This page is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
RfC on Scotland Act and Royal Assent links for the s35 note.
Personally I favour using the internal wikilinks rather than external links. I general I would think that is inline with usual practice but I'm not really one to go trawling around all the wiki policies, just an observation for how things normally are. Grinner (talk) 21:25, 24 January 2023 (UTC)
MOS:OVERLINK begins with the sentence External links normally should not be placed in the body of an article (see Wikipedia:External links).WP:EL has External links normally should not be placed in the body of an article. near the start. If we have an article on the topic, there is no reason why that shouldn't be linked. --Redrose64 🌹 (talk) 22:13, 24 January 2023 (UTC)
Yeah, without intending offense to anyone, I'm gonna assume this is a dispute between relatively inexperienced editors, because this is a clear WP:SNOW call: outside a handful of niche contexts, we do not utilize external links in the body of an article, and certainly not prose sections. Meanwhile, this would be an entirely appropriate application for internal links or pipes. SnowRise let's rap 22:48, 24 January 2023 (UTC)
It's a dispute between me and User:Soosrider3, who insists on treating this note like a poll source and putting external links to those topics to "provide clear data" and claiming that consensus is against using internal links. An obvious WP:SNOW, I know, but I don't know what else will convince them. Glide08 (talk) 05:53, 25 January 2023 (UTC)
With all due respect that is a very sweeping and comprehensive assertion "we do not utilise external links" especially when the wiki article referred to explicitly states the opposite using phrases such as "Wikipedia articles may include external links, links to web pages outside Wikipedia" "Some high-value external links are welcome" and most importantly "Besides those kinds of links listed in § Restrictions on linking, these external-link guidelines do not apply to citations to reliable sources within the body of the article.
Although I welcome other views those expressed so far are IMHO in danger of selectively quoting guidance without placing it in the context of the whole article and suspending 'common sense' . In this particular this is a highly technical matter regarding the law and is best served by linking directly to those statutes.
I look forward to your response. Soosider3 (talk) 08:23, 25 January 2023 (UTC)
Again, I'm afraid you are misreading the policy in question: WP:ELNEVER is about types of external links which are not permitted even in the dedicated external links section, certain infobox fields, or citations, where external links (those varieties which are permitted) belong. It does not in any way create an exception to the rule that external links are not permitted in the main body of the article (outside the previously mentioned niche exceptions). But yes, as to using external links inside citations, that is certainly permitted (expected even, when a link is available for a source). If you want to include the URLs inside citations here, that is probably permissible in this case, though probably you'll want to include secondary citations to support the statement in question as well. (see my response below).SnowRise let's rap 09:06, 25 January 2023 (UTC)
Thank you for your input, however I do believe that there is a danger of selectively quoting wiki guidance, further reading of the topic you linked to shows that far from being a hard and fast rule it is a guidance and common sense should be shown. In particular your reference goes on to say "Articles on technical subjects might demand a higher density of links than general-interest articles, because they are likely to contain more technical terms that general dictionaries are unlikely to explain in context." and would suggest to you that definitions of statutes are highly technical and are best linking directly to that statute rather than an opinion piece. The actual article is fairly technical being about polling on a particular subject, as such it is predominately contains external links. I belive that in this case teh use of external links is best practice, but would welcome your view Soosider3 (talk) 08:07, 25 January 2023 (UTC)
It is in fact about as hard and fast a rule as we have. While technically any editorial practice we have is amenable to WP:IAR, WP:guidelines represent established community consensus, and this particular guideline has existed for nearly two decades. I am unaware of a single example, since this became policy, where a WP:LOCALCONSENSUS has permitted the use of external links above the line in the main body, aside from the established carveouts for infoboxes and citations. Also, the section of the policy you quoted (WP:LINKSTYLE) is very unambiguously talking about internal links only. Candidly, there is functionally zero chance of the outcome you are advocating for here. That said, there's an argument to be made for using the links in question as WP:primary citations: although WP:secondary sources would be preferable here, using the primaries may be permissible as auxiliary. WP:verification. SnowRise let's rap 09:00, 25 January 2023 (UTC)
Right, now I am totally baffled, granted its early here and i may not have got my caffeine level up yet. However both the links you sent speak very specifically about there being no hard and fast rules.
The knub of this disagreement seems to hinge on internal/external links, where it has been stated that links MUST be internal, I find nothing to support that. Rather I find guidance that suggests that external links are perfectly acceptable and in many circumstances are preferable. The thread that runs through all these guidance is one of Common Sense and best practice, in this particular case I think common sense is to go straight to source providing reader with clear and unambiguous further data as to what the legislation actually says. That surely has to be best practice to provide readers with best and most reliable information. Soosider3 (talk) 09:46, 25 January 2023 (UTC)
In this case, the only parts where "what the legislation actually says" would be relevant is Section 35 specifically and the statement of reasons for the s35 order. Not for the Scotland Act 1998 in its entirety or the process of Royal Assent. Glide08 (talk) 11:16, 25 January 2023 (UTC)
Have to disagree all 4 components are legal processes, if link to statute is good enough for s35order and s35, then why would they not apply to the 1998 Act and Royal Assent, that seems a might inconsistent Soosider3 (talk) 11:30, 25 January 2023 (UTC)
Because a description of a stage in the law-making process and a full text of a law that includes a hundred sections is more tangential than the order blocking a law from being made and the specific section of a law that authorizes it. Glide08 (talk) 21:02, 25 January 2023 (UTC)
"Right, now I am totally baffled . . . both the links you sent speak very specifically about there being no hard and fast rules."
Yeah, I definitely understand why this could be confusing: the application of WP:IAR is very nuanced, and typically takes a fair bit of experience to fully understand. But I'll try to parse the distinction as expressly as I can. IAR has been maintained since the earliest days of the project as a reminder not to let arbitrary adherence to established rules stand in the way of development of novel approaches to new problems. However (and this is a big however), the longer an editorial principle has been enshrined in policy, and the more uniform its support in community WP:CONSENSUS at large, the higher the burden you face when you try to convince a WP:LOCALCONSENSUS to overturn that best practice guidance.
Here we are talking about a rule that was established very early in the project's history (for a bevy of reasons), and which is pretty much universally supported by the community at large. The likelihood that a variance will be supported here is exceptionally low, especially because there are alternatives for sharing this information, as I will discuss below. IAR is usually applied in highly novel circumstances, and there is nothing particularly novel here: the community long ago decided how it wants to handle situations like the one you are describing in this context, which is why the guideline reads as it does, and has done so for a very long time.
