Talk:2018 Wentworth by-election/Archive 1

Archive 1

Picture of Sharma in Infobox

I think it is problematic that a picture of Sharma and no one else is front and centre in the infobox. Until we have a picture of the ALP candidate we shouldn't have just Sharma. And since we don't know if it will come down to an LPA/ALP contest, we probably shouldn't imply that it will be.--Jack Upland (talk) 09:11, 14 September 2018 (UTC)

I strongly agree. It's a very unbalanced article right now. All about the Libs. With Kerryn Phelps confirming she will stand, and the Greens having been not been too far behind Labor last time round, implying this is a two party race is just plain wrong. HiLo48 (talk) 03:55, 16 September 2018 (UTC)

Agree also. Just because there is a photo of Sharma available does not mean it must be used. Its sole use hints at a breach of WP:NPOV. I'm removing it. WWGB (talk) 05:29, 16 September 2018 (UTC)

Implying that it is a Liberal-Labor contest is perfectly in line with how we've treated elections and by-elections in the past. Given that these two were the top two candidates in the last contest, we should wait for reliable sources to report otherwise. Onetwothreeip (talk) 05:48, 16 September 2018 (UTC)

The situation now is definitely NOT in line with elections and by-elections in the past. I learnt long ago that "We've always done it that way" is a terrible reason to do anything. HiLo48 (talk) 08:20, 16 September 2018 (UTC)

It's more about being consistent with our other articles. There has to be a good reason to do it some other way. Onetwothreeip (talk) 08:29, 16 September 2018 (UTC)

"It's more about being consistent with our other articles." That's really not much different from "We've always done it that way" WP:OSE is not an argument for always doing things the same way. HiLo48 (talk) 07:45, 17 September 2018 (UTC)

"We've always done it this way" is not "other stuff exists", which is what you've linked. My argument is neither of those things. I could very well be content with change, as long as it was applied consistently. Other than that, the change actually has to be argued as good. This is basic manual of style. I told you that my concern was in consistency after you made the accusation. Onetwothreeip (talk) 00:30, 18 September 2018 (UTC)

"...the change actually has to be argued as good". The current content is being argued as bad, with some excellent points being made. Almost anything would be better than what's there now. You really need to prove that's not the case. We don't do the same thing as elsewhere when it simply doesn't work here. HiLo48 (talk) 03:54, 18 September 2018 (UTC)

It's misleading. It looks like Sharma got 67.75 in the last election and Murray got 32.25. It looks like there's no other viable candidate. None of this is true. There's no point in following a formula if the formula is misleading.--Jack Upland (talk) 09:14, 16 September 2018 (UTC)

Polling

The ReachTEL poll commissioned by The Australia Institute is essentially two polls, one with the main parties and other, and one with the main parties and two noteworthy independents and other. How do we deal with this? Two entries? Onetwothreeip (talk) 08:01, 29 August 2018 (UTC)

The first question of polling intention is always the one that is used (exclusively), follow-up questions always deviate from basic polling intention methodology, most obviously and notably in this case, with a leading non-unbiased question asked of respondents. Additionally, these follow-up questions include hypothetical independent candidates (of which Alex Greenwich isn't going to run) and a specific hypothetical Liberal candidate, and furthermore provide no 2PP vote unlike the first question of polling intention. And Antony Green rubbishes the follow-up poll (and mentions that the two results have been confused in the media). Timeshift (talk) 08:52, 29 August 2018 (UTC)

Antony Green does not rubbish the second question.

"A second question asked about voting at a by-election following Malcolm Turnbull's resignation. The reported first preferences were Liberal 35%, Labor 20%, Greens 9%, Independent Alex Greenwich (who has ruled out running) on 11% and former AMA President Dr Kerryn Phelps (who is not sure about running) on 12% and Other/Independent on 13%. The poll did not attempt to produce a result after preferences. In the absence of an actual campaign and without knowing the names of candidates, this poll is highly speculative and probably unreliable."

