Talk:2008 Western Australian state election

Onselen's view
"I think it's a cheap political move to try and take advantage of the leadership change in the Liberal party, there's no doubt about that. But I think it's one that may backfire, because the Liberal party, their biggest problem was their leader. Their policy construction was going along perfectly fine, so Colin Barnett has now taken over a united team and he's also got bucketloads of credibility himself. So I think the Labor party are in for the fight of their lives at this election and they're bringing it on sooner than they should have." - Peter van Onselen.

Would there be WP:BALANCE/WP:WEIGHT/WP:OPINION issues if this were added/surmised in the article? Timeshift (talk) 05:10, 7 August 2008 (UTC)
 * Keeping in mind that van Onselen is a former Liberal staffer as well as a commentator and academic... what most publications do when citing him is quote him, but note his positioning. He's often used by the West for commentary, and he cowrote the recent Howard biography. The party is anything but united, too ... the Sunday Times had that famous graphic a couple of months ago on everyone's implacable disputes with each other. It's a pity we have such a shortage of commentators here - David Black is by far the best but is largely retired now. Orderinchaos 07:31, 8 August 2008 (UTC)

Seats held?
ALP 38 to Lib 19 in the infobox with total 59 in lead, however this says ALP holds 30 of 57. Where's the best place to find all the current numbers? Timeshift (talk) 05:18, 7 August 2008 (UTC)
 * The ALP hold 29 to the Libs' 15 and the Nats' 5, with 8 independents, at last count. I don't know where you'd confirm that, but 32 Labor people were elected. Note that Mackerras/Green and others disagree on the benefit of one vote one value - I went through all the calculations and as per them came up with 36 notionally Labor seats, whereas other commentators have come up with 38 or 39. Orderinchaos 07:35, 8 August 2008 (UTC)
 * On thinking about it, I've killed the infobox. There are just too many questions to justify its use:
 * Do we do it on notional seats or actual seats?
 * If notional, whose figures do we use - Mackerras or other?
 * If actual, how do we measure seats needed to win government given the change of apportionment and the size of the house?
 * Given the Coalition do not exist, does this mean the Liberals have to gain the full volume of seats to win Government?
 * Based on all these I find myself wandering so far into OR that I think we're better avoiding the entire issue by not having an infobox. When the entire foundation of the electoral system changes, it's almost as if every seat's vacated and we start all over again, so the sense of seat gains is quite imaginary in some ways. Orderinchaos 07:40, 8 August 2008 (UTC)

My basis btw for the numbers is:


 * Status at 2005 election
 * Members of the Western Australian Legislative Assembly, 2005–2008
 * Parliament members list which notes affiliation (ALP, LIB, NPA or IND)

32 Labor members elected, including two Labor members who replaced two 2005 members at by-elections (Victoria Park and Peel) minus John D'Orazio, John Bowler, Bob Kucera = 29

18 Liberal members elected, including one Liberal member who replaced a 2005 member at a by-election (Murdoch) minus Paul Omodei, Dan Barron-Sullivan, Sue Walker = 15

5 Nationals elected, no changes = 5

2 Independents (Churchlands, Alfred Cove) elected, no changes = 2 + the 6 named above = 8

Orderinchaos 07:48, 8 August 2008 (UTC)

New boundaries
I've only seen one comprehensive look at the post-redistribution battleground, and that's this analysis done by Antony Green.

He gets 38 ALP, 16 Lib, 3 Nat, 2 Ind.

Green concedes it's somewhat arbitrary whether Blackwood-Stirling and Moore are classified as Liberal or National seats. So the Lib/Nat seats could be any of 15/4, 16/3 or 17/2.

If you can get away with it, I would just say 38 Labor to 21 non-Labor.

I wouldn't bother with the independents who weren't elected as such in 2005. Bush shep (talk) 15:31, 11 August 2008 (UTC)
 * Yeah... there's probably an entire field of political science to deal with this. :) The Nats are apparently confident of winning 5 and possibly a 6th - almost unbelievable based on the former boundaries, but *possible* (if not probable) under the new ones. They have chalked in Blackwood-Stirling, Central Wheatbelt, Wagin, and Moore with reasonable confidence in Albany and Geraldton and some chance at the new Eyre seat, and have the first four based on the numbers at present. Albany, Bunbury and Geraldton are usually a lottery at any election though - the "bellwether status" given to Bunbury was more luck and local factors that happened to coincide with state polls, and I've heard some *very* dodgy ABC analysis of Albany (which was Liberal held for a very long time until 2001 - their WA bureau person referred to it as a swinging seat). Orderinchaos 20:23, 11 August 2008 (UTC)

Margin of error
It isn't 2 percent. Per this link, click on "27/06/08:Quarterly reading of Western Australia voting intention and leaders’ ratings". The bottom of the PDF clearly states a margin of error of 3.5 percent. Timeshift (talk) 13:56, 7 August 2008 (UTC)
 * I think if you see one of these issues, just change it and note that you've changed it here. I'm going to be so busy with the actual electorates and MPs and historical stuff that I will only be coming past here to sort out issues with wording and the like (I try to avoid the hypercolour boxes anyway :). Orderinchaos 07:42, 8 August 2008 (UTC)


 * Thanks Timeshift. I knew I'd seen that link somewhere but couldn't find it again, which was why I changed to 2% per the prev link.  My bad. Moondyne 07:52, 8 August 2008 (UTC)

Just so everyone knows what I'm doing...
I'm going to write sections under Background for each of the parties (a para each for the 4 majors, a quick one covering all others), and a brief summary of issues which arose in the 2005-2008 term. Distinct lack of refs in the sections I've added will be rectified within the next few days. I don't know how the Campaign section will be managed, as on other articles (Victoria etc) they have gotten quite chaotic and tangled. Any ideas welcome :) Orderinchaos 10:17, 8 August 2008 (UTC)
 * Mate, keep doing what you're doing. Kudos for you (and unnamed others) for getting this thing and the districts and divisions moving.  If you cop any flack for anything being unfinished, or have any specific tasks need doing quickly, sing out. Moondyne 10:47, 8 August 2008 (UTC)

Public funding?
Someone may care to check this with the electoral office--I've had hearsay that there will be public funding for the first time in this election, with a 4% threshhold. Cheers Bjenks (talk) 18:01, 20 August 2008 (UTC)
 * As far as I'm aware, it isn't the case, but I will check. Orderinchaos 21:50, 20 August 2008 (UTC)
 * Hmm... Wow, that one slipped through the net! After some searching this morning (it's not on SLP), tracked down the Electoral Reform (Electoral Funding) Act 2006, which came into effect on 27 October 2006, introducing amendments into Part VI of the Electoral Act 1907, providing for electoral funding of political parties and candidates. I can't find the actual text of the Act, will try and get that in the next couple of days. Interestingly, there was also no media coverage of it, other than a few opinion pieces in mid-August 2006, none of which mentioned the amending legislation. Orderinchaos 22:15, 20 August 2008 (UTC)

For the text see Part VI, Division 2 of the Electoral Act (sections 175LA to 175LK) Jmount (talk) 22:21, 20 August 2008 (UTC)
 * Yep, it seems to centre on 175LD and 175LF. (As a note for future ref, the supply of legislation to Austlii by the SLP was inexplicably terminated a couple of years ago, but seems to have been resumed at the beginning of this month. It's worth noting it could stop at any future time.) The figure of $1.39413 per vote as at July 2006, increased each year with March quarter CPI, apparently came from the Queensland legislation. Orderinchaos 22:28, 20 August 2008 (UTC)
 * Also found this press release from August, while the bill was still being debated. Orderinchaos 22:31, 20 August 2008 (UTC)
 * This WAEC page says it's currently $1.56888 per vote. Orderinchaos 22:49, 20 August 2008 (UTC)
 * Added section to Electoral reforms per our discussion here. I'd like to thank Bjenks for bringing this to our attention, especially after searching the media I'm amazed this one slipped under the radar. Orderinchaos 13:19, 21 August 2008 (UTC)