"The knub of this disagreement seems to hinge on internal/external links, where it has been stated that links MUST be internal, I find nothing to support that. Rather I find guidance that suggests that external links are perfectly acceptable and in many circumstances are preferable."
Again, I think you may need to re-read the policies in question, because you seem to be getting a skewed perception of what they actually say, and that might be confusing the feedback you are getting here from other editors as well. It's not that external links are forbidden from appearing in the article: we do find utility in many, many kinds of external links. However, policy is unambiguous that external links do not belong in the main body of the article. Rather, the vast, vast majority of external links appearing in an article occur in one of two contexts: 1) inside citations (and such links appear in the references section, but not in the prose of the article itself), or 2) In a dedicated external link section, which is one of a variety of sections that can be added below the main body of an article. This is expressly set out in the first sentence of WP:External link, as well as numerous other guidelines and WP:Manual of Style sections that reference it.
Pretty much the only exception to this rule about avoiding ELs in the main body is that we sometimes allow links to the official webpage of an organization or professional in the WP:infobox for the relevant article, if the article has an infobox. Other than this, there are no real exceptions to the rule that external links are proscribed in the main body: or at least none that I have ever seen survive scrutiny from the community. So where you are seeing language in the guidelines that is discussing which external links are permitted and which aren't, it is discussing all such linking exclusively within the contexts to which WP:EL limits external links. In other words, such language is discussing which external links should appear in the dedicated external links section or an infobox, and which should be omitted from articles altogether. And no, nowhere does any policy entertain that an external link can be considered as a valid alternative to an internal link. That's just not how prose works on this project: we use only internal links and pipes in the main body of the article (again, aside from limited uses in infoboxes).
All that said, appropriate URLs to the location of valid WP:RS are pretty much always permitted inside a citation. But as a matter of nomenclature, we almost never call that an "external link" but rather a "citation URL", "citational weblink" or something along those lines.
"in this particular case I think common sense is to go straight to source providing reader with clear and unambiguous further data as to what the legislation actually says. That surely has to be best practice to provide readers with best and most reliable information."
You're sort of correct here, but with a big and important caveat. We absolutely do put a premium on supplying our readers with directions to resources which they can follow up on to better understand our content, which is meant to be an encyclopedic summary of the topic, not the end-point. This is one of the reasons why Wikipedia's approach is so rooted in WP:reliable sources. However, at the risk of sounding like a broken record, we present those resources (including online documentation) in the afore-mentioned formats: references and external links sections. For many reasons (running the gamut from technical issues, to salience, to maintaining WP:Neutral tone and intuitive navigation), we want to clearly demarcate internal and external links, and this is the foremost means by which we accomplish that. I hope that clarifies matters for you and assuages concerns about making this info available to the reader: we absolutely can do that, but just not in the way that you are proposing. SnowRise let's rap 13:14, 25 January 2023 (UTC)
Firstly may I thank you for your full and detailed explanation, and your patience. While this entire article is festooned with external links all of which are necessary to the credibility of the page. I have perhaps become stuck in that thinking, your quote about wikipedia "which is meant to be an encyclopedic summary of the topic, not the end-point" has chimed with me I see the sense in that while at same time still thinking direct links have a very useful place.
I really do appreciate your input, I think teh correct response is that I will drop the stick. Soosider3 (talk) 13:40, 25 January 2023 (UTC)
You're very welcome: I'm glad the feedback came across as fulsome and helpful, rather than dense and inscrutable! For what it's worth, I understand why the guideline language might not make the distinctions at play in this case particularly clear, hence said wall-of-text reply. I wanted to make sure to emphasize that your underlying impulse (pointing the reader to the sources they may reasonably want to follow up with) is absolutely the correct one. In fact, itt's a cornerstone of the Wikipedia process that we summarize and report on what the sources (including the primary sources) say about a topic, rather than rendering what we ourselves believe (or even know). And a big part of implementing that principle is directing the reader to said sources to follow up with and draw their own conclusions. It's just that we've developed a very particular, siloed way of presenting such reference material. SnowRise let's rap 13:56, 25 January 2023 (UTC)
I endorse WP:ELExternal links normally should not be placed in the body of an article and I do not see cause for an exception in this case. I note that the article does contain a bunch of external links in the body - in the tables. Tables are one of the few cases we sometimes accept external links in the body. Those links are essentially serving as refs, and there would be questionable merit in converting them into proper refs. I'm neutral on that.
I interpret Soosider3's last comment as dropping/ending the dispute here, but I don't want to assume. If my impression is accurate then the issue is resolved, and Soosider3 or any editor can remove can remove the {{rfc}}-template at the top of this section. Alsee (talk) 22:46, 25 January 2023 (UTC)
Thank you for your input, the matter is resolved as far as I'm concerned, however one small comment, I note that tables can be acceptable for External Links and would point out that the disputed links are actually part of a table. But lets not go round again, a suitable position has been struck and I for one will leave it there. Soosider3 (talk) 05:31, 26 January 2023 (UTC)
Timing of a Second Referendum
I had previously noted that I thought this section was at best 'cluncky' and held the view that what data there was seemed to have been shoehorned into the table. I have drafted a new layout for this section and would welcome views on it before I publish
In essence there are only 2 companies polling on this matter Panelbase and Yougov, both ask different questions but with some overlap, I have amended table to reflect this, still a wee bit of shoehorning and I have ignored the additional questions asked at time of last Holyrood election. The column Lead; I have left blank in respect to Yougov as I could not see how this could be worked out
Your views would be appreciated.
I have published these in my sandbox and hope you can see them OK
https://en.wikipedia.org/wiki/User:Soosider3/sandbox#Support_or_oppose_a_second_referendumSoosider3 (talk) 13:56, 2 February 2023 (UTC)
Noted that earlier change to timing of second referendum had led other editor to move one poll from oppose/support to this table. Thinking that through I looked at remaining 2 polls in oppose/support table and realised that they were actually asking very specific questions on timing, realised that with only minor changes to new Timing Table we could incorporate these polls into one Table. Have done so in Draft form but feel that this layout removes much of the clunkiness of previous tables and although a wide table it does fully capture and display the actual questions asked and that the blank cells indicate where such category was not asked. It may also provide an element of future proofing
I would intend to make amendments to main article by replacing Oppose/support table and timing of second referendum table with this new layout. Soosider3 (talk) 20:55, 2 February 2023 (UTC)
I have amended the article as below, the issue was of accuracy of the support/oppose table was raised by me back in 20th december 2022.