Another two-party-preferred result is not necessary, it's already asked in the first question. 'Follow-up questions" are normally not about voting intention, normally the second question is a policy matter. We frequently include second questions, the two-party-preferred result is a result of a second question. If we're going to rule out polls because of their unreliability to relate with the result of the election, then we better off remove every single electorate poll on every single by-election page. Onetwothreeip (talk) 10:42, 29 August 2018 (UTC)

"The poll did not attempt to produce a result after preferences ... this poll is highly speculative and probably unreliable". The second question is not "normally" a policy matter, it is sometimes used to ask about alternative hypothetical voting intention under an alternative hypothetical leader(s). And we do not "frequently" include polling outside of voting intention, we never do AFAIK (apart from the regular standard leader ratings for non-by-election/federal election articles). Can you show me another Australian by-election article where we have included polling beyond the initial primary/2PP voting intention? Timeshift (talk) 12:07, 29 August 2018 (UTC)

The two-party-preferred voting intention is a second question to the primary voting intention. We comprehensively include the two-party-preferred result in polling in virtually every article where polls are included. The two-party result for the polling question which included Greenwich and Phelps would have been the exact same two-party question as the question without the independents, and would have the same answer.

To your claim about the second question not normally being policy, my experience with polling from The Australia Institute has been that they normally ask questions about policy along with their voting intention questions, and frequently others like Essential Research too.

Overall there's no reason why we can't incorporate multiple results from the same poll, as is often done in American polls, and in fact essentially what we do with two-party-preferred results anyway. Onetwothreeip (talk) 00:21, 30 August 2018 (UTC)

Incorrect - see what "Question 2" is at the ref. I'm afraid we must agree to disagree. Timeshift (talk) 00:48, 30 August 2018 (UTC)

I'm not sure what you're disagreeing on. It's undeniable that the two-party-preferred voting intention is a result of a question that comes after the primary voting intention question, they ask at least two questions for that, three including distributing undecided voters. They also regularly ask about policy in supplementary questions, as also can be seen in what The Australia Institute has released (Questions 3, 4, 5, 6, 7). Onetwothreeip (talk) 01:32, 30 August 2018 (UTC)

• Do we really need to document every poll? If you look at other by-election articles we have a list of polls and then the final result. It seems rather pointless.--Jack Upland (talk) 05:46, 16 September 2018 (UTC)

Yes. Onetwothreeip (talk) 05:53, 16 September 2018 (UTC)

Yes, it's pointless?--Jack Upland (talk) 05:57, 16 September 2018 (UTC)

No. Yes to your question. We should document every reputable poll, that's very much standard practice for election articles. Onetwothreeip (talk) 06:04, 16 September 2018 (UTC)

So, yes, it's pointless, but we've always done it.--Jack Upland (talk) 06:16, 16 September 2018 (UTC)

It's pointful. Opinion polling is a fundamental element of covering elections. Especially in a by-election like this! Onetwothreeip (talk) 06:36, 16 September 2018 (UTC)

Opinion polling is designed to predict the final result. Once we know the final result it's pointless.--Jack Upland (talk) 09:11, 16 September 2018 (UTC)

That is not the design of opinion polling. It is used to track sentiment over time. Onetwothreeip (talk) 23:50, 21 September 2018 (UTC)

Election results back to 2004???

Craziness.--Jack Upland (talk) 11:58, 26 September 2018 (UTC)

The reason I decided to add them is so there was a non-prose-text representation of the Wentworth result during the previous MP's period, particularly as it indicates that the previous MP had built up a very substantial personal vote, after being a mere marginal seat in 2004 and 2007. I added it to see what contributors think about it, so please feel free to actually discuss it. Timeshift (talk) 15:14, 26 September 2018 (UTC)

We can have tables for the last few elections of the two-party-preferred result, but including all these votes for all those elections in with the polling data is just WP:INDISCRIMINATE information. If this is because there wasn't background detail on Malcolm Turnbull in the seat of Wentworth, that is very easy to rectify. Onetwothreeip (talk) 21:40, 26 September 2018 (UTC)

Background?