My edit summary - an explanation
Hi. This is an apology in advance if previous editors of this article might be offended by my comment (in my edit summary) about lack of citations. Having looked at the article again I can't actually point to anything significant that is crying out for a citation. So - please ignore my comment. I probably came across as unnecessarily critical - though that wasn't my intention. My mention of citations was actually pretty stupid - especially when I look at the section on Electoral Changes - which I left well alone. That section is solid and also very well documented. In any event, if I needed to make a comment it should have gone on this talk page not in the edit summary! Its a habit I need to break!--Glen Dillon (talk) 12:02, 23 August 2008 (UTC)
 * No worries, we badly need constructive critics! If you do see something without a citation which most likely needs one, feel free to drop a note - sometimes one can get too close to one's own writing and fail to spot things a neutral observer would pick up. Orderinchaos 01:42, 24 August 2008 (UTC)

"government do not have 38 seats, so this is ridiculous"
Rather than insults, how about we refer to the reference? Timeshift (talk) 01:28, 24 August 2008 (UTC)
 * It's not an insult - it's a fact. The site quoted is not a reference for the Government having 38 seats - it's basically one person's educated opinion rather than fact. Blogs are generally not considered reliable sources, although we have certain exceptions to allow for information which is in effect peer-reviewed, such as Green's analysis of the electoral system. Also, Mackerras says they have 36, and I think his factual basis for doing so is considerably stronger based on the actual figures (my understanding is that the two men are not in disagreement, they just have slightly different models). When the election is called, the difference in seats from last election to this election is all that will matter, even though "nominally" they're now in a different position. It comes down to whether we are an opinion blog or an encyclopaedia, and I prefer the latter approach. Orderinchaos 01:36, 24 August 2008 (UTC)
 * Actually, Mackerras also has 38 Labor seats. Bush shep (talk) 15:59, 24 August 2008 (UTC)

Pictures
I'll write the parties section and some campaign notes in the next couple of days, but does anyone have any pictures worthy of use? I got a few at the Labor launch today (Penny Wong, Alan Carpenter, some MPs and a generic shot of the launch) but the quality is not great due to the darkness of the venue. It would be nice to have one from the Liberal launch too, as it'd be nice to see breaks from the text :) Orderinchaos 05:53, 31 August 2008 (UTC)

Hmm...
Storing here so I remember where I found it. Orderinchaos 10:48, 2 September 2008 (UTC)
 * The West's coverage of elections has been bad since the WA Inc days (of which the paper was strangely supportive). Writers with absolutely no political acumen (like self-styled funnyman Robert Taylor*) have been on the job for decades&mdash;not a great service to readers. This time, we have rampant anti-Carpenter bias (maybe because even Carps was a better journo than all of them rolled into one&mdash;van Onselen included). *(re Taylor, aka Sketch) Remember the old gag: "He thinks he's quite a wit, but he's only half-right". Cheers Bjenks (talk) 14:45, 2 September 2008 (UTC)
 * Mind you, after last night I'm beginning to wonder about the ABC's. The count told some of us so little that it was necessary to travel home and hit the WAEC website for information, while the evening seemed to be used by all sorts of odd former members, candidates and whatever else as a soapbox. Furthermore, they seem to implicitly treat the Libs and Nats as a coalition when they are not. Today they're saying the Nationals will hold the balance of power, when any sober analysis reveals the numbers don't strain that way. I think we're looking at 28 or 29 Labor seats, 22 or 23 Liberal, 4 National, 4 Independent, so it's actually John Bowler and Janet Woollard who will hold the real balance of power (as the two non ex-Liberal independents.) A minority government led by a Labor premier other than Carpenter now appears to be a likely outcome, but the Upper House is going to be hell - more Greens have been elected (5 or 6!), but as with their victory in 2004 fed, either the Libs and Nats will together hold a majority in the upper house, or hold exactly half with the Labor and Greens holding half as well. Therefore, any legislation to be passed will need the support of every or almost every party and at least two independents no matter who governs in the Lower House, and I can only see a fresh election 12-18 months away as being a solution. I freely admit by the way that I had predicted 31 seats and that the results in Southern River and North West screwed that up. At least one Labor win was a surprise to me, and it and three others I'd regarded as steadily uncertain cancelled out my other faulty predictions, although I hadn't put any of those in the "certain" column. Orderinchaos 04:43, 7 September 2008 (UTC)

More link storage. Orderinchaos 13:39, 14 September 2008 (UTC)

Polling hours?
Polling hours are presumably the same as Federal elections, 8:00AM to 6:00PM ? Tabletop (talk) 09:38, 6 September 2008 (UTC)
 * Yes that's correct. Orderinchaos 14:10, 6 September 2008 (UTC)

Formal vs Informal
Until last night I was confident that 1-2-blank-blank or 1-2-2-2 would be an informal vote in the WA state election, as it is in a federal election. This understanding appeared to be confirmed by the electoral commissioner in an ABC radio broadcast t'other day. However, I have learned from scrutineers at the count that such votes were being accepted as formal, backed by a presiding officer's manual, and without challenge. I wish the electoral commissioner had given a straight and honest answer about that. It would have prevented me from misinforming other voters and from myself casting an effective (preferential) vote for a candidate whom I would have much rather blanked. How many other voters were duped into voting for unwanted parties, I wonder? Cheers Bjenks (talk) 01:55, 7 September 2008 (UTC)
 * 1-2-3-3-4-5-6 is formal. The so-called Langer vote. I have seen 1-2-2-2-2-2 rejected as informal though, and 1-2-blank would definitely be discounted as an informal vote as it fails the legislative requirement that all squares must be numbered. There is a certain number of votes which show up as "exhausted" in 2PP counts, I believe One Nation were key in getting the message out there a few elections ago. The most by far in the 2005 election was in Bassendean where 179 votes didn't go into the count. Orderinchaos 04:26, 7 September 2008 (UTC)
 * People can correct me if I'm wrong, but it appears to be section 140A of the act, which was added by an amendment in 1996. Subsections 1b and 1d appear to allow a vote where all but the final number is nominated, Subsection 2 is a mangle of words but appears to allow a repeated numeral, so long as section 128 has been followed ("by placing the numeral “1” in the square opposite the name of the candidate for whom he votes as his first preference and consecutive numerals beginning with the numeral “2” in the squares opposite the names of the remaining candidates so as to indicate the order of his preference for all candidates") Orderinchaos 04:37, 7 September 2008 (UTC)

Here's the simplest you can put the LA ballot formality rule: all of the candidates (or all but one of the candidates) must be numbered, and one (and only one) of those candidates must be marked as the first preference. Where there is a break in the chain of preferences (like encountering a skipped number or duplicate numbers) the ballot is set aside as exhausted and not distributed any further. You also don't have to number the boxes - you can go "A-B-C-D" as well, or write "first preference" beside their name or whatever, so long as the order of preferences is made clear. At my training session I asked why "1-2-2-2" is formal when putting "1" and leaving the rest blank is not; they couldn't give me a satisfactory answer.