My amendemnets are
1 Remove Oppose/support table as wildly inaccurate, all 3 remaining polls have been merged into new Timing Table - so data not lost
2 Amended preamble
3 Published new Timing Table with more comprehensive timeframes
4 Moved UK polling to after Timing Table Soosider3 (talk) 12:37, 3 February 2023 (UTC)
Hello. I was the original creator of the table for timing of a second referendum .. the current one is IMHO a confusing mess, the table I provided was easy to read and similar enough to the main table above. No wonder then I'd like to revert to the easy to read table.. unless the new table can be seriously cleaned up and standardized.. possibly in the style of the easy to read table.
Hi and thanks for the comment, I have to say this was one of the more difficult tables which perhaps reflects the complexity of the questions asked and therefore requires this layout, at least if we want it to properly reflect the actual data.
The layout is far from being a confusing mess, it is more complex than other tables but fairly clear in its layout.
To be blunt the previous table was wildy inaccurate, not just a wee bit but seriously misreported the data. Have a look at the data for yougov poll of May 2022 in the old table it was presented as between 2-5 years and No referendum, when in fact the poll asked about in 2023, next 12 months and 2 to 5 years nowhere was the No Referendum question asked. Or look at the Yougov poll of Mar 2021 again presented as a straight choice between 2-5 years and no referendum when question asked about next 12 months and 2- 5 years. There were other examples of this sort of wild inaccurate reported of data.
There is no requirement to standardise tables, rather they have to reflect the data accurately and fairly, if that means a table like teh current one then so be it. Soosider3 (talk) 15:34, 16 February 2023 (UTC)
Hi and thanks for the comment, I have to say this was one of the more difficult tables which perhaps reflects the complexity of the questions asked and therefore requires this layout, at least if we want it to properly reflect the actual data.
The layout is far from being a confusing mess, it is more complex than other tables but fairly clear in its layout.
To be blunt the previous table was wildy inaccurate, not just a wee bit but seriously misreported the data. Have a look at the data for yougov poll of May 2022 in the old table it was presented as between 2-5 years and No referendum, when in fact the poll asked about in 2023, next 12 months and 2 to 5 years nowhere was the No Referendum question asked. Or look at the Yougov poll of Mar 2021 again presented as a straight choice between 2-5 years and no referendum when question asked about next 12 months and 2- 5 years. There were other examples of this sort of wild inaccurate reported of data.
There is no requirement to standardise tables, rather they have to reflect the data accurately and fairly, if that means a table like the current one then so be it.
Previous version of Timing TableSoosider3 (talk) 15:53, 16 February 2023 (UTC)
Hi. My original table was panelbase only. The inclusion of 2023, less and more than 10 years do not match the other questions and so I suggest make them non-standard and should be excluded. Perhaps a short paragraph talking about those polls results and pointing out they are not standard and uniform.
As for 'no referendum' not being asked.. the table was designed for the panelbase results. I took the other editor's additions in WP:AFG.
I welcome extra work to make what I begun better.. I just don't feel the current table is easy enough to read or succinct enough for such a simple question. Dava4444 (talk) 11:53, 22 February 2023 (UTC)
Hi again
I understand that Table may originally have been designed to capture Panelbase however the number of Polling Organisations asking questions re timing of next referendum has expanded since then, as have the time bands. Previouslu sticking to original formatting led to a very poor and misleading table.
We should be recording the data as asked by the pollsters to do so we need to have the options available as per table.
The idea of a question as standard and therefore all others as non standard makes no sense, the questions asked are not the binary Yes/No and to attempt to do so makes no sense and in fact would bring us back to the previous unacceptable position where data was shoehorned in to such an extent that it, put bluntly, misrepresented the data. I gave previous example where it was recoded as "no referendum" when the Q was not asked, furthermore it was not clear what data was amalgamated to produce the number recorded on the table.
Having looked at every set of data on this table I am confident that this layout is a good and clear representation of the actual data and the questions asked, far from being a simple question it is a complex and varied one.
I really don't understand your rationale for considering excluding polls, these have been completed by reputable Polling companies and deserve to be recorded. Soosider3 (talk) 05:23, 23 February 2023 (UTC)
Hi Been reflecting on your suggestions and looking again at data
There may be some merit in looking again at the 10 year columns, given that so far only one pollster has asked questions in this range and the framing of the question I believe we could comfortably create a column headed say 'More than 5 years' and amalgamate these 2 columns into it, we might need a wee note to explain this, which would require an additional column, this may be beneficial and allow a space for editors to explain any merging of data they introduce.
I agree.. and it's what I meant by standardization. Dava4444 (talk) 17:19, 1 March 2023 (UTC)
I also have doubts about the lead column, do we actually require it? could this become the space for a notes column?
Sadly while true, it would clean up the table a bit.. Many Unionist politicians used this data when it first came out to lie to the public, saying only one of the two columns percentage versus the no percentage. This is why I put the column here in the first place.. those politicians were misleading the public and it is our duty as wiki editors to present facts. Dava4444 (talk) 17:19, 1 March 2023 (UTC)
The 2023 column is tricky, Q first asked over 2 years ago and as we have moved closer to 2023 and now in it the response to this Q and its meaning has changed, I don't believe we could consistently move it or merge this data with any other time scale, also given that Yougov poll of March 2022 only asked this question and very specifically, then we are going to have to retain a 2023 column in which case we should use it. Soosider3 (talk) 08:44, 26 February 2023 (UTC)
I believe for unusual questions/answers we should leave them as a note below the table for people to see for themselves. 2023 is far to unique to need its own column. Like the 10 year column.. I feel very few pollsters will use this going forward. Dava4444 (talk) 17:19, 1 March 2023 (UTC)
Okay, if not in complete agreement we do seem to be thinking that a further tidy up is required, I would suggest
1 We remove both the 10 years columns and replace them with more than 5 years, where longer time sacles are asked about they can easily be amalgamated into this column and clarification put in a notes field.