When does the "Background" give way to the "Campaign"?--Jack Upland (talk) 04:51, 29 September 2018 (UTC)

At last, a section on the campaign!--Jack Upland (talk) 07:38, 13 October 2018 (UTC)

Regarding infobox

As polling has revealed that this will be a by-election with a close three-way race rather than the ordinary two-party races, should Kerryn Phelps be added as a third candidate to the infobox? Catiline52 (talk) 02:00, 5 October 2018 (UTC)

Absolutely, and I'd do it myself if I wasn't so busy this weekend. Hopefully someone else does, and you're free to do it. Onetwothreeip (talk) 02:31, 5 October 2018 (UTC)

We've discussed this before. I agree that the current format is problematic, but adding Phelps in based on a couple of opinion polls is also problematic. Polling hasn't "revealed" anything.--Jack Upland (talk) 08:30, 5 October 2018 (UTC)

Polling indicates Kerryn Phelps has a significant likelihood of winning, for example. Onetwothreeip (talk) 21:01, 7 October 2018 (UTC)

Now that there are 6 polls and news articles suggesting it will likely be a Liberal/Phelps race rather than a Liberal/Labor race, should Phelps be added to the infobox? Catiline52 (talk) 06:54, 17 October 2018 (UTC)

Well, she's there now, but it's still a very clumsy presentation, implying that both the Lib and Labor candidates stood last time round. There are times when simplistic Infoboxes simply don't work. HiLo48 (talk) 06:43, 19 October 2018 (UTC)

CORRECTION: While I was writing the above, someone who agrees with me removed the Infobox completely, and that is the right thing to do. HiLo48 (talk) 06:45, 19 October 2018 (UTC)

Antony Green's called it for Phelps

Antony Green has called the election for Phelps, so when can the 'results' section get started? --122.108.141.214 (talk) 08:37, 20 October 2018 (UTC)

Percentage changes

There cannot be a percentage change for a candidate who has not previously stood. Assuming zero is incorrect as some candidates who previously stood may have got zero or close to zero. For those candidates there can be a change. There cannot be a change if there is nothing to be a comparison too. As such it should be denoted as N/A. 91.110.126.37 (talk) 20:20, 21 October 2018 (UTC)

Of course there can. This exactly is how the Australian Electoral Commission shows the results, and the ABC, and pretty much every use of the election results templates on Wikipedia in dozens of countries. And it's not referring to candidates, but to political party. If a candidate stands as an independent in consecutive elections, then the change in their vote percentage is reflected from the previous election. Your assertion and assumptions make no sense and are out of step with established consensus and practice on tens of thousands of articles, as well as many government electoral authorities. --Canley (talk) 21:11, 21 October 2018 (UTC)

It's not a percent change in increase of their last vote. It's just how many more percent they got this time. So if they got 2% last time and now they have 10%, that's +8%, not 500%. Onetwothreeip (talk) 21:13, 21 October 2018 (UTC)

I don't think this is a percentage points issue. I think the editor is expecting the change column to be a zero-sum thing, that parties/candidates not running are also listed as minus values so that the changes add up to 0. This is ridiculous and will make thousands of election tables difficult to manage and incomprehensible. --Canley (talk) 21:21, 21 October 2018 (UTC)

Actually, adding N/A won't do that, so I guess their undo comment just meant because we don't do zero-sum like the AEC does, then we shouldn't also do vote percent change from zero like the AEC also does. --Canley (talk) 21:38, 21 October 2018 (UTC)

Looking again I think you’re right, maybe they’re saying you can’t have a % change from 0 (which is true but this is showing a percentage point change). Hmmm, I got told to take it to the talk page but no further comment from the IP. —Canley (talk) 22:01, 21 October 2018 (UTC)

Election of Phelps becoming uncertain after postal votes

With 73% of the votes counted, the election is becoming quickly 50/50. The article shouldn't assume Phelps should win, as although she won by a significant margin in the polling booths, pre-polls and postals seem to be significantly rising Sharma's vote. We shouldn't assume until this is dealt with. Catiline52 (talk) 23:05, 20 October 2018 (UTC)

Reliable sources are saying Phelps has won, particularly the ABC which is our highest priority source, and of course the Liberal candidate has conceded also. Reputable analysts Kevin Bonham and William Bowe believe that while the vote will be closer than last night, these won't be enough to make the difference. There are also two booths which are likely underreporting preferences for Phelps, so it's not all going to be in Sharma's favour. Onetwothreeip (talk) 23:12, 20 October 2018 (UTC)