In the upper house, if someone has put a "1" on the left, and also fully numbered the right hand side, the right hand side takes precedence. But if they number both sides and one side is done wrong, then the other side kicks in as formal. They can also fully rank the ticket/party boxes on the left: you just ignore all the numbers except "1". Ticks and crosses are also formal on the left hand side instead of a "1". - Mark 10:46, 7 September 2008 (UTC)
 * Thanks for that :) Answers a few questions. Orderinchaos 10:54, 7 September 2008 (UTC)
 * Hmm. So I have to suspect that one presiding officer might (correctly) rule 1-2-blank-blank informal while another elsewhere might admit such votes to the count. And to the formal-informal dichotomy, the WAEC adds another category--"early exhaustible=formal". I would like to see the written instructions which guide presiding officers in such matters. I would also like to know why the electoral commissioner publicly (on radio) declined to agree that 1-2-3-3 would be accepted as a formal vote. Perhaps the commission should be asked to provide a detailed analysis and explanation of (a) informal and (b) early-exhausting formal votes since the public was misinformed just prior to this very crucial election. Bjenks (talk) 14:27, 7 September 2008 (UTC)
 * 1-2-blank wouldn't be admitted under any circumstances as it's a breach of the Act. I'd say you're more likely to have officers if anything unintentionally applying the Federal legislation and ruling Langer votes informal when in fact in WA they are formal. Keep in mind it goes through entirely another process at the counting centre so irregularities between processes used to determine informal votes in the booth counts are ironed out at that stage - in an election I ran in, the final count was quite different to the night count. As for "misinformed", I think there's a situation where there is a proper way to vote and have it counted, and a loophole which allows a different method, although apparently not intentionally (it meets the requirements of the law, but the requirements were not drafted with the situation in mind). The 2001 election had thousands of such ballots, with a significantly reduced number in 2005. It's not the commission's job to publicise every possible loophole but to tell people how to vote. Orderinchaos 15:04, 7 September 2008 (UTC)

Reading all this above here make me not want admit i was working yesterday in any capacity - my manual has gone back in the trailer :( SatuSuro 14:35, 7 September 2008 (UTC)

Results
The online results are somewhat unreliable. The way they are copied into this article make it seem they are far less ambigious than they actually are. For example a lot more than 2 seats are still in doubt. One of the problems seems to be that no reliable prefernece counts are being provided. Have a read of http://blogs.abc.net.au/antonygreen/2008/09/updating-wester.html Jmount (talk) 07:09, 7 September 2008 (UTC)
 * Agreed, but I think the intention is to update as news becomes available. A note above them indicating they are anything but final would be useful I think - my opinion is that Morley, Eyre and Wanneroo are much more in play than is being acknowledged, and possibly several others. Orderinchaos 08:27, 7 September 2008 (UTC)


 * Can someone in the know add a quick summary about how the elections are counted and just what the source of information for provisional results are? Timrollpickering (talk) 11:42, 7 September 2008 (UTC)


 * Issuing officers first unfold and sort the Lesiglative Assembly ballot papers into their stated first preferences, and any questionable or objected-to ballots get decided upon in relation to formality by the presiding officer (Polling Place Manager). Next, the presiding officer sets aside the first preferences for two candidates most likely to win (according to the Electoral Commissioner; usually Liberal & Labor), and takes a minor candidate's first preferences and looks to which of the two main candidates is preferred. After doing this for all of a minor candidate's ballots, they are tallied, recorded, and returned to their pile as that candidate's first preferences. This is repeated for all remaining minor candidates. Finally, all of the Legislative Council ballots are sorted by the issuing officers into their first preferences (whether that first preference be a party ticket or an individual candidate on the right hand side). This means there's gazillions of piles of these ballots. The first preferences are tallied. After each of these stages, the results get called through to the returning officer who passes it on to the electoral commission. Finally, the polling booth is packed up and bags filled with votes and materials, sealed and taken to the returning officer (often around 10pm or later).
 * On the Sunday the returning officer checks through the materials to make sure everything is hunky dory, signed and everything. Then the returning officer whacks it all in the car and several trailers and drives it in to the electoral commission. Monday morning at 8am, staff at the electoral commission crack open all the declaration envelopes (absent and provisional ballots) after checking the voter was enrolled. This is why voting always seems to stall at around 75% after election night: it doesn't mean that a quarter of polling booths gave up and stopped counting for the night. The provisional vote count will probably be quite high this election because of a lot of problems with the electoral rolls and people enrolling at the last minute, due to the short campaign. I'm a bit cloudy on what goes on at the electoral commission though, so I can't really help you there. - Mark 12:59, 7 September 2008 (UTC)

Having been working at a booth yesterday I can vouch that most of the above is true :( SatuSuro 13:06, 7 September 2008 (UTC)


 * We spent more time at my booth unfolding the ballot papers than we spent actually counting them. It's a real time waster. :-\ - Mark 13:22, 7 September 2008 (UTC)


 * I am not sharing some of the hairier aspects of where things go wrong on this page - but the ideal situation as outlined above seems almost too good to be true - (it does happen that smoothly at most booths, but...) - some federal and state situations vary, and the point about provisionals - from the electorates that I know about may well see something brought up about that, either in disputes or court cases if either side do a bush and send in heavies on re-counts etc. At the commission - it all gets re-counted and checked and archived. SatuSuro 13:35, 7 September 2008 (UTC)

