Talk:1950 South Australian state election

Differing state wide 2PP estimations - which source to go with?

Antony Green of the ABC has a page with an LCL 2PP of 51.3% in 1950 and 47.0% in 1953, however this publication (pg 10-11 tables) of three members from the "Association of Professional Historians" give an LCL 2PP of 52.6% in 1950 and 46.1% in 1953, and additionally provides 1944 and 1947 2PPs, while Antony only starts from 1950. It was only recently I came across and added the 1944 and 1947 2PP data from the latter publication. Currently, the latter link is used for 1944 and 1947, the former link from 1950 onward, and I came up with the 1950 2PP swing by taking the 1950 2PP from the 1947 2PP (as done federally, pre-1949). I tend to think of Antony as the go-to for election stats which is why I left the 2PPs in the 1950 and 1953 articles unchanged, however upon further consideration, Antony doesn't specialise in SA elections whereas the three members of the "Association of Professional Historians" appear to be based in Adelaide. So my question is, should we consider changing to the latter ref... where the LCL 2PP in 1950 would change from 51.3% to 52.6% and 1953 from 47.0% to 46.1%, and adjust the 2PP swings for 1950, 1953 and 1956 accordingly? Timeshift (talk) 15:04, 31 January 2016 (UTC)

I think the APH paper makes a compelling suggestion that the state 2PP is manifestly useless for a lot of this period: you've still got swathes of uncontested seats: indeed, they point out that in 1956 a third of the electors didn't vote due to uncontested candidates. The broad 2PP is a far more modern obsession than this, and only really functions unproblematically once places got into the habit of both contesting every seat and doing so in a stable two-party system.

My preference would be to cut the state 2PP from the infoboxes until SA politics evolved to a point where it made more sense as a statistic, which that paper at a glance pushes into the 1960s or 1970s, and focus our results data on Dean Jaensch's figures (much as I whinge about his attention to detail about party status in the early years, he's an unimpeachable statistician to my knowledge). I would use that as a starting point, but it might also be worth asking Antony what his opinion of the APH figures are - he seems to be increasingly accessible on social media and his take on those would obviously be relevant (and, equally obviously, supersede my judgment). The Drover's Wife (talk) 16:35, 31 January 2016 (UTC)

Have you read the second publication (pg 9-10 just before the tables) that explains how they got to the 2PP figures they did? And this was back in the day when it tended to be that Labor voters were Labor voters, Liberal voters were Liberal voters, and swinging voters were few and far between. I wonder how Antony came up with his? I'm certainly up for adding a more detailed sub-note calculation explanation to the earlier statewide 2PPs from 1944 up to and including 1973, however they are certainly valid for what they are, just like the federal nationwide 2PPs from 1937 up to and including 1969. Timeshift (talk) 17:42, 31 January 2016 (UTC)

I'm not arguing with how they got to the figures they did (statistics is not my forte and they provided a perfectly reasonable explanation of the challenges inherent in collating this figure), but I do think they're not valid in the current infobox-level context, even with a footnote. It is completely reasonable to include a paragraph in the results section stating the APH's figures, Antony's figures, and explaining the headaches for providing such a figure at that election, since they were many and varied and the end results are still likely to be of interest. Adding it to the infobox as if it carried the same sort of clarity it does now back then is misleading, even with a footnote, and involves dismissing Antony Green's figures without solid reason when they are obviously the subject of reasonable dispute. The Drover's Wife (talk) 21:26, 31 January 2016 (UTC)