"We now have the first batch of postal votes – about 5,463 votes. Dave Sharma is winning 64.4% of them after preferences. This suggests he is roughly polling enough in the postal vote to have a chance of winning – the big question is how many postal votes are left to be counted, and whether adjustments in the election-day vote might give Phelps some breathing room." - Ben Raue from the Tally Room (One of the sites used in the 'external links' section). Reliable sources are saying that the announcement of the result may have been premature last night. It's probably safer to not report a victor with such close results than accidentally make a "Dewey beats Truman" mistake. Catiline52 (talk) 23:25, 20 October 2018 (UTC)

The Libs would need 70%+ of postal votes to win, and with half of them included this morning, they are only getting 64% of postals, so they can't (won't) win from here. Timeshift (talk) 23:33, 20 October 2018 (UTC)

"In the last of my updates in the section below, I calculated that Sharma would need 70% of postals to close the gap, but it seems this was an overestimate." and "To cut the following long story short: this isn’t over." - William Bowe from the Poll Bludger. Several analysts are starting to pull away from the certainty of a Phelps victory. An electoral victor shouldn't be declared unless it is certain. Catiline52 (talk) 23:56, 20 October 2018 (UTC)

Why was the information about the postal swing to Sharma removed? It's being discussed in multiple reliable sources, and even if the postals don't carry Sharma home, it's worth mentioning as part of the results. --122.108.141.214 (talk) 22:11, 21 October 2018 (UTC)

It's the final outcome of the election that matters, not so much the chronological machinations of the count (unless a recount is triggered which is worth mentioning). Also, it wasn't just a postal swing—there were a few transcription errors which made the two-candidate margin look closer than it was, but this was detected and fixed up in the standard check count. All in all, these swings in votes received at a particular stage of the count (and counting or transcription errors) happen all the time and this isn't a particularly notable instance of that. --Canley (talk) 22:24, 21 October 2018 (UTC)

I agree that it's the final outcome that matters. The count doesn't matter unless there's a major incident.--Jack Upland (talk) 07:45, 22 October 2018 (UTC)

Incredible detail in this entry

I appreciate the detail in this coverage of what turned out to be an historic by-election, the largest swing against a Federal gov't in Australian History. TheBustopher (talk) 11:59, 20 October 2018 (UTC)

Tim Murray and several journalists have said that tonight, but I don't think it's true. In the Wills by-election, 1992, the swing was 23.6 against the ALP, and Lyne by-election, 2008 was 32.5 against the Nationals. --Canley (talk) 14:04, 20 October 2018 (UTC)

The Wills by-election was declared void by the High Court, so I don't think it really counts. And the swing in the Lyne by-election corresponds to the Nationals' loss, who at the time were not a party in the federal government (in fact, Labor did not contest it). Impru20^talk 14:10, 20 October 2018 (UTC)

Good point re: Wills but debatable I think. Oops, ignore Lyne. Also Wentworth by-election, 1956 (but there were a lot of Independent Liberal candidates), and Wakefield by-election, 1938 (although the swing seems to have settled later in the night and did exceed this one earlier). --Canley (talk) 22:44, 20 October 2018 (UTC)

Could be, though Wentworth 1956 was not an actual loss. As for Wakefield, the swing was fairly similar (20.1 as opposed to 19.7). I also guess that media claims for the current election to constitute the largest swing against a sitting government was done on the basis of early returns, when the swing was even larger (about 23.0 or so). Nonetheless, if not the largest it is definitely one of the largest. Impru20^talk 22:50, 20 October 2018 (UTC)

I think Wills still counts, and is being mentioned by commentators.--Jack Upland (talk) 07:47, 22 October 2018 (UTC)

Polling issues with infobox

Some polls are starting to do TPP results between Phelps and the Liberals rather than Liberal/Labor. How should this be dealt with in the polling infobox? I added a new column for Phelps/Liberals alongside Labor/Liberals, I'm unsure if there's currently a solution to a three-person race.Catiline52 (talk) 05:25, 13 October 2018 (UTC)