Legislative Council ballots are then counted (although depending on internal arrangements inside each polling place, the LC count might start before the LA is 'wrapped up' for the night. I'm not sure of this, but my best guess is that the LC ballots are "notionally counted" - that is they are counted sufficiently to give a snapshot of probable outcomes for each region. I'd guess that they are first sorted into three groups: ticket, non-ticket and obviously informal.  The papers where the voter has chosen a ticket-vote (left hand side) are then sorted by party and a count obtained (for each party) and this would be then phoned through to the tally room by the polling place manager then entered into the WAEC database.  The number crunching at this preliminary phase is pretty straight forward.  A quick and dirty "first-pass" result for the upper house could be obtained just by applying the pre-ordained ticket algorithm to the prelim. ticket-vote results phoned through on the night. In centres with small turn-out, its all over by 9:00-9:30pm - everthing packed-up - ballots bundled into bags (which are then sealed and immediately driven to each District Returning Officer on Saturday night). He/she delivers all ballot papers to counting HQ on Sunday morning. From there I'd guess that whatever counting is required (meaning full dist. of preferences for LA and a comprehensive re-count of all LC ballot papers) starts on Sunday arvo or Monday - with preference given to those districts where the count is close (ie result is very much in doubt such as Riverton/Alfred Cove for example). (...hope this long-winded explanation answers your question...I'm by no means expert in these matters, although just out of curiosity, this year I signed up for temp. employment with the WAEC and scored a gig yesterday as a Declaration Vote Issuing Officer taing care of absentee & provisional votes at a polling place. (Still don't quite know why I did it...certainly wasn't for the money...but after 20yrs working in an industry very far removed from the public sector, I can now at least say that I was a govt. employee for 14 hours...I found the whole thing interesting to be honest and might even do it again in 4yrs time) The reason why I can't be definitive about some aspects in my answer is because I didn't participate in the counting process last night. I was in a different room from 6pm till around 9:00pm dealing with Absentee and Provisonal ballot papers - well sealed envelopes CONTAINING absentee and provisonal votes to be precise. At close of polls at 6pm, temporary polling officials (like me) dealing with Abs & Prov votes have a fair bit of paper-work to do.  For those issuing ordinary votes, balancing the books is easy cause there's ony two types - 1 LA and 1 LC.  Accounting for each individual ballot paper is easy. Basically total papers issued to you on Sat morning = total unused + total spoilt/discarded + total unused + total contents of ballot box. EASY. By 6:30pm this basic equation is finalised and confirmed. For Absentee and Provisional (ie Declaration votes) it takes longer because we were each issued with ballot papers for the whole state (most of which end up being unused) - lets say for example 40 ballot papers for each of 59 LA Districts and say 100 papers for each of the 6 regions. 'Balancing the books' just takes longer because the accounting is a bit cumbursome A) because its all done manually (unbelievable though this may seem) and B) its on a District and Region basis. Also - while  reconciling papers issued with papers left over/spoilt/discarded is straightforward - we don't have the actual votes to lok at - they are of course in sealed envelopes and are not opened till probably tomorrow (Monday). In fact the envelopes are not even opened until the legitimacy of each envelope (= the individual voters elegibility to cast a vote) is determined and confirmed by reference to numerous sources - such as the marked-off rolls for the entire state, address checks etc etc. Sealed declcaration vote envelopes are emptied from separate ballot boxes by Dec. officers (like me), counted, grouped by LA District (59 of these). Envelopes for each district then sorted alphabetically by surname of the voter, then bundled (with elastic bands) counted again, a label attached then put into a large bag and offically sealed. Absentee and Provisional votes, by definition, are DECLARATION VOTES - where the voter DECLARES before me and signs on the dotted line that he/she IS ELIGIBLE to vote. The WAEC confirms this eligibility during the course of the following week. Thats why Absentee/Provisional counting can take sometimes two weeks to finalise, despite being only 5% of total votes. Dealing with ABS/PROV votes, as far as the WA Electoral Commission is concerned, is a problemmatic area. (this is only my opinion). Apart from a voter claiming to be eligible but who turns out not not to be for whatever reason, some absentee/provisional votes are rejected because the Declaration Issuing Officer has stuffed up in some way - through carelessness (eg. giving the voter the ballot paper for the wrong district) or just plain stupidity (like not asking the voter the right sequence of questions etc etc). As for me, even though I worked as a Declaration officer on Saturday just for "something different", I took the whole thing very seriously & managed to get my "books balanced" perfectly at the end of the night. I believe I made two errors during the course of the day, and thoroughly enjoyed interacting with hundreds of people from all walks of life that I would never have otherwise encountered, and I learned a lot about the election process. I intended to provide a brief answer but I've now done an essay-length expose out of all proportion to what the  person asking the question intended. I've just noticed that you only asked for a SUMMARY! I'll crunch this down to 100 words and post on the article page tomorrow morning. Apologies! --Glen Dillon (talk) 15:34, 7 September 2008 (UTC)
 * took me so long to answer the question - others have snuck in and answered far more succinctly... anyway...for what its worth: here goes...at precisely 6:00pm, the board in the main tally room (was it in one of the big sheds at Fremantle wharf?) kicks off immediately with the pre-loaded results of early-in-person and early-by-post votes (pretty small proportion - prob 3%?). Menawhile, back at each of the 700 or so polling places across the state, within 5 minutes of doors closing, the counting commences.  Ballot boxes containing ordinary votes (which is 90% of all votes) are unsealed and ripped open.  Each polling place conducts a count of lower house votes.  If required (and in almost all cases it is) a distribution of preferences is conducted.  There may not be a comprehensive distribution of all preferences on the night for a number of reasons (insufficient staff available/it looks like it'll go down to the wire and there are ten candidates in the seat etc).  In all cases, however, a notional distribution of preferences is undertaken (I take this to mean dist. of first, maybe second pref's), and the results are phoned through to tally room by the mananger of each polling place (Scrutineers who have departed the polling place are often able to informally pass on info to candidates/party officials/even journo's if so inclined, even before the 'official' results are phoned through to central tally room).
 * What I wrote about early votes being pre-counted was a big assumption on my part. Reading page 23 of the 2005 election reportleads me to believe that while the eligibility of early votes (in person or by post) might be pre-confirmed prior to election day, the envelopes aren't opened until after the close of polls on Saturday. So although it would be possible to have these votes already counted, I guess the potential for leaking of results prior to polling day would make it a bad idea - especially this year given that an unprecedented 7.5% of the population voted prior to 6 September  Glen Dillon (talk) 18:06, 7 September 2008 (UTC)
 * Just to elaborate on a couple of points you made, a full distribution of Legislative Assembly preferences is never done on the night at the polling place: only a notional distribution of preferences (if requested by the Electoral Commissioner for that electorate) - this means taking Greens etc. first preference ballots and seeing which of the two candidates are ranked higher. The full one doesn't get done until they go in to the WAEC offices. Also, the tally room this time was at the ABC studios in East Perth. - Mark 02:26, 8 September 2008 (UTC)
 * As a signed off PPM I would certainly take issue with lots of small points everywhere above but hey it would be nitpicking - not necessary - there are always disasters at certain booths and with certain issues - I was lucky enough to have a clean bill of health for my report but i was employed less than 5 days before the election due to a no-shower. I would say that the AEC and the WAEC have different standards and my AEC training from previous elections put me in a better perspective. Some of glens guesses are not far from the mark, but the issues of provisionals is one that may or may not come out during the week or later or find itself filed away in the annals of the cockups of the methods of the waec and its regulations re validity of enrolment. SatuSuro 02:44, 8 September 2008 (UTC)
 * At the end of all this we are probably going to have the basis for an article about the mechanics of the electoral process at at least C class :P Orderinchaos 02:58, 8 September 2008 (UTC)
 * I accept that my huge spiel above was off the mark in a number of areas. After realising there was an edit conflict and others had answered succinctly in the meantime I nearly hit the delete key but saved it on the reasoning that this is a talk page. I hope the new article section added this morning hits the spot ok.--Glen Dillon (talk) 03:38, 8 September 2008 (UTC)
 * You're fine :) We're all non experts here and I think across the breadth of us we'll get it right, with each being able to correct other narratives if need be. Orderinchaos 06:47, 8 September 2008 (UTC)

"Secondly, only the 'ticket votes' for the Legislative Council are counted on election night" -- that's not entirely accurate. Ticket votes are counted, as are first (non-ticket) preferences for specific candidates. - Mark 06:59, 8 September 2008 (UTC)
 * Good pick-up Mark and I've fixed the text accordingly, though you'll see I've written "in some cases" because according to 2005 report there were some procedural variations between the regions on election night which I assume would apply to how the preliminary LC count is done. Hope you're ok with that.--Glen Dillon (talk) 07:35, 8 September 2008 (UTC)
 * Good work :) - Mark 07:47, 8 September 2008 (UTC)
 * Electronic counting of upper house ballot papers? Had nothing definitive re:2008 so I got that bit from the 2005 report; it was just an assumption because its hard to imagine the WAEC reverting to manual counting. I'm not so sure now after seeing news footage of the Freo count today. Every close-up shot showed real people examining the big pink sheets. In 2005 the scanning was apparently done off-site at the contractor's premises with WAEC staff on hand to make rulings when the software couldn't decide re:informals. Anyone know what the deal is this year?--Glen Dillon (talk) 09:32, 8 September 2008 (UTC)
 * A source that i have been asking things is a boss just slightly up the food chain - they spend less each election compared to previous (reducing staff where in the past they helped get some results quicker) - so I would suggest that the compulsory recount on completed items next weekend is person based from what I have heard so far - but dont quote me! some info i have been getting is contradictory :) I was always under the impression that OIO names lists were the only items scanned - its a new one to me that they do it with actual ballot papers SatuSuro 09:51, 8 September 2008 (UTC)

Question re related article
Re: List of Western Australian Legislative Assembly elections - someone removed the colour for the total seats field. Do you think it should remain colourless, have a colour but maybe not light green, or simply go back to the way it was? I changed it in the first place to make it more readable rather than a mass of figures, but am open to ideas. Orderinchaos 06:46, 8 September 2008 (UTC)
 * I've been trying to see how it looked with colour. With the previous version, at least on my screen, that particular column had black text against black background which might explain why the other editor made the change. I'm not up to speed on the technical aspects of colour formatting, but is it possible that your monitor was displaying differently? Having said that, and seeing your other tables/charts, all of which are first class, I'd say go ahead and put the colour back in for the reasons you've suggested.--Glen Dillon (talk) 10:08, 8 September 2008 (UTC)

Graphic
I dunno if this goes without saying, but since we're certain to have a minority government here (unless we go back to the polls), it would be good to have a graphic like the one at the right to diagrammatically show the make-up of the chamber. - Mark 07:47, 8 September 2008 (UTC)

Tally Room
Would the Tally Room located at ABC's Perth studios refer to the old ABC studios, since the ABC has moved elsewhere? Tabletop (talk) 10:05, 8 September 2008 (UTC)
 * yep, fixed.--Glen Dillon (talk) 10:11, 8 September 2008 (UTC)