Are you referring to just the SA 2PPs or federal as well? Federally, for example 1969, we've got on just fine with having 2PPs for an entire decade without dispute. Even 1980, the AEC did not undertake a full distribution of preferences for statistical purposes. The sources appear to be reliable and non-contradictory, barring the two exceptions I raised above for SA state 1950 and 1953. Perhaps if you're so keen to see a change after an entire decade, you might be interested in contacting both sources and see if they cant come to some sort of agreement. The last thing we would want to do is remove the 2PP from the result table and the infobox on every election article federally pre-1983 and SA state pre-1973 and relegate them in to a blur of paragraphs that most viewers likely do not read. The 2PPs, estimates as they may be, are too important for such a relegation. There's various sources out there that say in a fair system, based on votes, the LCL should not have won 1944, but if you just look at the primary votes you wouldn't come to that conclusion. And if we extend your theory, at least the 2PPs are a statewide estimate which more accurately tell the tale than primary votes. The most extreme example is 1956 where 26 of 39 seats (66% of seats) were not contested by both major parties, resulting in just 39.2 percent of voters throughout the state being given a choice between the two major parties, and in 1953, leaving aside the 9 uncontested seats, 15 of 39 seats were not contested by both parties. How can you possibly say that under these conditions, the primary vote is more meaningful or more result table worthy than the estimated statewide 2PP? Whether you want to look at primary votes or the pendulums in the 40s/50s/60s, it's a mess, but it is what it is. They are so disproportionate that somehow the estimated 2PP is the most indicative/telling/whatever of the lot! Think about it for a bit longer if you don't think I haven't given you some points to reflect on that you previously didn't consider. I fail to see the concern of adding some form of unmissable asteriskish character next to the 2PP, linking it to a footnote indicating it's an estimate. On that logic, why don't we mark every primary vote figure as they had even less meaning? Over 60 percent of all SA voters not getting a choice between the two parties, sorry but that has to take the cake for an example of primary votes being grossly more meaningless than 2PP by a country mile! But, if it comes down to just SA 1950 and 1953 for no other reason than because we have two WP:RS, Antony Green is by far more authoritative simply based on how many media outlets have quoted him for such a long period of time, and I would have no issues using his figures as the default for SA 1950 and 1953 - I was just asking the question, I certainly wasn't advocating switching to the other source. Ultimately, remember that it is an educated estimate by WP:RS professional academics. It's not a guess, we are not engaging in WP:OR, and we are referencing what seem to be WP:RS. Timeshift (talk) 22:19, 31 January 2016 (UTC)

I am not hugely invested in what we do about estimated 2PP figures where reliable sources don't disagree: if we cite the agreed figure and footnote it noting that it is an estimate, that is completely fine by me. However, we can't do that where reliable sources do disagree, and it's hardly a great sacrifice to leave it out of the infobox and explain what the sources have to say in text in the results section. In this case, they also disagree for really obvious (and reasonable) reasons, which makes it all the more important to explain the differing figures and a brief summary of why they're different (based on the verifiable reasoning in the APH paper). Otherwise, we are getting into sheer original research - we're picking sources based not on any evidential basis but on whether it makes our infoboxes look pretty.

There are absolutely and obviously sources that point out that the 1944 result was questionable, but again that's something we need to explain in text (and do a good job of because it's important!) unless reliable sources agree on a specific figure we can use to simplify things. In those cases of very few contested seats, the primary vote is the only definitively useful thing because it's undisputed fact; the circumstances necessitate that there is no obvious way to come up with an undisputed estimate, and that means we explain the issue and the differing conclusions. The end result gives by far the most informative and useful article, containing all the necessary information - it just doesn't provide a pretty figure for the infobox. The Drover's Wife (talk) 23:24, 31 January 2016 (UTC)