I know we have traditions and all that, but I would rather see no attempt at all to put simple and potentially misleading polling figures in the Infobox, and leave it to the text to contain words very similar to above from Catiline52. HiLo48 (talk) 06:14, 13 October 2018 (UTC)

What do you think is potentially misleading? Onetwothreeip (talk) 06:38, 13 October 2018 (UTC)

All my longish life the whole concept of "Two Party Preferred" has meant a comparison between the potential success of the Liberal/National Party and the Labor Party, at both a local level and a national level. That does not apply here. And an independent candidate is obviously not logically referenced by the first P in "Two Party Preferred". Not what TPP was invented for. The massive number of candidates also makes predicting these things somewhat doubtful. HiLo48 (talk) 06:45, 13 October 2018 (UTC)

Fair enough, we'll say two-candidate preferred. Onetwothreeip (talk) 06:51, 13 October 2018 (UTC)

But that's not accurate either. Yes, people who vote for an independent probably "prefer" that particular candidate, but that's not the case for a Liberal candidate. In fact, that's a major issue here. People who have "always voted Liberal" (a quote from ABC News today), no matter who the candidate actually was, are considering not doing so this time round. Obviously for some Lib voters it still won't be about the candidate, just the party. HiLo48 (talk) 06:57, 13 October 2018 (UTC)

They're still voting for candidates. The polling organisation asked them which candidate they preferred out of two, for two sets of two final candidates. Onetwothreeip (talk) 07:00, 13 October 2018 (UTC)

You're ignoring reality. HiLo48 (talk) 07:33, 13 October 2018 (UTC)

Would you like to explain how? Onetwothreeip (talk) 21:28, 13 October 2018 (UTC)

Because of the easily sourced fact that many voters always vote for the same party, no matter who the candidate is. HiLo48 (talk) 06:40, 19 October 2018 (UTC)

The votes are still for candidates, regardless of the reason. Onetwothreeip (talk) 05:47, 23 October 2018 (UTC)

2PP

Timeshift9 would you like to explain what you mean by saying that the 2PP swing against the Liberals is 67.75, and the 2PP swing against Labor is 32.25? Or please provide an alternative description for this. Onetwothreeip (talk) 08:49, 24 October 2018 (UTC)

That's not what it means. See Mayo by-election. For 2PP field, having Liberal (-67.75) and Labor (-32.25) means that is what they were at the last election and is awaiting the commencement of counting at this election. If you don't understand it, that's not my problem. Timeshift (talk) 09:35, 24 October 2018 (UTC)

But that's not what it says in the table and those aren't what the swings are going to be. If this happened in the Mayo by-election it would be wrong there too. The row is for swing, not previous 2PP. Onetwothreeip (talk) 10:42, 24 October 2018 (UTC)

That's why there are brackets and it adds up to 100%. The AEC/ABC do the same thing after adding the section but before the count commencement... it's a pity there's nowhere I can presently show it to you. But it's correct, and it's how it has been done before. Why do you think it wasn't reverted on the Mayo by-election article...? Timeshift (talk) 11:34, 24 October 2018 (UTC)

This seems like a weird way of doing things, and I've never seen it done before this year. Let's just not list a 2PP until we actually have one. (This reminds me that we agreed a while ago that 2PPs should go below 2CPs and the "Independent gain", but then I never got around to implementing it. Will have to stick it on the to-do list.) Frickeg (talk) 13:03, 24 October 2018 (UTC)

If I was aware of it on the Mayo by-election article I most certainly would have removed that. Given there were five by-elections that day, it's no surprise that there was less diligence. I think what you are referring to from the ABC and the AEC is that when there are 0 votes for a candidate, as it shows for all candidates before any votes are entered into the computer, their share of the vote defaults to 0% and the swings are all shown as if they have 0% of the vote. Pointing to that article is simply saying WP:OTHERSTUFFEXISTS. Since I don't want to be in an edit war, I appeal to another user to make the relevant change if necessary. Onetwothreeip (talk) 23:21, 24 October 2018 (UTC)