Subpages or new articles
Just looking thru theres appears to be a big area of unnecessary information that more suited to a seperate article the sections on the Electoral System and counting the votes are more generic detail that applies to many articles. Also reading the above discussion I think maybe people who worked on the day could be a little more cautious on some of the details, otherwise there may be concerns over OR or RL issues with the commission. Gnangarra 12:34, 8 September 2008 (UTC)
 * Just a matter of getting the publications off the website and a copy of the Electoral Regulations off SLP or Austlii and sourcing each point to them. Anything that can't be, we can decide whether to drop it or not. Re the electoral system, I think it was critical enough to explaining the once-in-a-lifetime nature of this election that it was necessary. Orderinchaos 12:36, 8 September 2008 (UTC)

Hmm - surely there have been elections of the past that have been as significant that could have had similar issues that would technically require similar detail? SatuSuro 13:02, 8 September 2008 (UTC)


 * Sep.article on electoral system & counting?-I agree. OR issues? perhaps yes on the talk page but article itself looks pretty well verified (I've cited additional documents which resolved issues raised by others in recent hours). Vote Counting will be a hot issue in the real world with this election so perhaps the lengthy treatment is justified but I don't hold any firm views on whether or not big cuts are required. I'll defer to the consensus view on that. Re: Past elections - no doubt SatuSuro is right. Perhaps the path forward starts with sep. article on elec. system & historical election-specific detail can spread from there to the articles for each election?--Glen Dillon (talk) 13:17, 8 September 2008 (UTC)

When orderinchaos is back from his well earned rest i think he might want to alert fellow wikipedians of his researches of earlier elections that may have required either lengthy count times or similar issues perhaps that might need identifying and elaborating upon SatuSuro 13:25, 8 September 2008 (UTC)
 * The big ones are 1901, 1911, 1933, and arguably 1989. 1901 noone actually won and it took 10 months to form a government stable enough to govern (only after several byelections and people switching sides in both directions), 1904 was almost identical to this one (i.e. 2008) except that Labor came from the crossbenches (they previously held 8 seats, 2 gained at byelections since 1901) to minority government with Independent support, only for the Independents to abandon them. 1911 was the biggest majority ever held by a government in WA (68% of seats). 1914 was interesting in that the government only had a one seat majority, which it then lost when an MP disappeared and his seat was won by a Liberal, and another one quit the Labor party for the crossbenches. 1924 was interesting due to a split in the Country Party which effectively brought Labor to power, 1933 was the first and so far only time a major party got reduced to minor party status in WA, 1947 was won by the Country Party but they allowed a Liberal to be Premier, 1953 and 1971 were one-seat majorities for Labor - the first time they lost it, but with an even number of seats (50, not 57/59 as per recently) they were able to hang on, while the second they narrowly maintained it with the Balcatta by-election of 1973. 1989 was a case where they won a minority of votes but a majority of seats (31 of 57), although went into minority government during the term because of three defections, and a lost by-election in Geraldton. The 1989 election was also critical for the complete redesign of the Legislative Council. 1996 was unusual in that Labor gained votes yet lost 1/4 of its seats - then 2001 was the record number of seat changes in a single election, as well as the best performance by a minor party until this election (it appears the Greens have easily beaten One Nation's 2001 vote). Ones involving very major changes in boundaries were 1911, 1930, 1950, 1983 and 1989. (1996 just renamed many of the 1989 ones while 2005 had little impact outside of certain defined regions.) Curiously, 1911, 1930 and 1983 were lost by the government implementing them (going further back, so were 1901 and 1904.) This is all off memory as I'm typing quickly. Oh and a complete review of every redistribution ever held in WA, although incomplete after 1966 for the LC, is at User:Orderinchaos/Electorates - it's been checked and rechecked and re-re-checked against different sources, I use it heavily as a reference. Orderinchaos 14:12, 8 September 2008 (UTC)

Well thats an article your post you have just done? SatuSuro 14:27, 8 September 2008 (UTC)

Results 2
Interesting - the postals being counted are not following the traditional pattern where you can say 0.7% go from Labor to Liberal. At present, the Liberal and National votes are almost unchanged from election night counting, the CDP has grown by 0.03% at the expense of Family First, the Labor vote has slipped by under 0.1%, while the Greens have gone up 0.15% and independents have fallen about 0.2%. Curious to know what anyone who knows more than me about the procedural side of things has as an opinion on it. Orderinchaos 05:17, 10 September 2008 (UTC)
 * Not sure, and I'm not an insider but only suggestion I can offer is that the postal votes you've mentioned are maybe not strictly postals but could include absentee/provisional votes cast on Saturday - which were at unprecedented levels. I'd guess that preferences on these ballots would therfore follow the trends for ordinary votes, rather than having the standard 'postal vote' skewing (normally explained by socio-economic/demographic factors or whatever). In summary, the 2007 redistibution caught some electors by surprise, inadvertantly swelling the total numbers of absentee votes, which, when counted, end up looking much like ordinary votes in terms of voter preference.GlenDillon 05:36, 10 September 2008 (UTC) more thoughts...about 17.5% of all votes in 2005 were "not ordinary" (absent/postal/early/provisional). A rough guess-this year it could be 21% with all the extra absents. I'm not suggesting that people vote any differently when they are out of their districts, but some might, especially if irritated about it! This is a very unique election -at the macro level re:current state of play in terms of seats, and micro re:voter behaviour on polling day, the prelim Count, poss. delays in the official count etc most of which can be linked, in some sense to the 2007 reditribution. GlenDillon 06:41, 10 September 2008 (UTC)
 * You're quite likely right. 106,000 132,000 additional valid votes have been counted since I first updated the tally mid-Sunday (although it says 959,000 approx, the actual figures added to 913,011 and this was an error they shortly later corrected). That represents almost exactly 8% of enrolled voters. My guess is turnout will be down on last time due to voter apathy/"pox on both your houses"/electoral redistribution causing confusion etc. The decrease in informal voting, despite the lack of clear issues to vote on, is probably an indicator of this as well. (This isn't a reply, it's as much me getting my thoughts together!) Orderinchaos 06:51, 10 September 2008 (UTC)
 * Where are you pulling the number of informal votes from? If it's from the WAEC website I'd be careful... - Mark 07:15, 10 September 2008 (UTC)
 * I was told by WAEC to subtract total valid votes from total votes, rather than rely on their own informal figure. All figures are provisional anyway (as has been seen in the past hour when 600+ ballots disappeared from the count!) Orderinchaos 07:19, 10 September 2008 (UTC)


 * From my booth perspective - I'd say there might be potential changes regarding (a) provisional votes disallowed (my dec wasnt trained sufficiently to discern whether a provisional had a hope in hell) (b) informals being reviewed (we passed ours by our scrutes and they had no probs with what we put in there - but they were around only for the assembly) either way in or out (c) im interested to see where the council votes and prefs go as our perspective on sat night was so preliminary compared to what might happen SatuSuro 08:34, 10 September 2008 (UTC)
 * In case anyone's curious, no I don't have OCD. :P I'm updating the results regularly to recordably observe the patterns, it may be useful for some research I'm doing offline and the edit history is (as far as I know) citable per GFDL in any outside work I do on the subject. Orderinchaos 14:32, 10 September 2008 (UTC)
 * Not sure what you meant, so just spent 15 mins searching for a Wiki policy on OCD without success, then it dawned on me. Hey - even if you did, it's probably a useful trait for an editor to have! Just noticed that the text above your table still referred to the table as Sat. night's results so fixed it. All the good work you're doing might've be ignored otherwise.GlenDillon 15:23, 10 September 2008 (UTC)
 * Ah good point - thanks for noticing that. :) And yeah, I have some friends with it, it doesn't sound much fun. It probably seems funny to outsiders that we have a lot of technical stuff and not much on the actual campaign and the parties (something I intend to write once my current avalanche of uni assignments has been cleared) but it's going to be difficult given that it's event- rather than policy-driven. Orderinchaos 15:45, 10 September 2008 (UTC)
 * Sure there's more that can be added to article - campaign, media treatment, the shennanigans this week (and prob. next) over alliances/role of National Party etc, but while the election is still a 'happening thing' those sub-topics are possibly more the province of blogs, papers & electronic media than an encyclopedia...they also bring more contentious/NPOV issues to this space which take time/effort to manage. The article content is solid thanks to all editors & also many others who've been contributing comments/questions/suggestions via this Talkpage & I reckon it's already served as a very useful reference for many hundreds of readers. The extras you mention can be added in coming weeks with all the benefits of hindsight and availability of good verifiable source material.GlenDillon 16:34, 10 September 2008 (UTC)
 * Wow... 1,100 people a day visiting this page! (And yes, I solidly agree on all points above.) Orderinchaos 17:23, 10 September 2008 (UTC)

