Talk:Opinion polling for the 2013 Australian federal election

Galaxy etc

I've restructured the article to start to add in Galaxy polling. Hopefully the format works but please feel free to change things. Rsloch (talk) 15:47, 4 February 2013 (UTC)

Why not simply add the Galaxy and Essential polls to the other tables? Readers compare polls, and that would be far easier without having to scroll through huge chunks of the article. -Rrius (talk) 17:28, 4 February 2013 (UTC)

I was following the format used in Opinion polling for the Australian federal election, 2010 for consistencies sake. Perhaps creating an all in one table for all election polls with the pre election polls below would be a solution. Rsloch (talk) 18:03, 4 February 2013 (UTC)

Why does the table only include Newspoll in the first place? It's a random choice, it's biased, and it's pretty misleading as to the actual situation of polling when you're not including 3/4 of the data (Morgan, Galaxy, Essential Research). The Drover's Wife (talk) 12:02, 5 February 2013 (UTC)

Below it a suggested template.

House of Representatives

Date	Pollster	Political parties				Two-party-preferred
		ALP	Coa	Grn	Oth	ALP	Coa
5 Feb 2013	Guesspoll	40%	40%	5%	5%	50%	50%

Rsloch (talk) 14:32, 5 February 2013 (UTC)

Seconded. This table is much closer to parallel articles format. --4idaho (talk) 22:53, 11 March 2013 (UTC)

If there is no objection and no one has already started one, I intend to start incorporating the data from Galaxy, Nielsen, and Essential in a table along the format proposed in about 20 hours (noon AEST). -Rrius (talk) 06:46, 18 March 2013 (UTC)

In my user space I've incorporated Newspoll and Nielsen for BPM/approval, and those two plus as much Galaxy as I could find (nothing before August 2011) for party preference. If someone wants to make a start on Essential or Morgan, please use {{in use}} while you do, so I don't duplicate your effort. -Rrius (talk) 08:02, 19 March 2013 (UTC)

Graphs

The graphs in the article are out-dated. I've tagged the article. I don't see an active Wikipedia account for the user who made the graphs. Somedifferentstuff (talk) 08:16, 19 March 2013 (UTC)

I have updated the data at User:GoForMoe/polling. Anyone who understands Python, Matplotlib, and numpy can easily create the SVGs. I have no idea how to work with those things. -Rrius (talk) 10:30, 19 March 2013 (UTC)

I've placed an edit warring warning on your talk page. By removing the tag as you did, users/readers are no longer alerted that there is an issue. I advise you to reinsert the tag. Somedifferentstuff (talk) 10:42, 19 March 2013 (UTC)

The article is not out of date. The article contains the most recent polling. That the graphs are out of date is a separate issue and is being dealt with. So stop being such a drama queen. -Rrius (talk) 10:51, 19 March 2013 (UTC)

And no, I most certainly won't. In addition to the tag's being overkill, after publicly noting your actions at my talk page, you'd have to be mad to expect I would do anything just because you want me to. The issue is being discussed and dealt with here and elsewhere, so there is no reason why editors need to be alerted (not that there are a ton who would be alerted by the tag anyway). -Rrius (talk) 10:55, 19 March 2013 (UTC)

By removing the tag as you did, users/readers are no longer alerted that there is an issue, which creates a neutrality issue due to the inaccuracy of the information. I've added specific tags to the graphs instead of the general one I was using. Somedifferentstuff (talk) 14:05, 19 March 2013 (UTC)

The tables do reflect the information. The graphs not reflecting it neither renders the article out of date nor POV. Where on earth you get the idea that out of date templates make affect neutrality is beyond me. And once again, the template was completely useless for accomplishing what you wanted. Especially since it there was no discussion here. The right move would have been to simply leave a note here or at the main article's talk page. I have done the latter, and an editor has committed to updating the template around lunchtime today. And before you try to argue that your addition of the template led to the outcome, the same exact thing would have happened had you simply left a note on the talk page here. -Rrius (talk) 01:44, 20 March 2013 (UTC)

Leaving a note on the talk page alerts absolutely no one who is viewing the main page that there is an issue. For some reason you are unable to grasp this concept. I recommend you seek assistance from a more experienced editor. In the meantime the tags will remain in place until the problem is resolved. I looked at the recent changes made to the graphs by User:Canley. The first graph here[1] contains the wrong thumbnail so the graph we see in the article is the outdated version with the February 1-3 date. The second graph[2] appears not to have had any changes made to it and also has the wrong date. There are problems with the next 4 graphs as well. Somedifferentstuff (talk) 09:31, 20 March 2013 (UTC)

The Primary Vote graph has been updated and is suffering from the same thumbnail problem as the others. The thumbnail issue is a Wikimedia problem of several months' standing that we mere editors can't do anything about. It will be fixed when it is fixed.