They are both correct in their own methodology, and as such, this is a case where the two sources don't need to agree on a specific figure. If they used the same methodology but got a different figure, you'd be right. But different methodologies give different figures, and both are correct. Every 2PP in SA pre-1975 and federally pre-1984 is disputable and an estimate. But considering we use Antony Green's methodology for every SA election from 1950 onward, in comparison to the Adelaide "Association of Professional Historians" that doesn't even get a mention on news.google.com.au (enter Antony Green in to news.google.com.au and you get "About 129,000 results"), it's more than reasonable for both consistency and widespread ref use on wikipedia to use Antony Green's 2PPs as the main 2PP figures where available, which in the case of SA goes back to 1950 inclusive. Why would we relegate Antony Green's 2PP figures for 1950/53 from the result table and infobox to hide amongst the text paragraphs just because I came across the Adelaide "Association of Professional Historians" who used a different methodology to get a different figure? I'm sure there are several methodologies that could be used to come up with several 2PPs for any given result. Are we really thinking of relegating the 1950/53 2PPs to hidden obscure body text status just because someone as comparatively unknown as the Adelaide "Association of Professional Historians" has used a different methodology? The more I think about it, the more I shouldn't have raised this. One thing wikipedia is all about is consistency. We have Antony Green's SA 2PPs going right back to 1950, and I challenge you to say that because of the Adelaide "Association of Professional Historians" differing methodology, neither should be in the election table or infobox. We have the almighty Antony right back to 1950, and two elections prior to that we have another source. What is incorrect, misleading or improper about using Antony over whoever they are as the main figure? As I said, I have no problem with a big asterisky mark and a footnote with the Assoc Pro His 2PP, or perhaps even directly under the election table where it says "The LCL did not contest eight seats, Labor did not contest ten seats. The primary vote was counted on seats contested, while the two-party vote was estimated for all seats" and provides the far-lesser known source's 2PP? I'm happy to compromise in any sensible way, but i'm drawing the line at the very unreasonable exclusion of any 2PP from the election table and infobox altogether. We have Antony Green's data, published by the Australian Broadcasting Corporation who were founded in 1929, is our national equivalent to the BBC, which attracts millions of viewers (a day?) through various methods. All this 2PP complete relegation nonsense because for some now-clearly-stupid reason I thought i'd bring to attention the whoever-they-are's 2PP figures which covers just a couple of elections. In reflection, at most, it is something of a secondary factoid nature in comparison to what we already have. As I said, happy to compromise in pretty much any way, except complete relegation of the 2PP from the result table and infobox in to the abyss of the body text, when the source we've used for a decade, Antony Green's data, published by the Australian Broadcasting Corporation, which attracts millions of viewers (a day?) through various methods. Apart from electoral commissions, if you want a source as authoritative as Antony, you're out of luck. Let's get real... to let the Adelaide "Association of Professional Historians" de-rail both their and Antony Green's data is ludicrous. Please, compromise. Or better yet, wait a day or three, because i'm getting a headache over this. Timeshift (talk) 01:24, 1 February 2016 (UTC)

I am thoroughly confused as to why you brought this up, in this case. They are a reliable source. Antony is a reliable source. The obvious thing to do is to mention both and mention their different results and methodologies, and no one is going to be terribly distressed that there isn't a 2PP figure in the infobox. If you want to ignore them in the name of simplicity, I could care less - at least with Antony we're working off reliable estimates - but then why bring the alternative dataset up if you didn't want them mentioned? The Drover's Wife (talk) 09:13, 1 February 2016 (UTC)

I suppose I gave them exposure because it is extremely rare to find election-wide 2PPs outside of electoral commissions and Antony Green. I thought it was worthy of exposure here. Perhaps I wanted to see if, unlike me, others knew of this "Association of Professional Historians" and whether or not there were arguments not immediately obvious to me as to worthiness of some form of inclusion. At no time did I expect anyone to respond to their existance with "let's not list either 2PP figure in either the result table or infobox anymore". Talk pages are a good place for interesting, on-topic discussion of discoveries like this, and to collaborate which can lead to ways of inclusion. I'm open on various ways to incorporate both sets of data. It seems you're set on your one desired aim of result table and infobox exclusion. Nobody has indicated any knowledge of the "Association of Professional Historians", confirming what I suspected, which is that Antony is inherently more reliable. As we do with every other state and federal election article, on reflection I think we should continue the status quo and use Antony's 2PP data where available. But i'll suggest yet another proposal. How about we use Antony's 2PP in the infobox, remove the 2PP from the result table, but add another table directly below the result table - a very simple one of just two rows. The first row with Antony's figure the second row with the "Association of Professional Historians". We could then explain the situation either directly below as an add-on to "The LCL did not contest ten seats, Labor did not contest thirteen seats. The primary vote was counted on seats contested, while the two-party vote was estimated for all seats", or we could use a footnote. Objection? Timeshift (talk) 10:30, 1 February 2016 (UTC)