Although it affects no particular counts - wow, the Nats lost 0.1% of their vote and 993 votes this morning, and it seems the previous numbers didn't add up to the total provided for the last 3 updates yesterday (they do now). Orderinchaos 05:36, 11 September 2008 (UTC)
 * Unless something radically changes - and it hasn't all day - I think my musing about the turnout earlier will prove to be correct. Orderinchaos 08:23, 12 September 2008 (UTC)

Grylls goes with Libs
Breaking news media conference on Sky News. Timeshift (talk) 03:33, 14 September 2008 (UTC)

Kingsley
Orderinchaos, in your edit summary you state, "9 is correct - it goes on the basis of them having had 37 nominal seats. there was an error in the ABC's calculation of kingsley."

Could you please substantiate this? I point out once again that both Antony Green and Malcolm Mackerras—who presumably calculated the new margins independently—had Kingsley on the Labor side of the pendlum and therefore gave Labor a starting point of 38 seats. 

--Bush shep (talk) 14:08, 14 September 2008 (UTC)
 * By 48 votes. It made an assumption based on what happens if you take the CFP vote out of the equation, but didn't reflect the *actual* CFP vote and preference distribution in the bit which stayed and the bit which went. I was relying on both Black and Phillips for that assessment, can't find it in print though. Orderinchaos 14:12, 14 September 2008 (UTC)
 * It shouldn't matter if its 48 votes or 4800 votes. This sounds to me like OR. What is CFP? Bush shep (talk) 14:17, 14 September 2008 (UTC)
 * Community First Party (Marie Evans, wife of former Liberal MHR for the area Richard Evans) which got 11.5% at the election - . Not OR at all, it reflects the parlous state of most of the sites we use as a basis for analysis and the fact they almost never get updated (ABC is STILL stuck on Thursday night's figures, for instance, and the document we're using as a basis for this hasn't been modified since it was first written in 2007.) Orderinchaos 14:24, 14 September 2008 (UTC)
 * I strongly disagree. It has never been standard practice to recalculate seat margins by readjusting preference flows. Not least because it involves some very brave assumptions. (How can we possibly know what the result would have been in the absence of a Community First candidate?) If this standard was adopted consistently, you'd be putting a lot of 2001 Labor seats back in the Liberal column ahead of 2005 because of One Nation preferences. (e.g. Albany, Bunbury, Geraldton...) Bush shep (talk) 14:32, 14 September 2008 (UTC)
 * There was no assumptions made. Black and Phillips had the actual figures from the WAEC at booth level. The assumptions made by Green (and presumably Mackerras) placed Kingsley on the Labor side by a wafer-thin margin assuming that each booth had exactly the same proportion of ALP/LIB preference from the CFP, which never happens no matter what electorate you're looking at. The other seats have nothing to do with it because One Nation was not a factor in 2005. In reality, anyway, redistribution effects are a very inexact science as they always go off booth results only (out of necessity) when a growing percentage of people vote postal/prepoll/absent and their votes have certain key differences (not always the same in each seat) to the booth votes which may affect any notional calculation, especially if it's ultra-close. Orderinchaos 14:47, 14 September 2008 (UTC)
 * This just isn't right. Firstly, there were no assumptions made about preferences flows. None. Green and Mackerras operate strictly on the two-party preferred vote. It is you who is assuming Labor would not have received preferences that they did in fact receive. Secondly, I wasn't saying One Nation was a factor in 2005. I said One Nation was a factor in 2001 that, using your approach, would need to be corrected for in analysing the 2005 results. (N.B. I am not advocating this.) You are correct that post-redistribution calculation is an inexact science. You are however wrong about declaration votes; they do have a way of estimating where the declaration votes came from. This is all the more reason to defer to the esteemed psephologists, rather than forming our own judgements. Furthermore, the factor you cite isn't to do with the redistribution; its to do with preference flows. Bush shep (talk) 17:32, 14 September 2008 (UTC)
 * As it would be illegal for the WAEC to reveal sufficient information about postals and counts to any psephologist to make the sort of analysis of declaration votes that you describe, and I believe not even WAEC link the individual declarations to the individual votes once they are added to the count, I'm genuinely curious as to how you believe they would estimate where declaration votes come from. Additionally, I don't "esteem" anyone more than their professional capabilities permit - noone is perfect, and indeed they have made mistakes before and would readily admit to that. I should probably note, too, that I am not without qualification in the field myself, and have been studying the geographic implications of voting since around 2005 and have hit (and discussed with the relevant authorities) many of the aforementioned limitations on such analysis. Orderinchaos 17:59, 14 September 2008 (UTC)

Of course the WAEC does not provide details of postal voters, that's why I said it was an estimate. However, your claim was that they work off "booth votes only." Not true. Declaration votes are mapped to booth votes via a straightforward calculation, and are therefore included in the new margins. You can see an example of such calculations in this paper p4 (p10 in adobe).

As for Green and Mackerras, I don't why you think calculating post-redistribution margins would be "more than their professional capabilities permit." In any case, they arrived at the same figure, which surely confirms that there is no mistake.

Finally, could you please provide a cite for the Black/Phillips margins? By the sound of things, their method involves more assumptions than Green/Mackerras, and is therefore less scientific.

--Bush shep (talk) 19:00, 14 September 2008 (UTC)
 * p10 of that document is *exactly what I meant* about assumptive reasoning. It's just simply a weighted average, and doesn't (actually, can't) take into account the actual nature of the votes, so simply assumes they all fit a prescribed pattern. I'm well aware of that technique - I use it myself when working federal electorates into state ones and vice versa for comparatives in my own research. But it doesn't actually tell you what the votes were, and that's where it falls apart. There's no mathematics to it, it is basically a good educated guess. My comment you picked up on can basically be interpreted to mean "they are not mind readers, nor do they have access to the individual votes mapped per booth/suburb etc." Re the Black/Phillips comments, I assume they have not yet been published given the election is only just over. I'm still waiting for the Jan-June political chronicle to be published, that occurs in December 2008. The July-Dec one will come out in June 2009. Unless they write a peer reviewed journal article before then (which is where both authors tend to be presented most), I assume that it will be a while before any meaningful analysis emerges of these sorts of factors, and I'd put money on this rather unusual electoral result pushing someone into researching it as soon as the WAEC make the full distribution figures available (that hasn't happened yet either - in fact, most of the figures will not be available until the Election Report is produced in, I'd guess, March next year). This is why writing about a lot of these things as they happen seems like a great idea at the time but leads to a lot of guesswork and very patchy analysis. One big reason why we haven't generally proceeded with an analysis of the election in this article (quite apart from the parlous state of the WA media, scant coverage on the ABC and eastern papers, and an ill-timed Fairfax strike which deprived us of facilities of the one paper - the AFR - which has always provided reasonably astute analysis of WA elections since it started covering them.) In the meantime I think we should be very careful how we handle these sorts of issues. Anyway, I have an assignment to do, so I'll be offline until this evening. Orderinchaos 19:20, 14 September 2008 (UTC)