Now, it is the people who actually watch the page, and therefore will see your edits in their watchlists, who you need alerted when an issue like this comes up. Not random people who come along to view the article. What's more, simply slapping up a template, as you did multiple times, gave no indication whatsoever what your problem was. That is doubly the case where the template doesn't really match the problem. It is you who need the help of a more experienced editor. For myself, I'll rely on my eight years and nearly 42,000 edits. Leaving a note on the talk page was essential, yet you repeatedly added the template without bothering to do the thing you actually needed to do. That leaving a note on the talk page is essential is so basic a piece of knowledge that it is astounding you still can't grasp it. It's alluded to in the template itself for chrissake! It should have occurred to you that something was up when absolutely no one had responded to your problem after two weeks. Then again, your excuse about the datedness is not what you initially claimed. Your edit summary said, "Added POV tag. This article relies on ONE polling firm." That has fuck all to do with how up to date the graphs are. And it was also wrong. Aside from the fact that there is no obvious neutrality problem, the article did and does provide some Galaxy polls, expansion has been under discussion for some time, and strides are being made to include much more polling. And you would know all this if you actually read the article (or even just the table of contents) and looked at the talk page. Instead you did the laziest thing you could have done, adding a template without starting a discussion, yet expected your template to sit there as a monument to your snit for all time. -Rrius (talk) 10:45, 20 March 2013 (UTC)

It's amazing that you have that many edits and still don't understand how tags function. The tag I applied later was precisely the tag required (dated info) showing that the graphs were not current. This alerts editors that there is a problem and more info can be found on the talk page. This alerts readers by letting them know that the graphs are not accurate. Even the user who recently worked on the graphs stated on my talk page, "...just remove the tags when you are satisfied and can see the updated versions..." - Get a clue already. Somedifferentstuff (talk) 11:28, 20 March 2013 (UTC)

Nonsense. The article message templates you used do not say anything about the graphs. One says the article violates Wikipedia's Neutral Point of View policy, and the other says the article is inaccurate because of outdated information. The information in the article was up to date and accurate. The graphs that incorporate that information had not been updated for about a month. Those are two different things. There is no overriding reason why the graphs ever need to be updated so long as they clearly mark what data they represent, and there is certainly no requirement that the graphs be updated with the addition of each new poll. Adding a tag that claimed the article violated NPOV was so wildly off base that it is shocking you are still defending that move. Moreover, you did not start a talk page discussion until well after the tags had been added, so there was nowhere for people to read about your beef. Alerting editors to a problem and then not explaining it is stupid unless the problem is self-evident, which it clearly wasn't. And why do you keep pretending this is all about the graphs? Your initial move was based on, according to your edit summary, the asserted fact that there was only one poll reflected in the article (which was untrue even then). And despite the fact that you seemed to have a problem with the number of polling firms, you didn't bother to participate in the discussion above, which was started before you added your first tag. You have a pretty bizarre understanding of the power of article message templates. You seem to think that by merely adding one everyone can see inside your head and understand what you have a problem with and why. There was never a POV problem, and the accuracy of the article was never in doubt. After all of the effort you have put into this, you have still failed to ever explain how using a single firm's data expresses a point of view (even if it were true that only one firm was represented). And you have failed to explain how, despite the actual data (the tables) being up to date, the whole article was made inaccurate by the graph images not having reflected the two or three most recent polls (depending on the particular graph). So I would put your own advice back to you: Get a clue. -Rrius (talk) 00:36, 21 March 2013 (UTC)

The only tags on the article are directly attached to the graphs and state "dated info"[3] so not only are they appropriate but are still necessary as explained above. Your lack of understanding is freaky and if I come across someone who can assist you I'll send him your way. Somedifferentstuff (talk) 07:45, 21 March 2013 (UTC)

No one was ever disputing those tags. This discussion has always been about the article warning templates you added at the top of the article here, here, here, and here. I've never commented on the image-specific tags you added only later. It was only after your third contribution here you even added them. You clearly have no idea what the hell is going on around you, let alone what you are talking about, but I'm done trying to explain it to you. -Rrius (talk) 07:57, 21 March 2013 (UTC)