I just said I didn't have a problem with this? You're reading an aggression into all my posts today that is really not there. You asked on my talk page for my opinion, I gave it, but since you have a strong objection to said opinion this is a completely acceptable alternative. The Drover's Wife (talk) 10:33, 1 February 2016 (UTC)

Sorry but that's incorrect. You were set on one method - having any and all 2PP figures removed from the result tables and infoboxes, and have them relegated to often-unread article body text ONLY which is often glanced over/not read. What I just proposed is we use Antony's 2PP in the infobox, and create a second table of two rows for both 2PP figures which almost attaches to the bottom of the election result table. This would include both 2PP figures, but still appear where they're expected to appear, directly underneath the primary vote rows, and in the infobox, and as such would basically remain in the positions they currently are, and therefore impossible to miss. If you want to claim you have no problem with this, then fantastic! The SA and fed issues have been resolved within minutes of each other. Fantastic outcome. Thank you. Timeshift (talk) 10:55, 1 February 2016 (UTC)

Hon, you came to my talk page and asked me for my opinion, and I gave it. It just happens to be an issue I don't have strong opinions on, and am in no way going to die in a ditch over, so "let's not do that" is...fine? The Drover's Wife (talk) 11:03, 1 February 2016 (UTC)

"Let's not do that" would indicate you want to keep the status quo. Forgive me, i'm getting heavy wiki-fatigue and should log off for a while, but i'm sure that in this discussion, your position is that we should remove Antony's 2PP figures from the result tables and infoboxes, and relegate them in to article body text along with an inclusion of the new 2PP calculations. I'm saying we should keep the 2PP, or the 2PPs, prominent, and have given suggestions for keeping 2PP figures in the infobox and visually remain part of but not technically remain part of the same result table, with two table rows indicating Antony's 2PP in the first row, and the "Association of Professional Historians" 2PP figure in the second row, with a footnote, or failing that a sentence added to the end of the "The LCL did not contest ten seats, Labor did not contest thirteen seats. The primary vote was counted on seats contested, while the two-party vote was estimated for all seats" bullet-point directly below the result table, distinguishing the two. Bearing in mind my fatigue, please bear with me... am I still misinterpreting what you said... I read that you wanted to remove the 2PP figure from the result tables and infoboxes for 1950 and 1953 entirely, and move them to the article body text (ie: below the TOC). I want to keep the 2PP prominent in the result tables and infoboxes and suggested ways to do that above. Is that a reasonably accurate summation, or have I completely lost the plot and would be best logging off for the night? Timeshift (talk) 11:46, 1 February 2016 (UTC)

It means "you've correctly interpreted my first preference, but it's just a preference and I don't particularly care either way, that in this case I probably wouldn't have bothered chipping in if you hadn't asked, I gave an opinion since you did, and...am not particularly bothered if you'd rather do something different!" The Drover's Wife (talk) 11:49, 1 February 2016 (UTC)

Fair enough. I asked for your opinion and I got it. Perhaps I was being too narrow in what I was asking in the original post. I've provided rare 2PP calculations for the very small interested segment out there and for now i'm more than happy to move on. Timeshift (talk) 12:21, 1 February 2016 (UTC)

My opinion on the matter is that whichever 2PP figure is chosen by the most people, that would be fine. It seems like it was simply an educated estimate in both cases of Green and the AoPH (although that paper cites a journal article), and it only seems to be out by 1% or so so it doesn't bother me either way. Probably the most tried and true of estimating the vote in seats uncontested by both parties would be to look at how they voted in Federal Senate elections around the same time. Whoever did it with that method would probably be the most accurate (although even then, there's the difference of state and federal issues but it's just simply too tricky to adjust estimations for that).

I apologize if I missed that in the conversations above, it was a lot to read through. Kirsdarke01 (talk) 07:26, 2 February 2016 (UTC)