 * A few things:
 * 1. How else are they supposed to map the declaration votes? You yourself rubbished the idea of getting more a accurate figure for declaration votes. You can't claim this to be a weakness of Green's and Mackerras's methods when Phillips's and Black's analysis will suffer from the very same thing. Therefore, its a red herring.
 * 2. I never claimed that Green's and Mackerras's methods didn't contain assumptions. Of course they do, they HAVE to. The key point however is that they involve as few assumptions as is necessary.
 * 3. You attacked Green's figures for not taking into account the Community First vote. But to Green (and Mackerras) its irrelevant; they operate strictly on the two party preferred vote. What you're actually demanding is more assumptions, not fewer.
 * 4. Without a cite, the 37 seats claim is unsourced.
 * --Bush shep (talk) 19:32, 14 September 2008 (UTC)
 * I can't believe we have just wasted two screens debating a single figure. My point stands, and I'd much rather spend my limited Wiki time improving the article instead of engaging in pointless debate over questionable sources. As far as I am concerned, Green's and Mackerras's estimates, while generally on the mark, on close calls like Kingsley are simply one person's opinion, and I've seen reasonable evidence that they are not infallible opinions. Neither individual would claim such a high stand, either - they're both pretty professional about things and willing to admit their limitations, from my experience. (Green has told me his website is basically a well-informed blog and has given me some indication as to the time pressures involved in creating it and the relative inability to go back and check/fix things after they're in because of the need to move onto other things.) If a more detailed rationale was provided by them on this one seat, I might well concede the point, but I don't see the point in depending so tightly on erroneous non-peer-reviewed sources (see WP:RS) simply because they're able to publish on a tighter timeframe on a website, and that the genuinely reliable sources which depend on appropriate fact-checking processes take longer. Orderinchaos 20:06, 14 September 2008 (UTC)
 * No doubt Green's and Mackerras's 0.1% is well within the margin of error of these things. But that's not the point. A 50/50 seat is still going to be one or the other, and their calculations place it on the Labor side. It is arbitrary to simply flip the seat from the Labor column to the Liberal column. When the only two published sources (unpublished sources are in violation of WP:RS) have it as a Labor seat, we should do likewise. Bush shep (talk) 20:17, 14 September 2008 (UTC)
 * I disagree, they're not peer reviewed sources, and it's not 0.1%, it's 0.03%. But you know my opinion, I know yours, and we've done it to death. Orderinchaos 20:36, 14 September 2008 (UTC)
 * It could be 0.000003%. Doesn't matter. I addressed this point in my previous reply. Secondly, a peer review is a desirable but not an essential facet of a reliable source. A published source IS essential. Thus Green and Mackerras trump Black and Phillips. These are wikipedia guidelines, not a matter of opinion. Bush shep (talk) 20:48, 14 September 2008 (UTC)

Orderinchaos says: "I was relying on both Black and Phillips for that assessment, can't find it in print though." That's because it doesn't exist. David Black and Harry Phillips didn't do any calculations. Every pronouncement they made about the post-redistribution margins and seat totals, including those about Kingsley, came directly from Antony Green's determination - which was not a "blog post", but a paper published by the Western Australian Parliamentary Library.William Bowe (talk) 11:05, 16 April 2009 (UTC)
 * This is wrong too: "Black and Phillips had the actual figures from the WAEC at booth level. The assumptions made by Green (and presumably Mackerras) placed Kingsley on the Labor side by a wafer-thin margin assuming that each booth had exactly the same proportion of ALP/LIB preference from the CFP, which never happens no matter what electorate you're looking at." The Statistical Returns published by the WAEC after the 2005 elections provided individual two-party results for EACH BOOTH, and it was these figures which Antony Green used for his calculations.William Bowe (talk) 14:11, 16 April 2009 (UTC)
 * I was loathe to re-ignite this old debate, but it looks like Antony Green and William Bowe have done that for me. William's points only strengthen my insistence that Kingsley ought to be considered a Labor loss. Digestible (nee Bush shep) 13:25, 17 April 2009 (UTC)
 * OK let's revisit this to show why it cannot be determined to the significant figure level that would produce a 0.03%:
 * There was a protest candidate, the former wife of the two-term former Liberal candidate for the area, and the protest specifically was against the mode of appointment for the replacement Liberal candidate in a safe seat. A protest vote of 11.50%, whose precise split per booth has *never* been identified, was therefore generated. (All we know is that 2,859 votes plus 1,259 she received from other voters, in total, split about 55/45 for Labor, but that the difference was 447 votes and the Greens at their elimination had transferred 623 to her - so I'd assume it was about 50/50.)
 * The two Warwick booths and the Edgewater booth were removed from the electorate at the redistribution. At pure booth level, this meant a drop from 51.37% to 50.62%. There were 3,019 votes (of 21,146 booth votes total) cast at the two Warwick booths and Edgewater - 14.27%. (Green seems to have rounded this up to 15% in his analysis for the postals.)
 * There were 3,705 formal "other" votes cast with a Lib 2PP of 52.63%. The drop produced by the "other votes" in the 2005 results is from 51.37% to 50.77%, a drop of 0.6%.
 * The reported 2pp is 50.03. It appears to me that the way this was come to was to multiply the postal votes by 2005 votes in 2008 area/2005 votes in total area (to get 3176), take 0.65% off the postal vote for Labor (i.e. LIB 53.38%) and feed the two in together. This produces 10651 vs 10642 - 9 votes difference, or 0.14% at postals level alone. i.e. If the postals had have been LIB 53.52% instead of LIB 53.38% - and keeping in mind the entire basis for calculating the 53.38 was a guess - the seat would be on the LIB side of the ledger. (My error of 48 above was based on a calculation error - I still have the figures I used to make that statement.)
 * Argh - a calculation error, but one I'm going to leave in as it is an interesting point. "This meant a drop from 51.37% to 50.62%" = 0.75%. Plug that result into the postals and what do you get - 10647 to 10646. ONE vote. Even the difference in exhausted votes (6, down from 10) is greater. Orderinchaos 00:40, 18 April 2009 (UTC)
 * The above assumes there is noone who lives in the 2005 area that is now excluded that voted at a 2008 area booth, and that everyone who lived in Greenwood voted in Greenwood (and not at the Warwick Church of Christ hall, which attracted 1,140 valid votes and is located immediately behind Centro Warwick, the electorate's lone shopping centre of any size.) Long term trends at both federal and state show that Warwick and Edgewater are both more Labor than the regional average.
 * The above also ignores attrition rates, or people moving out of the area (or this world, arguably). While many areas in the northern suburbs are growing, the ABS tells us that Greenwood, Kingsley and Warwick have an attrition rate of 0.48% which is consistent over the last 15 years at censuses. Most likely factors would tend to suggest Labor voters leaving the area in small numbers, with Liberal ones staying put. I have *no* survey data at all to base this on, but I am not trying to prove a point, I am trying to demonstrate a level of uncertainty.
 * My summary - it's fine for a psephology website, but there are so many factors involved and, in this case, such a TINY margin of error ( 9 votes 1 vote out of 21303!!) that Wikipedia should not be making definitive statements on it. Whoever may have published it, it is a *best guess by one expert*. If it's 5% in the balance, the basis used by Green and Mackerras to calculate the results is more than adequate. If it's 0.03%, the academically rigorous answer is "impossible to determine". I understand that the pressures on the research are such that every seat needs to have a stated result, but the difficulties in estimation are not adequately expressed in footnotes and therefore the ability to interrogate the results is reduced. Even the "margin of error" (which you referred to in an earlier post on this topic above) is not stated, nor how it is determined. Orderinchaos 00:17, 18 April 2009 (UTC)
 * Just found the publication to which Bowe is referring (the link to Parl Lib publications was broken for months, they seem to have fixed it recently) - Green's introduction contains the following: "The estimates should be treated with some caution, as they can only be a guide to the political impact of new boundaries. The estimates cannot take into account differences in candidate and campaign effort that would have occurred if the election had been fought on the new boundaries. The inaccuracies introduced by splitting polling place catchment areas also mean that the estimated new margins should be treated as approximate rather than exact margins." He notes on page 3: "Three new metropolitan seats are calculated to have margins under 1.0%, well within the margin of error used in calculating estimated margins. Kingsley is calculated to have a Labor majority of 11 votes." Orderinchaos 00:58, 18 April 2009 (UTC)
 * Oh, and I entirely forgot to answer the point made by WB re Black/Phillips, I heard a radio interview, read a piece in a community newspaper and saw them at Curtin Uni during a forum, and in *all* of those occasions, they'd noted Kingsley was a Liberal seat. They never once talked about Labor losing it. The comment re "actual booth votes" was related to something Phillips said at the forum (incidentally, he predicted a Liberal win, while Black was probably closer than anyone as he forecast a hung parliament with the Nationals and an Independent deciding the result). I apologise if I was unclear at the time with regard to what I meant they had said. Orderinchaos 00:28, 18 April 2009 (UTC)
 * I didn't round the Kingsley booth votes to 15%. 15% is the enrolment transfer. All the declaration votes transfers in my publications are calculated according to the enrolments transferred, as supplied by the Electoral Commission, NOT the votes in the booths transferred. I've been doing these redistribution books for Parliamentary Libraries for two decades and always use the same methodology. Yes Kingsley was close, but the numbers plonked it just on the Labor side of the pendulum, using exactly the same method as for every other seat. It had a sitting Labor MP, which meant calling it a marginal Labor seat was always the easiest solution if that was what the calculations produced. All these calculations produce approximate margins, in the end, the only alternative in close seats is to simply say it's not held by anyone, but where does that help anyone's analysis?  AntonyGreen (talk) 02:26, 18 April 2009 (UTC)
 * Ah cool, makes sense re enrolment transfer. Although the point still stands that we may know where the corpus of declaration voters are located through EC data, but for obvious reasons we do not know how each voted (e.g. enough to exclude Warwick and Edgewater appropriately) as the envelope and the vote are separated when it enters the count.
 * I hope you have not interpreted my comments above as being an attack or criticism of your work - I think it's one of the best resources we have and for the most part we'd be lost for quality analysis without it. One thing it is particularly useful for, and which I use it for consistently, is calculating swings - i.e. we can say a particular seat swung a certain percentage to whichever party and be more correct than simply comparing the performance of a previous election in what was in some ways a very different electorate.
 * However, as your last sentence preluded, part of this is about different audiences - producing an analytical information work for the general public and to assist Parliament and the WA Electoral Commission, and writing an encyclopaedic article about it for Wikipedia are allied but not 100% compatible aims. One example - with regards to "not being held by anyone", that *is* possible for our purposes, but isn't for yours. This is because it makes more sense for our purposes to look at the overall balance of Parliament then and now rather than notional seats, with other analysis of the results being discursive in nature in the article text. Consistency over time and sparse historical data means we can't compare apples and oranges over, say, an 80 year period.
 * I guess though my academic training makes me permanently wary of accepting anything as gospel especially when it's subject to random social and demographic factors - there's never a final, or a single correct, answer for any of this, no matter how hard we try or how solid our source info is. And one, nine or eleven votes out of 20,000+ are very, very small numbers from a statistical angle, and if the margin of error is 1.0%, we're looking at 200+ voters. What is provided is always just an indication based on theories which are refined over time.
 * Thanks for your considered response. Orderinchaos 03:26, 18 April 2009 (UTC)
 * The point that the declaration is cast and where it comes from is recorded is true, but the point also stands that this data is not released by the Electoral Commission so it can't be made use of. A scrutineer in a specific seat may have more information on this from access to written documents and bundle wrappers, but the data is not generally released. And the point that David Black and Harry Phillips were making about Kingsley is a Liberal electorate is also perfectly true. It is an electorate that should be Liberal held. But in 2005 it was won by the Labor Party, and taking those votes, it isn't a Liberal electorate. When I do these redistributions, the point is all the booth transfers become the input data for my next election broadcast. You always use the result of the last election as your starting point.
 * As for being for different audiences, that is entirely appropriate. I don't think the way this page classifies the seat change from 2005 is very meaningful. Comparing this 59 seat Parliament with the 57 seat Parliament in 2005 doesn't illustrate how many seats Labor lost. I adopted a different approach in my results book, which makes more sense in explaining a specific election. But if you were looking at data over time, you don't always look at before and after boundariues, but rather the snapshots of election results. Again depends on what you are trying to illustrate.
 * But that so much time has been wasted on Kingsley baffles me. It was so damn close however you calculated it, just classify it by the sitting member. Seats where a sitting member represents a seat notionally held by another party are the bane of my life and an endless source of confusion as people argue whether the seat is gained or retained. Just deal with it in footnotes is my view. I sometimes make an alternate choice for television because we don't have the luxury of footnotes, But I also don't have a category of seats calle 'Unknown', so you just decide. Kingsley came out on the Labor side by 11 votes and it had a Labor MP, easy decision for me. 11 votes on the Liberal side and a sitting Labor MP, I might have made a decision based on trying to avoid confusion rather than just sticking to the numbers. 07:42, 18 April 2009 (UTC)