Incorporation of other polling firms

In line with the consensus developed above, I have incorporated the Nielsen BPM and satisfaction numbers in with the Newspoll numbers. I have also added Nielsen, Essential, Morgan, and Galaxy voter intention and 2PP in with Newspoll. The new tables are here: User:Rrius/Sandbox 1. I may be missing a good many Galaxy polls, but I think I have found all of the individual polls for the others. I think we need to add a methodology section explaining the methodology for each, but I have no intention of doing so in the short to medium term. Before replacing the tables, I'll leave things as they stand for a few days to give everyone a chance to leave their comments here.

There is also the matter of graphs. The question is whether we do graphs for each firm, graphs that incorporate all firms and distinguish with different lines or markers, or graphs that include all data without distinguishing between firms. -Rrius (talk) 08:23, 6 April 2013 (UTC)

Further to the graphing issue, I have created some using MS Excel. They include polling from all of the polling firms. I can also make BPM and satisfaction graphs if these are deemed acceptable. -Rrius (talk) 11:56, 6 April 2013 (UTC)

The colour on the first graph is wrong, ALP should be red, and Coalition blue. — Preceding unsigned comment added by 174.19.224.63 (talk) 07:23, 20 April 2013 (UTC)

Sample sizes + mode please?

Obviously some of you have pretty detailed records of the polls. Can we please add sample sizes as a column? This is important to understand the relative impact statistically.

Additionally, it would be helpful if the mode in which the survey was conducted was listed. Some of Morgan's polls are telephone, some are face to face. This is important because of house effects. Thanks Velocidex (talk) 22:54, 16 April 2013 (UTC)

Each pollster needs to be separate in the graphs

Each pollster is a completely different series. The graph should not connect the results of all of the pollsters. That is not at all accurate and is misleading. Apples and oranges. The graphs in the 2007 and 2010 election articles are correct. The graphs for 2013 should be amended accordingly. 58.106.64.38 (talk) 00:46, 11 June 2013 (UTC)

Separate graphs would result in too many. While the polls do not necessarily use the same methods, they produce broadly similar results, and presenting them this way is fairly standard where graphs are employed on Wikipedia. Providing separate lines for each firm in one graph would be too busy, and providing separate graphs would result in excessive images. I wouldn't object in principle to having each firm's data points marked with distinctive symbols, but I don't know how to do that, and I'm not sure the symbols would be visible. Incidentally, trying to compare to 2007 and 2010 is pretty silly, since Newspoll is the only firm whose data is used, at the article and in the graphs. -Rrius (talk) 06:15, 11 June 2013 (UTC)

The graphs used in the 2007/2010 elections I linked to in my previous comment. Look at them, it's not just Newspoll. 58.106.64.38 (talk) 22:05, 11 June 2013 (UTC)

I'd never seen that before, but it doesn't really matter. I have no idea what the sources for those are as this is the first time we've included other firms in the table. I also have no idea how to make those charts. Even if I could, it could only be used for the TPP chart. With TPP, you can use just one party's TPP value, meaning you would only have five lines on the graph. For primary voting, you'd need 20 lines to show the five polling firms' results for the four political groupings, half of them bunched around the 10% line for Greens and Other. And again, I don't know how to create what you are looking for. I will try to work out how to do what you want for TPP, but it is not a high priority because there is little added value. What you are asking for are multiple lines moving more or less in the same direction rather combining them into a more squiggly line showing the numbers more or less moving in the same direction. But I'll try to work it out over the next week or so (when I'd normally next update the graphs anyway). -Rrius (talk) 00:25, 12 June 2013 (UTC)

I've managed it. The graph at the right includes the latest Nielsen, but not Galaxy. I've really only uploaded it to ensure it meets your request, so I don't propose to put it in the article until it includes this week's Morgan, Essential and Newspoll. Given that the dates of polling are often not released until a day or so after the headline results, it may be a few days. -Rrius (talk) 06:54, 17 June 2013 (UTC)

The solution to this is not to join the points up with lines, which would remove the implication that it is a series of polls which should be directly compared. — Preceding unsigned comment added by 129.219.108.243 (talk) 14:06, 27 June 2013 (UTC)

That sounds like nonsense to me. The whole point of a chart that shows the polls over time is to compare them over time. -Rrius (talk) 10:32, 1 July 2013 (UTC)