2 We remove the Lead column and replace it with a Notes column
3 The 2023 column as we both recognise is awkward, however question has been asked as recently as Feb 23, obviously its use will stop by end of year. My thoughts are to just leave it, the Q has been specifically asked and we should respect that.
If this seems okay to you I will make the changes as suggested. Soosider3 (talk) 18:08, 1 March 2023 (UTC)
Hi
I agree with point 1, but do not agree on points 2 and 3.
The Lead column needs to stay due to the lies of Unionist politicians.. but in the interest of good faith.. I would be open to reforming it in to something we both agree on (although I can't imagine anything more simpler than just tallying the totals of the two columns and showing the results)
Point 3
Yes.. we both are seeing this the same way, it's awkward. 2023 is too specific, and if the polls of 2022 are included such as 12 month polls (which currently they are also 1 2021, this doesn't bear relevancy say in 12 months from now in the same those 2022)
I see the problem of the '12 month' vs 'yearly' polls.. and would like to suggest again we stick to 12 months only. You can see how a poll for say 2023 and 12 months from 2022 could overlap right? And then it becomes a rolling problem. Just leaving it seems like just ignoring it.. any suggestions? Dava4444 (talk) 20:12, 1 March 2023 (UTC)
Glad we have agreed on some of the changes, I will merge 10 years plus into more than 5 years and introduce a notes column, and lets see how that looks.
I believe in principle we should limit how much the data is merged/amalgamated and keep it to the absolute minimum, the merging of 10 years is easy and at moment a one off. Doing so with 2023 column is not IMHO in the same and would introduce a degree of interpretation that might make it harder for readers to follow the data. At present we layout teh data and readers can see instantly what timeframes were asked and most importantly whether the timeframes collectively make 100% or if each timeframe 100% unto itself.
2023 is a specific Q and one that will (obviously) peter out, we should just leave it to do so and leave it as an example of how an issue can be clear at a point in time but become irrelevant Soosider3 (talk) 08:48, 2 March 2023 (UTC)
Missing panelbase poll
This panelbase poll from late Feb seems to be missing. I can’t see the actual data but has been reported by What Scotland Thinks
Not sure what is going on here but at this time we do not appear to have any clear data that poll actually happened. Soosider3 (talk) 11:29, 9 March 2023 (UTC)
Yeah I had a quick look and couldn’t find the actual data tables either but I’m on my mobile so wasn’t sure if it just wasn’t working for some other reason! Very strange! Dunk the Lunk (talk) 12:22, 9 March 2023 (UTC)
I kind of suspect that the client may not have released all the information from the poll and is holding back publishing until they do, I think the Times did something similar a few months ago. Soosider3 (talk) 14:19, 9 March 2023 (UTC)
Survation poll 8th-10th March 2023
Surely this Survation poll everyone keeps vaguely referencing (or deleting) must be this one funded by DC Thomson:
Which What Scotland Thinks gives as Yes 40% No 48% Dunk the Lunk (talk) 11:27, 13 March 2023 (UTC)
Shrugs, there maybe 2 polls ghosting around at this moment. It really doesn't help when different news agencies give a few bit of data then conflate it with data from another poll. At this moment (please note the caveat) There appears to be a poll commissioned by DC Thomson that may or may not have actually asked the Indy question. Whatscotlandthinks has it up but link brings you back to same page !!! The other appears to have been commissioned by Diffney partnership which may be the start of a 3 monthly poll from them. We are still waiting for Panelbase poll that had also been ghosting around last week.
Perhaps we need to just wait until we have clear data from whatever source. Soosider3 (talk) 12:09, 13 March 2023 (UTC)
Very likely the same poll as reported in media as it shares the same sample size, dates and polling agency. I suspect the accompanying press release for the poll misquoted the figure excluding Don't Knows as 53% No 45% Yes. AlloDoon (talk) 14:02, 13 March 2023 (UTC)
Graph
Could someone advise me what the criteria is for a poll to be included in the graph. I had a look at the files for it, and if I am reading them correctly, (always a big if) not all polls are used, for example none of the polls in November are included.
I know @RERTwiki does a lot of work on this perhaps he could take the time to advise me
Many Thanks
Gavin Soosider3 (talk) 09:26, 23 January 2023 (UTC)
Hi - Sorry I missed this, I’ve been busy with other…charts! Every time there is a new poll, I paste the whole of the table into a spreadsheet. I go through all the notes, and any poll marked as ‘non-standard question’ is tagged and should be excluded from the charts and averages.
If you’ve seen anything that doesn’t seem to correspond to that, let me know here.
(I paste the whole thing in again to ensure consistency between the two, and catch any historic edits.) — Preceding unsigned comment added by 31.111.26.26 (talk) 19:47, 8 February 2023 (UTC)
Thanks for getting back to me, much appreciated. I was just curious but also wonder if we should have a very brief note preamble to graphs so that readers are clear what they are looking at, reluctant to do that without fully understanding what is in them.
Can a suggest a small tweak to the 'margin since 2014' graph, would it be possible to include the Undecideds? Soosider3 (talk) 05:52, 14 February 2023 (UTC)
@[[User:RERTwiki sorry to bother again, but could you explain a bit more about the charts, especially about the lead since 2014 chart. The current version uploaded by you yesterday shows almost no lead for either side, I am curious as to how this was arrived as simple average of last 10 polls shows a 2% difference, I assume you are using some weighting and would appreciate knowing what they are, many thanks Soosider3 (talk) 06:40, 17 February 2023 (UTC)
All of the averages in charts are weighted by sample size. The date for each individual poll is the mid-date of the polling interval. The date for the moving average is the weighted average date of the individual poll dates.
The Undecideds are shown in two of the charts, so I think the margin ex-undecided is a useful addition. It is also the only thing roughly comparable to the referendum, since you can't vote 'don't know'. RERTwiki (talk) 11:36, 14 March 2023 (UTC)
Graphs misbehaving for me
Hi. Having done the usual graph update, two of the charts are failing to display, at least for me.
I've put the margin chart back as default, since that does display.