Reverts in mainspace
The west australian newspaper had a story today (19th September) that says all counting has been completed and the count declared - please do not revert Orderinchaos.Rebecca edits - thank you SatuSuro 14:51, 19 September 2008 (UTC)
 * Then cite that story, if it includes the numbers. The WAEC site shows only 86% counted right now (http://www.waec.wa.gov.au/elections/state_elections/election_results/2008_State_General_Election/). 64.32.30.2 (talk) 14:59, 19 September 2008 (UTC)
 * Which is the turnout now counting is complete. 14% of those on the roll didn't vote. Jmount (talk)
 * That's correct. The count hasn't even been declared yet btw (although the seats individually have, so we know who won what but in some cases not by how much) - there's still about 11 seats to go. There will, from my experience, be continuing minor adjustments but not major ones - the edits I've made to the tables probably give some idea (generally messing around with informal/major party votes in the order of 50-100 at a time, with the odd surprise thrown in). Orderinchaos 15:26, 19 September 2008 (UTC)
 * If counting is complete, why does it explicitly say "Count In Progress" for several seats? And if the seats have individually been declared, then you still need to cite a source for that. They sure haven't been declared on the WAEC site. 64.32.30.2 (talk) 15:41, 19 September 2008 (UTC)
 * According to here, 8 seats have been declared, 40 have been finished counting and are awaiting declaration, and 11 are still being counted. - Mark 15:59, 19 September 2008 (UTC)
 * The "% Counted" on the WAEC website might be the number of ballots counted out of the number of enrolled electors. Or perhaps they declare the result once the number of ballots left to be counted is less than the current winning margin of the leader, and hence the trailing candidate cannot possibly win. - Mark 16:02, 19 September 2008 (UTC)
 * That's correct - one can in fact verify that by dividing the total number of votes counted by the total electoral population. In Western Australia, a turnout of over 90% is very rare indeed, and it was generally anticipated that the turnout for this one would be lower due to the blandness of the campaign and antipathy towards both the government and opposition. The West Australian Electoral Commission liaison rep told ABC radio this morning that all primary votes had been counted and the declarations were simply the formal close of the procedure. Incidentally, a by-election in Victoria Park in 2006 had the lowest turnout - just 64% - of any election in WA in the past 50 years. Orderinchaos 16:16, 19 September 2008 (UTC)
 * Also note from The Australian newspaper which gives the seat totals. Orderinchaos 16:34, 19 September 2008 (UTC)

OK mY misread - apologies I misread - declared but not count complete - and the subsequent chatter above shows the west report was probably re what orderinchaos points out about primary votes counted was the issue - not the rest of what the conversation has gone through SatuSuro 01:29, 20 September 2008 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 1 one external link on Western Australian state election, 2008. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive http://web.archive.org/web/20080722024823/http://www.waec.wa.gov.au/download/September+enrolment+statistics+2.pdf to http://www.waec.wa.gov.au/download/September%20enrolment%20statistics%202.pdf

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.—cyberbot II  Talk to my owner :Online 18:25, 31 March 2016 (UTC)