Charts for latest polling

I understand a Nielsen poll is supposed to come out very soon. I intend to wait 12 hours or so before updating the charts. If by that point no Nielsen results are forthcoming, I will update the primary and 2PP charts, but not the other two. The reason is that a single data point does not show up on its own, so there is really no point until there are two data points. But if no Nielsen poll comes out over the next few days, I'll just use Gillard's last result as a dummy for Rudd until such time as there is a second genuine data point for Rudd. -Rrius (talk) 10:35, 1 July 2013 (UTC)

Glossary

I have added a glossary to explain the names and abbreviations. I know this page is meant to be a kind of "data sheet" to support the main article about the election, but lay readers do follow external links directly to here without going to the election article, and it would be helpful to them to understand some of the terms used, with appropriate links to main articles. --PalaceGuard008 (Talk) 09:15, 5 July 2013 (UTC)

"Non-regular" polling should be included

I loathe to use another country's example with Nationwide opinion polling for the United States presidential election, 2012, but seeing as that's the way this article is heading, why not. Speaking of why not, why are we choosing only to include "regular" pollsters? Timeshift (talk) 07:56, 22 July 2013 (UTC)

Actually, it was out of respect for you. It was a big move to start including the other firms after just including Newspoll, so I've understood the consensus as only extending to the firms we actually agreed to include. If we do add others, we should also put the new firms in the "Polling firms" section, which I intend to make into a "Methodology" section, but I keep procrastinating.

If we are going to include ReachTEL, I would like to find sources for their polls going back to the last election. I made a quick attempt and only found the most recent table, which included polling back to February or March of this year. I tried pouring through Poll Bludger, but wasn't successful.

This does add the question of whether I should crawl back through the Essential tables and pluck out the leadership polling for the instances they actually ask the questions. Thoughts? -Rrius (talk) 03:12, 23 July 2013 (UTC)

A related question is how to handle the other firms in terms of the TPP graph, which separates out the various firms. For this election in particular there is a danger of the table (especially the key) becoming unwieldy because of the multiple varieties of Morgan polls used until the multi-mode was rolled out a few months ago. -Rrius (talk) 03:21, 23 July 2013 (UTC)

The graphs are an obvious exception to including all. Otherwise you would indeed have too many lines, and inconsistent timelines for them. Timeshift (talk) 07:09, 23 July 2013 (UTC)

What about the other questions? These are—

1. Should the "Polling firms" section be expanded into a full-on treatment of methodology?
2. Do you have any ideas about finding old ReachTEL polls?
3. Should I look through the Essential polls for leadership polling and include that?

Thanks. -Rrius (talk) 11:06, 25 July 2013 (UTC) -Rrius (talk) 11:06, 25 July 2013 (UTC)

Campaign period tables

One thing I've seen elsewhere are separate tables showing polling after the writs drop. This helps highlight the changes during the official campaign period, when polling is more often more frequent. It is also helpful because it brings polling during that period into sharper focus. The idea here is not to cut out the full tables, but rather to supplement them. I foresee doing this only for the voter intention tables (or even just the TPP table if that is what people want) because I don't predict much call for separate leadership tables, but I'd be perfectly willing to do all four if there is interest. -Rrius (talk) 03:19, 23 July 2013 (UTC)

I have created the tables, and I will add them unless there is an objection. I will add them faster if there is support, though I don't anticipated adding them to the article before the next polling round. The images are, obviously, to the left, and "Campaign" will become "Election" on the primary chart at the next upload. -Rrius (talk) 08:29, 5 August 2013 (UTC)

Why are there acronyms in the table?

The acronyms in the table look bad. There is enough room for coalition, Labor, Greens, Other etc. Please fix it.

131.217.6.8 (talk) 03:53, 25 July 2013 (UTC)

Disagree. All Australian election polling tables on wikipedia are abbreviated. Timeshift (talk) 07:04, 25 July 2013 (UTC)

There are no acronyms used in the tables. Abbreviations (including the initialism "ALP") are used because headings with full names unnecessarily widen the columns, putting too much white space between the numbers. That, in turn, makes the important part of the table (the data) harder to compare across a given row. -Rrius (talk) 11:01, 25 July 2013 (UTC)

I have inserted explanations of these abbreviations precisely to address the issue that readers unfamiliar with the context would not immediately pick up what "Coa" means, or "Gre". Rrius, I have an issue with the way you revert positive improvements using the "revert" button and without explanation. If you have an issue with the explanations, you are welcome to discuss them here, or at least leave a comment in your edit summary. The way you have gone about reverting others' contributions without explanation is not helpful. --PalaceGuard008 (Talk) 12:44, 29 July 2013 (UTC)