I'm puzzled by this: I'll look for some tech support, but might take a while - especially if it's my PC at fault! RERTwiki (talk) 11:25, 14 March 2023 (UTC)
Works now, no idea what was up. I'll leave the margin chart as default for now, rather than change it back. If anybody feels differently, fine by me, just do it. RERTwiki (talk) 11:42, 14 March 2023 (UTC)
Coding of last poll entry
Hi -
Before I edited it to correspond to previous usage, the latest entry in the main table as edited in by KeyRing666, contained some wiki coding for the data sort value and date range using the "opdrts" function. I can't insert the exact code here because the editor is interpreting it, but it displays as you would want: |data-sort-value="2023-03-13"|9–13 Mar 2023
But sadly it generates text which contains an unusual character (non-breaking space between Mar and 2) which Excel barfs at, and so needs manual intervention to make the charts work.
Is it possible to avoid using this format? Thanks, RERTwiki (talk) 12:17, 14 March 2023 (UTC)
Latest poll is questionable
The FindoutnowUK poll for March 2023 is questionable as the individual blogger whom commissioned it is on record as saying they’ll deliberately pick pollsters and questions that favour yea Newmate12 (talk) 00:23, 11 March 2023 (UTC)
Yep. Remove it. Completely unreliable poll from a wildly biased, amateur online outlet. Ridiculous that it's still sitting in tht table as of now. 88.110.117.122 (talk) 19:40, 11 March 2023 (UTC)
Poll is by reliable source, member of BPC and complies with its rules etc. To start to remove polls because you suspect the commissioner may have an agenda would lead to a massive reworking of this whole article, think 'these island' or 'Scotland in Union' 'Belive in Scotlans'. We either trust Polling companies to act in a professional and unbiased way or we don't, me I trust polling companies not to risk there reputations on a single commission remember that for most of them political polling is a very small part of what they do. Soosider3 (talk) 12:00, 13 March 2023 (UTC)
I agree with Soosider3 that currently the poll should not be removed.
However, I do think it is one to watch very closely and may need to be removed (at least to another section) in future. Unlike the 'The Islands'/'Scotland in Union' polls which ask a completely different question (so are in a different section on the article) AND/OR use a mainstream pollster (Survation/YouGov) Pick Media Ltd trading as Findoutnow are NOT a mainstream pollster. While they are a member of the BPC, they were set up as a digital consultancy, then started running a postcode lottery Pick Media Ltd trading as Pick My Postcode, https://pickmypostcode.com, and now have started doing polling. While nothing about this is directly concerning, this lack of experience when combined with the very high Yes response in their last two polls (when compared to other pollsters), does suggest their methodology is perhaps not a robust as other pollsters. I would certainly be interested to see professional analysis of their polling, but until it happens there is no reason to remove them. Dunk the Lunk (talk) 13:14, 13 March 2023 (UTC)
Taking a look through the table, FindOutNow seems to consistently have polls showing fairly inflated Yes votes in comparison to all the polls conducted around the same time. This is a consistent theme throughout the table. I think you really have to call into question the reliability of a separatist blog that repeatedly keeps churning out polls that are quite at odds with all mainstream, impartial polls being conducted around the same timeframe. 88.110.121.113 (talk) 23:59, 15 March 2023 (UTC)
I think we need to be very careful about getting into 'judging' a poll and classifying them as "impartial" or not, lets be clear this is what we need to do. We would need to develop a whole set of criteria for that, we struggle at times to find agreement on much more straightforward matters. It is perhaps no more than an example of how different polling companies tend to produce slightly different results, a well known phenomenon. Soosider3 (talk) 05:36, 16 March 2023 (UTC)
Okay. So you'd have no problem with someone adding a poll from some amateur British nationalist blog that kept showing the Yes vote a good 5-10% lower than it generally was in mainstream, impartial polls? You don't think that would be a little odd? You wouldn't question that data and wonder whether it should be included in the article?
They're not producing different results though. FindOutNow (a separatist online blog) is producing different results consistently and repeatedly which are at odds with the results from every other poll conducted around the same time (which all find very, very similar and consistent Yes/No ratios.
You can try and bend the word impartial all you like, but I don't think you can ever bend it enough to fit FindOutNow as 'impartial'.
I think you may be conflating the commissioner with the actual Polling company, they are 2 separate things, you may not like either the blogger or the result but neither is a reason for removal. As previously stated some pollers seem to consistently (by a small margin) fall one way or another, this is not a sign of bias or over due influence from the commissioner, rather it is a consequence of each using slightly different methodology. I find it most helpful to only really compare polls from the same company with each other to try to see any trend. With Findout now they have done 3 polls in this area since Dec 2022, showing a fairly consistent return Yes 50/52/50 No 46/44/43 Undecided 4/3/6 Soosider3 (talk) 08:45, 18 March 2023 (UTC)
Okay. I can see you're completely beyond reason. I'll stop wasting my time here. 88.110.121.113 (talk) 14:37, 18 March 2023 (UTC)
Request for Comment: Inclusion of Lord Ashcroft polls
Request for comment on whether opinion polls conducted by Lord Ashcroft Polls, JL Partner and other agencies which are not members of the British Polling Councilshould be included in the article with a note specifying that they are not BPC polls, or should not be included in the article. AlloDoon (talk) 00:10, 14 February 2023 (UTC)
Should be included in the article with a note. Lord Ashcroft Polls are included in Professor John Curtice's polling aggregator of Scottish independence polls. John Curtice is the head of the British Polling Council. Lord Ashcroft Polls, for all intents and purposes, follow the same methodological weightings of British Polling Council polls. I don't see an adequate reason for their exclusion. AlloDoon (talk) 00:13, 14 February 2023 (UTC)
The consensus reached in this article is that only polls from members of BPC should be included, it says so right at the start of the article, second sentence "Only polling companies that are members of the British Polling Council and therefore fully disclose their findings and methodology, are shown in this section." Like most articles anomalies can creep in over the years I believe it is wise to remove the anomalies rather than use them as a precedent. KISS keep it simple and straightforward. Soosider3 (talk) 05:38, 14 February 2023 (UTC)
This is a break with precedent. Lord Ashcroft polls have been included for years, with a note to the effect that he is not a BPC member. Again, I believe his polls are listed at Curtice's website. It's hard to see this move to remove those polls as a reaction to the poor showing of the Yes vote in this poll. RERTwiki (talk) 11:02, 15 February 2023 (UTC)
And needless to say, because it is a break with precedent, the Ashcroft and JL partners polls should remain in the list until the conversation here has reached a conclusion. RERTwiki (talk) 11:12, 15 February 2023 (UTC)
Really depends on whether you see there previous inclusion as a precedence or an historical anomaly. In trawling through the whole article, initially to add links to data, I came across many anomalies and have worked to correct these into a more coherent and consistent article, I think with some success. I am surprised and disappointed you would make such an unseemly last comment Soosider3 (talk) 11:13, 15 February 2023 (UTC)
Yes, discussing motivation likely won't help. However, we do constantly engage in discussions on 'enhancements' when polls show the Yes vote starting with a 3. But we are all working hard on the page, and we need to listen to eachother.
