

As previously the model assumes that the Independent Andrew Wilkie in Denison, Adam Bandt of the Greens in Melbourne, and Bob Katter of an Australian Party of the same ilk in Kennedy will all win reelection. This is not an entirely safe assumption.

New South Wales

Things are deteriorating badly for Labor in New South Wales. There is evidence that swings against Labor in Western Sydney are going to be stronger then elsewhere in the state. The model isn't accounting for that (assuming it is true) and I wouldn't be surprised if Labor lose more seats in Western Sydney than is predicted.

In 2007 New South Wales elected three Coalition and three Labor Senators. Labor won't elect three senators again. The third Labor seat could be won by the Greens, or Family First, or the Shooters and Fishers, or even the legendary Pauline Hanson.

Victoria

The Liberal member for the seat of Indi (they are rather personally unpopular) is receiving a strong challenge from an independent. I've changed the rating of Indi to Likely Lib, though this is to reflect the chance of the independent winning not any chance for Labor.

In 2007 Victoria elected three Senators from both the Coalition and Labor. Labor will not poll well enough to repeat the feat and will almost certainly lose a seat to the Greens. The Coalition may struggle to hold their third seat against a flood of minor parties (they failed to win a third seat in 2010, losing out to the DLP (socially conservative, economically left-wing)).

Queensland

Labor are struggling in Queensland. Like everywhere. Seat based polling has even Prime Minister Rudd behind in Griffith.

Labor have negotiated Katter's Australian Party (KAP) preferences across northern Queensland. This will presumably give Labor a better chance in these seats (Leichhardt, Herbert, Capricornia, Flynn, Dawson) than the model predicts. I've got neither past election data, or polling, to help me model this.

In 2007 Queensland boringly elected three LNP Senators and three Labor Senators. That won't be repeated. Labor have chosen to preference KAP ahead of the Greens in Queensland. They might not have many votes to give out though. Even so the most likely outcome will be three LNP, two Labor, one KAP. It remains possible the Greens could win a seat (Palmer's Australian Party is preferencing them). The Fishing and Lifestyle (they like fishing and outdoorsy stuff) and Family First (socially conservative Christian) parties also have very good preference flows.

South Australia

Labor are in danger of losing a seat or two in South Australia in the House. Also importantly Labor are in danger of losing a Senate seat.

In 2007 South Australia elected two Labor, two Liberal, one Green and the independent Nick Xenophon to the Senate. Xenophon is a moderate centrist who will work with both major parties and draws significant votes from them both as well. The general perception is Xenophon will again win a Senate seat in his own right but it's anyone guess exactly how he will poll.

I'm confident that two Liberals and one Labor senator will be elected from South Australia. I strongly suspect Nick Xenophon will be reelected. I've absolutely no idea who will win the final two seats. They may be both from the left (one Green, one Labor if they are). They might be both from the right (One extra Liberal, one minor random party (Family First and the No Carbon Tax Climate Skeptics (guess their policies) both have great preference flows)).

Western Australia

Previously the model had rather liked Labor's chances in Western Australia. This is no longer the case. I personally suspect the model may be a little too pessimistic about Labor's chances in Perth for candidate specific reasons.

In 2007 Western Australia elected three Liberals, two Labor, and one Green Senator.

Labor struggled in Western Australia in 2010. Even a fairly small swing against them could lead to Labor being reduce to a single seat. The Greens could also potentially lose their seat despite their traditional strength in this state. However both parties will under no circumstances lose a seat each.

The Liberals should be considered very likely to retain their three seats. If Labor or the Greens were to lose a seat that seat would be most likely won by the WA Nationals (who technically aren't part of the Coalition but would be nearly as good for the Coalition). It is also possible, though markedly less likely, a minor party from the right or even the Liberals could win the last Senate seat.

Tasmania

Polling has Labor looking like receiving a huge swing against them. Labor may very well lose all of their House seats though there is still hope of saving Franklin and Lyons.

In 2007 Tasmania voted strongly for the left in the Senate sending three Labor, two Liberal, and one Green to represent them. Polling is very clear that the Liberals will be electing three senators this year. It is most likely that the Greens will receive enough of a primary vote to reelect their Senator, leaving Labor the remaining two seats. It is possible the Labor may be able to beat the Greens and send three senators again, but in my view less likely.

Territories

Labor is probably in trouble in Lingiari and isn't very likely to reclaim Solomon. Though polling in the Northern Territory is of poor quality and anything could happen.

Both Territories always elect one Coalition senator and one Labor Senator. This could change this year.

In the Northern Territory Julia Gillard (when she was Prime Minister) declared that the long-time Labor Senator from the Northern Territory was to be replaced at the election by an Aboriginal candidate. This met with a mixed response. Even a small swing against Labor could lead to the election of a AFNPP (First Nations) Senator. It's also theoretically possible a Sex party candidate could win election.

In the Australian Capital Territory the Liberals have been living dangerously close to losing their senate seat for years. This might not be the election they finally lose their seat to the Greens (though their candidate cut down the sitting senator in a "primary" challenge and is way less moderate so maybe it is).

Summary

The most likely Senatorial outcomes in my view follow;

The Coalition would appear to be a long way short of winning the five extra seats they would need to win control of the Senate in their own right. However the Left (Labor + Greens) look quite likely, though not certain, to lose at least three (net) seats and control of the Senate. The putative Coalition government will likely be able to pass legislation with the assistance of a selection of Right-aligned Senators. It wouldn't be particularly smooth sailing but it would be much easier than if the Left were to retain control.

Newer
Older

RSS

Alternate Siena Models

Western Australian Election Modelling

Modelling the 2016 Australian Federal Election: Final Week

Modelling the 2016 Australian Federal Election

Final Model of Today's New South Wales Election (Australia)

Next Weekend's Victorian State Election Modeled with Venom

Victorian State Election Modeling - With Venom

Modelling the Victorian State Election - With Venom

Australian Federal Election (2016?) Modelling

Analysis of Modelling the Australian Federal Election

Australian Federal Election - Final Predictions

The Australian Federal Election (Senate Edition)