First of all, IP's problem wasn't that readers wouldn't know what the the abbreviations were, it was an aesthetic problem. Second, your "key" is only useful is someone looks at the section lead, and anyone who looks at the section lead doesn't need help to figure out what "Coa", "Gre", and "Oth" are short for, or if they do, they lack the basic mental faculties to make use of the table. Finally, I take issue with the way you generalise one revert into a pattern of behaviour. If you have a problem with the way a particular thing was handled in a particular edit, that's fine, but if you are going to make a broad claim, you should probably look at the other edit summaries first.

Won't comment on the personal dispute there, but I would definitely prefer better versions of the party names. Hard core aficionados such as regular editors here are often lousy judges of what mental faculties our typical readers have and should have. Make it easy for our readers. Don't assume much prior knowledge. HiLo48 (talk) 05:05, 30 July 2013 (UTC)

What exactly are you proposing? I maintain that putting the abbreviations in the section lead is useless. If we genuinely think people will not understand the headings, then filling out the headings or adding a proper key are the only answers. A person who doesn't understand them who then goes and reads the paragraph above the table will come to understand what the abbreviations mean regardless of whether PalaceGuard's abbreviations are there. This is what I am talking about when I talk about people's mental faculties, not whether they would understand the abbreviations in the first instance. I'm not sure that anyone who comes to this article would do so without already knowing what the parties are (or at least gaining understanding from the colours used in the table). But again, if we are to assume otherwise, PalaceGuard's solution is not the right one. -Rrius (talk) 05:13, 30 July 2013 (UTC)

Probably not. I'd prefer clearer, more complete headings. Can we squeeze more detail in with a smaller font? HiLo48 (talk) 05:23, 30 July 2013 (UTC)

Smaller font.......? Timeshift (talk) 07:48, 30 July 2013 (UTC)

Call me a pedant, but I have issues with using abbreviations without explaining them. A key is one solution, explaining it in prose is another. I'm happy with either, but leaving them unexplained is not ideal.

The issue is not about the mental faculties of readers. These articles are not IQ problems. They need to be adequately self-contained or equipped with links so that readers, even those less familiar with the context, can absorb the information without having to stop and solve word puzzles.

One issue is that abbreviations like "Coa" and "Gre" are relative neologisms for the typical reader. Informed Australian voters (who are the main demographic likely to be interested in this article) are used to the way the parties are abbreviated on TV on election night - i.e. ALP (for Labor), LIB (for Liberals), NAT (for Nationals), DEM (for Democrats), GRN (for Greens), and OTH (for others).

Compared to this more usual set of abbreviations, using "Coa" for Liberals and Nationals combined, and using "Gre" for Greens, especially, are relative neologisms. You may be shocked to hear this, but there are plenty of people, even Australian voters, who are not fully aware that they are voting between the ALP and the Coalition - they think of the major parties as Labor and Liberal. Many less informed voters aren't even familiar with the term "Coalition". To them, figuring out that "Coa" means (what they think of as) "Liberals" is, I would suggest, one step too many.

(I am leaving the point critiquing your editing style - it was intended constructively, and now that you are aaware of it, I won't labour the point.) --PalaceGuard008 (Talk) 13:56, 30 July 2013 (UTC)

For the second or third time, the question of mental faculties was strictly about people who read the section lead. If they read in the prose that the polling involves the Coalition, Greens, and Other, then they are going to get what "Coa", "Gre", and "Oth" refer to. Thus, adding an explanation in the prose is completely useless: people who read the prose will understand anyway, and people who don't won't be helped.

If we want to be pedants, I'm not sure the abbreviations are exactly neologisms because they aren't actually words. But I take your point; they are atypical. I didn't come up with them and don't really care what abbreviations we use, if any. Would "Labor, L/NP, Grn, Oth" need explanation? Anyway, I'll just repeat my point that if we use them and feel the need to explain them, it should be in a key.

HiLo asked whether a smaller font for the headings would be a way to use the full names without creating problems. I don't think so. For one thing, the headings would either be smaller than the data or the data would have to shrink as well. I don't think either is a good option.

I've tried using both full names (Labor, Coalition, Green, and Other) and using L/NP for the Coalition. The former doesn't look too bad, but in the latter version, the columns look equal in width. The single abbreviation could be marked with a footnote along the lines of the "Remainder uncommitted" note in the BPM/satisfaction table.