I don't see how we explain to someone why they have seen a poll splattered all over the press, they can go to the polls website and download comprehensive data, and yet we have decided that they ought not to see it here. Oh, and polls by the same organisation were fit to see a few months ago. What will they conclude?
Just because we can dream up an excuse to hide some data doesn't make it the right thing to do. We can provide relevant info, like BPC status, and let the readers decide. Just because someone has incorrectly said in the article that all the polls in the table are from BPC members doesn't give us licence to change history to exclude them.
Lastly, I looked up the BPC website, and JL Partners are now members. RERTwiki (talk) 11:37, 15 February 2023 (UTC)
the Note on JL partner is more nuanced than that, "not a member at time of poll"
The rest of your comment feels very much as if you are projecting rather than rationally discussing, some of your phraseology is eye catching if perhaps irrelevant. Because a mistake may have been made in the past should not bind us to containing to repeat that same mistake, the wise thing to do would be recognise it as such and follow through with what ever corrective measures are required. I really don't accept your undertones about partisanship motivations, that is really very unhelpful especially as I am sure you could maintain a discussion without lapsing into that behaviour. Every change I have made or proposed has been based solely on trying to improve the article. Soosider3 (talk) 12:33, 15 February 2023 (UTC)
My position is the same as in previous discussions. Selectively removing items from the tables leaves the page wide open to accusations of bias. We should be seeking near unanimity that a poll is spurious or suspect, especially from well-known polling organisations, before we exclude it. Editing history to remove polls from organisations which have previously been admitted is doubly wrong.
The purpose here is tell people what polls have been conducted on Scottish Independence, and their results, not to follow arcane rules which permit the editors to censor public discourse at will.
The polls which have been subjected to this treatment are: Ashcroft, Yes 37; Techne, Yes 39%; Yougov/TheseIslands Yes 39%. Earlier Ashcroft, Yes 44% and Ashcroft, Yes 46% didn't raise a peep. Hanbury Strategy, not now BPC members, with Yes 50% and Yes 52%, raised not a peep. JL Partners with yes 51% raised not a peep, when they were not BPC members. What is going on barely needs to be articulated, which is a good thing because that would be considered impolite, or possibly disappointing. RERTwiki (talk) 08:55, 16 February 2023 (UTC)
should not be included in the article The consensus reached in this article is that only polls from members of BPC should be included, it says so right at the start of the article, second sentence "Only polling companies that are members of the British Polling Council and therefore fully disclose their findings and methodology, are shown in this section." Like most articles anomalies can creep in over the years I believe it is wise to remove the anomalies rather than use them as a precedent. KISS keep it simple and straightforward. Soosider3 (talk) 10:24, 14 February 2023 (UTC)
Should be included in the article. As per other pages on UK polling, with the appropriate caveat. The poll was also widely reported (including in papers both against and for independence). I’d also be happy if there were separate lists for BPC members/non-members. Dunk the Lunk (talk) 07:14, 14 February 2023 (UTC)
Previous discussion on content of page had rejected the idea of what other pages do as not being relevant, to that I would add that where the poll is reported is likewise not relevant. Idea of a section for non BPC might have some credence but we would need to work out some ground rules as I would be concerned that we could open the door for some rather dodgy polling. Soosider3 (talk) 10:30, 14 February 2023 (UTC)
I don't accept that what other pages do is irrelevant. What the rest of Wikipedia does is highly relevant to whether we are following established practice, or away with the fairies.
I personally think that a section for non-BPC polls is a bridge too far. The page is complicated enough as it is. RERTwiki (talk) 11:10, 15 February 2023 (UTC)
You may be in danger of conflating 2 different matters, of course we comply to Wikipedia however that is not the same as following other articles standards for content.
It would appear as if the discussion is following the usual pattern with the usual contributors and that the consensus is to include them. I would think we need to help readers understand what they are actually reading in this article.
Therefore, somewhat tongue in cheek
"We only record polls from BPC members - except when we don't"
" WE usually only record polls that follow BPC standards RE sample size- except when we don't"
"We only record polls that ask the "independence" question - except when we don't (Leave/remain)"
Lastly the page is far from too complicated, in fact it is much more straightforward now than when I began contributing last July, the content is better focused, the layout is clearer and easier to navigate. Soosider3 (talk) 11:31, 15 February 2023 (UTC)
"We only record polls from BPC members - except when we don't"
Yes, someone made an error making a flat statement that only BPC polls are in the table. It's now proposed that that error requires us to exclude some polls with perfectly adequate data disclosure, or polls which have been in the table for years.
" WE usually only record polls that follow BPC standards RE sample size- except when we don't"
We never made any statement at all about 'BPC standards re sample size', it became an issue when it was used as an excuse to exclude some polls.
"We only record polls that ask the "independence" question - except when we don't (Leave/remain)"
We have never said 'Only Independence question': leave/remain were in the main table until it became obvious they were significantly different from the others. A separate table for them is a good solution. RERTwiki (talk) 11:47, 15 February 2023 (UTC)
We are now just circling, the difference being as I said earlier whether precedence or anomaly, interesting that you see it as alright to correct what you see as an 'error' but would deny doing so with other errors. Soosider3 (talk) 12:24, 15 February 2023 (UTC)
If we write text in the article which is not consistent with the data we present, it is an error we can easily correct by fixing the text. We don't fix the inconsistency it by deleting data from the table to make the text correct. (!)
We need a compelling reason to not publish a poll, to remove data from the tables, or to 'de-admit' a polling organisation.
Otherwise, I'm pleased to see you don't contest my recollection above. RERTwiki (talk) 09:07, 16 February 2023 (UTC)
Please don't assume that a short reply is that I don't think your position is contestable, it most definitely is. However it would appear as I stated earlier that the usual few people are the only ones who comment, although I have only been active on this page since last July that much is very very obvious. The BPC standard has clearly been very evident, in the main table, as shown by the number of polls from BPC members and the size of the sample. The text reflects the reality of the article, I really do find it surprising that this obvious and very clear position escapes you.
As I said we are circling, so little point continuing. Soosider3 (talk) 10:25, 16 February 2023 (UTC)
This has been discussed previously, in archive 1. This is Professor John Curtice's view on Lord Ashcroft's polling: "As it happens, Lord Ashcroft Polls is not a member of the BPC (and as an organisation that does not do work for multiple clients is not eligible to be a member), but as it happens it publishes full details of its polls in much the same way as a BPC member would be expected to do." In other words, Ashcroft behaves in the same way as a BPC member, and the only reason he isn't a member is based on a technicality: all of the work he does is paid for out of his own pocket, rather than being paid for by clients. It should be included. Jmorrison230582 (talk) 10:05, 17 February 2023 (UTC)
Saw that quote the other day which is why I had backed away from my position, still think im right but if big John says then thats good enough for me Soosider3 (talk) 10:11, 17 February 2023 (UTC)
Had noticed your text re 'missing' polls at least Whatscotland thinks has them as polls re Independence, I have had a look at as far as I can, following link from WST gives us general Panelbase page, for both dates there are 2/3 polls conducted none of which ask the Independence question. Very odd but suspect somewhere either Panelbase have not posted them, or even more unlikely WST is wrong Soosider3 (talk) 10:46, 17 February 2023 (UTC)
Graph - can we have an explanation?
The graph provide a very useful visual start to the article, however I have some questions as I am unclear as to how the data is actually used to produce the graphs and hope we could have a brief explanation ( perhaps as a footnote)
Are all polls listed used in producing the poll?
I understand that we use a 10 poll average and assume this is to smooth out the graph and minimise the impact of outliers, however I think we might also be using some form of weighting re date and size of poll. If so can we have an explanation of what weighting is being used and why that one?
I make these requests on the basis of ensuring transparency and that all sources are clear and concise.
In trying to duplicate the graphs, in using Excel and datawrapper I have found it impossible to reproduce even a close proximity to them, I have checked them against other aggregators such as Whatscotlandthinks and Ballotboxscotland and again find a significant difference between their graphs and ours. Soosider3 (talk) 08:20, 3 April 2023 (UTC)
To me it looks like all this is in the relevant file descriptions on Wikimedia? Could you not use those? Otherwise I’d message the author directly. Certainly the What Scotland Thinks graph is just a dot to dot so won’t look anything like these graphs which a weighted. Dunk the Lunk (talk) 11:36, 4 April 2023 (UTC)
In no way is this a criticism of graph or author rather trying to understand what is happening, had a look at metadata which is what actually got me thinking "The margin and date are weighted by sample size. " is a direct quote, what does that mean? how much weighting? etc etc.
I'l take your suggestion and message author Soosider3 (talk) 19:02, 4 April 2023 (UTC)
As is noted under the headings, polls are those from the main table, excluding those which carry a note to say that they use a non-standard question.
For each poll, I take the range of survey dates, abd compute its mid-point, ie (start+end)/2. That is the date of the poll, which is the x-axis for the data points for individual polls, for all charts which have them.
For the moving averages, I take the last 10 polls which are in the main table and ask standard questions. I weight the polls by their sample size, and use that to create weighted averages of yes, no, and undecided. That gives the y-axis measure for each series. To get the x-axis measure for the moving averages, I take the weighted average of the poll dates, as described above.
(Other polling pages tend to use Loess smoothing. I prefer to avoid that because it leads to average lines which are not stable, in that a new poll can ‘change history’ and move the earlier average around. The centred moving average method, as above, is fixed forever once it eventually appears.)
To chart the margin ex-undecideds, I use the same averaging scheme and x-value. The y-value averaged is (no-yes)/(no+yes).
The final chart is a frequency distribution. The y-value is the percentage of a given organisations polls which lie in the x-axis bucket. I compute the x-value for a poll by linearly interpolating the moving average margin ex-undecideds to the date of the poll, as described above, and take the x-value as poll margin ex-undecideds minus interpolated average ex-undecideds.
When written out (as requested) it looks complicated, but actually each step is simple and pretty compelling.
That if course the above is the intention: mistakes can happen. I have to exclude the non-standard polls by hand, once or twice I’ve got that wrong.
If there is any cause to believe something is wrong, happy to discuss. 213.31.88.127 (talk) 22:17, 6 April 2023 (UTC)
Many thanks for taking the time to give such a full and helpful reply, you certainly have improved my understanding of not just how the graphs are done but just as importantly why they are done in that particular way.
I would suggest that your reply could form the basis for a footnote and would welcome other folks views on this Soosider3 (talk) 11:20, 7 April 2023 (UTC)
How about a button in switcher to access the methodology? RERTwiki (talk) 11:36, 7 April 2023 (UTC)
Now your just showing off ;-)
Seriously that would be great, anything to enhance transparency and understanding Soosider3 (talk) 11:45, 7 April 2023 (UTC)
I was just thinking of having a 'chart' in the Excel file which is a page of explanatory text which I can upload (thankfully only once!). I've just updated the charts, I'll look at that next time. RERTwiki (talk) 18:09, 7 April 2023 (UTC)
Actually, playing around with it, I can put editable text into switcher, so I should be able to make the button text look quite natural on the page. RERTwiki (talk) 19:19, 11 April 2023 (UTC)
That's done now, no need to wait for the next poll. RERTwiki (talk) 09:47, 12 April 2023 (UTC)
Murrell News
I couldn't help but notice the recent revert-sport going on about the inclusion of a line on the news of the Murrell arrest.
I thought it might be good if we could have a reasoned conversation about it here instead.
The exact definition of what does and does not warrant a mention in the table has not really been articulated, to my recollection at least.
It seems to me that what makes sense to include is reference to any event which may tend to explain apparent changes in the meat of the article, which is the polls in the main table. Which is really any big news relevant to Scottish politics.
I can't think of much if any bigger news than the Murrell arrest. It certainly has the potential to affect polling.
My 2 cents is that it should be included. Could we possibly try and see what the consensus is by replying here?
I think the 'standard of proof' isn't as high as required to exclude a poll, since there would otherwise be endless trivia included. I think we should simply include or exclude this news according to majority feeling. RERTwiki (talk) 18:22, 7 April 2023 (UTC)
Personally and for the sake of avoiding argument I would probably not have anything mentioned in the table apart from the appointment/resignation of party leaders/first ministers, and elections.