(Finally, on the "editing style" point, there is nothing for me to be aware of. One edit does not a style make.) -Rrius (talk) 10:41, 1 August 2013 (UTC)

Both of those sound like good options (although I'd be interested to hear what HiLo48 or anyone else thinks). Also, I think that if we are able to wikilink the table headings to the relevant party articles that may obviate the need for a key. Your views? --PalaceGuard008 (Talk) 08:47, 5 August 2013 (UTC)

I went ahead and did the "Labor, L/NP^*, Green, Other" version since no one objected. I'm really not sure if linking is enough. And would we link "L/NP" to both parties or to Coalition (Australia). At the moment, all three are linked in the foot note at the bottom of the table. -Rrius (talk) 09:03, 5 August 2013 (UTC)

Thanks, I like what you've done. --PalaceGuard008 (Talk) 09:59, 5 August 2013 (UTC)

Line or scatterpoint?

I would like to know which version is preferred, the current line chart, or a scatter point. -Rrius (talk) 08:03, 29 July 2013 (UTC)

Ideally (the most conservative approach) the data points from each source should be its own series and joined.

The line chart implies that the data points from all sources form the same time series, which just on observation appears to be violated at least in recent times, and there are systematic differences between sources that are, prima facie, attributable to different methods used.

On balance I prefer the line chart, because the data taken as a whole is more or less a time series, and we are mainly analysing for trend not bivariate correlation when we look at these charts. --PalaceGuard008 (Talk) 14:21, 30 July 2013 (UTC)

What you call the most conservative approach would be possible, but difficult, as it would mean hours of work moving data around. The variety of Morgan polls (and the need to go back and separate the phone from the face-to-face polls) only compounds that. For the most part, the upper part of the chart, where Labor and the Coalition results are, would be readable. The portions for late 2010 and the current post-Ruddemption polling would likely be difficult to read. For a taste, imagine the current TPP chart, but with twice as many lines sitting in the same space, and about half as large (because the graph would cover 0–60% rather than 35–60%). The area around the 10% line would be intolerably cluttered as it would have all of those lines occupying a very small territory. That's why I didn't change the primary graph when I changed the TTP one.

What was possible to do without much effort was to retain the data points and add the lines back in. Adding black outlines to the point markers help keep them readable (otherwise they blend with or are even covered by the lines). If you're interested, I can upload it tomorrow as I don't have the time just now. -Rrius (talk) 10:20, 1 August 2013 (UTC)

When I said "ideally", I only meant from the perspective of keeping statistical assumptions to a minimum. I totally agree that many lines for many data sources would not be practical here.

I've given more thought to the issue of scatterplot vs line graph. One issue also is that the surveys are conducted over different periods, which often overlap with each other. That's a consideration that weighs towards scatterplot and against line graph.

The dot-and-line sounds like it could be the best combination. --PalaceGuard008 (Talk) 08:50, 5 August 2013 (UTC)

I just did the update with a dot-and-line version, so let me know what you think. -Rrius (talk) 09:00, 5 August 2013 (UTC)

It definitely makes the data points easier to see. Thanks! --PalaceGuard008 (Talk) 19:39, 6 August 2013 (UTC)

Another alternative would be a scatterpoint with trend lines instead of connect-the-dot lines. -Rrius (talk) 01:03, 7 August 2013 (UTC)

That Morgan phone poll

One editor added, and another editor has removed, the Morgan phone poll conducted on 12/13 August and released the 14th; it was the first Morgan phone poll released since November 2012. It showed primary votes of 41 (Labor), 52 (Coalition), 9 (Green), and 8 (Other). Its voter-response TPP was 57/43 to the Coalition, with no TPP based on last election preference flows reported. This is possibly because those flows are 59/41 to Labor.

The first question is whether the poll should be in the table. The arguments I see are (1) it is clearly a rogue; and (2) it does not report a last election TPP. The arguments for are that it we include polls, regardless of our opinions (however valid) about their quality. As for the TPP issue, that can be handled with a note or by calculating it ourselves. Applying a math formula is not OR, so it shouldn't be a problem.

That leads to a second question, if we include it, what do we do about TPP? And finally, if we include it in the table (especially if we do so with the voter-response TPP), should it stay out of the charts? -Rrius (talk) 12:55, 15 August 2013 (UTC)