With regards to the Murrell news it certainly seems politically more relevant that the war in Ukraine or COVID19 arriving in Scotland, but the tone did need rewording to make neutral. Dunk the Lunk (talk) 11:39, 8 April 2023 (UTC)
I have waited a few days before putting my tuppence worth in.
Looking at other wiki articles re polling across the UK, they do appear to take a very simple view, changes in political leadership/elections but rarely anything else. I believe that not only creates a cleaner look but avoids an area where there could be lots of scope for wrangling about significance or not of an issue. There are so many things that could be taken as 'political' and therefore worthy of mention that perhaps it is best to avoid that potential quagmire. lets keep it simple and straightforward 81.96.97.29 (talk) 12:03, 11 April 2023 (UTC)
Oops I forget to sign in before posting previous comment Soosider3 (talk) 12:06, 11 April 2023 (UTC)
I also do not understand why Murrell's arrest, which is highly relevant to Scottish politics and has received extensive coverage in Scottish media, has been excluded from the table while the war in Ukraine has been included. AlloDoon (talk) 16:47, 11 April 2023 (UTC)
Soosider3 makes a fair point about avoiding what are essentially judgement calls, but in this case I'm afraid the call is pretty simple. The news has barely been off the front pages for days. Even some SNP officials (sorry don't remember who) have said that this will set back independence for some time. Hard to see how it isn't relevant to polling.
While I appreciate this isn't what some people would like to see, it is very significant news and ought to be included. RERTwiki (talk) 19:18, 11 April 2023 (UTC)
Interesting discussion, which from what I can see appears to be fairly evenly split.
Is the Murrell story a big story - Yes.
Has it had an impact on Polling - who knows, we would need to predict the future to know that
Is this article about Polling on Scottish Independence or has it become something else? where we speculate about what might or might not impact on Polling.
The inclusion of 'News', stories is a fairly recent phenomenon, prior to Covid only one incidence back in 2017 (about s30 Application)
From what I can see of other wiki articles on polling in UK - they do not include 'News' stories.
Perhaps we should use this discussion to focus on what we want this article to be and think hard about that bearing in mind that once you set the precedence you will find it hard to set the bar, and that's before I get into issues such as Reliable Sources Wikipeadia has good guidance on what it sees as Reliable sources and Social Media and newspapers (especially red tops) are seen as being unreliable. if we are to include we would need to find reliable sources for our inclusions.
To be clear, I am against the inclusion of 'News' stories in this article and would go further and suggest for consistence we remove the few that are there such as War in Ukraine, Covid 19, and possibly even Supreme Court and S35
As I said earlier my preference would be for everything to be removed apart from elections (local, Scottish, UK, Scottish and UK referendums, but not by-elections), appointments/resignations of party leaders and FMs, and head of state changes (ie Queen dying and Charles becoming King). Everything else starts to become subjective, for example there is mention of Westminster using S35 to block legislation but no reference to the Scotland Act 2016 which transferred a range of powers including full control of income tax to the Scottish Parliament. Dunk the Lunk (talk) 05:29, 12 April 2023 (UTC)
I think the point about red-tops is spurious: this story has been everywhere.
I would argue for inclusion of items which a substantial body of editors opinions suggest merits inclusion. We are not talking about 'cat stuck up tree' stories. This is one of the most dramatic and shocking political events I can remember, and I remember Harold Wilson as PM.
FWIW I also think the biggest pandemic in 100 years and the first invasion of a European country in many decades merit inclusion.
Both sides in this debate have got to respect the other. We can't go on with endlessly debating the censorship of reasonable inclusions on the (unspoken) basis that they are embarrassing for one side, using arcane technical excuses. Often this comes with downgrading the page by having to remove things previously considered valuable on the basis of 'consistency'.
If only a brute-force revert war can create some balance, then so be it, but it would be better avoided. RERTwiki (talk) 10:10, 12 April 2023 (UTC)
Yes I agree with that RERTwiki.
I also think Soosider needs to appreciate WP:Ownership guidelines and stop relentlessly edit-warring.
I think we need to step back a bit, what had been a very constructive conversation has lapsed into merely personal attacks and allusions about intent, this achieves nothing.
Reading the comments it is clear we have an impasse, with 2 for inclusion of these sort of incidents and 2 against it.
Lets take some cooling off time. Soosider3 (talk) 15:52, 12 April 2023 (UTC)
Dunkthelunk said: With regards to the Murrell news it certainly seems politically more relevant that the war in Ukraine or COVID19 arriving in Scotland, but the tone did need rewording to make neutral.
This has been amended, and therefore a majority want to include the news, not to mention the original editor who added it in the first place.
It would greatly assist discussion if you could desist from misquoting contributors or quoting out of context Dunk the Lunk has twice stated his preference, very clearly and very concisely.
"Personally and for the sake of avoiding argument I would probably not have anything mentioned in the table apart from the appointment/resignation of party leaders/first ministers, and elections.
With regards to the Murrell news it certainly seems politically more relevant that the war in Ukraine or COVID19 arriving in Scotland, but the tone did need rewording to make neutral." and" As I said earlier my preference would be for everything to be removed apart from elections (local, Scottish, UK, Scottish and UK referendums, but not by-elections), appointments/resignations of party leaders and FMs, and head of state changes (ie Queen dying and Charles becoming King). Everything else starts to become subjective, for example there is mention of Westminster using S35 to block legislation but no reference to the Scotland Act 2016 which transferred a range of powers including full control of income tax to the Scottish Parliament"
It really could not be clearer, for you to interpret this as supporting your position is, well, bizarre. Until or if Dunkthelunk amends his position or clarifies it then please respect his view. Soosider3 (talk) 19:25, 12 April 2023 (UTC)
My position is unchanged. Given how this discussion is going it rather makes my point about only including clear cut events (which I previous noted), rather than the wider events included at present which are ambiguous and open to edit wars etc. Dunk the Lunk (talk) 16:18, 13 April 2023 (UTC)
Yes to cooling off, I have to consider my Blood Pressure (capitalised).
2-2 is more than a little unhelpful. Ideas which occur: seek some arbitration, which we've done before or possibly craft some more anodyne statement which might be viewed as a compromise. The first has more prospect of success I think.
Possibly "Extensive coverage of Police probe into SNP financial affairs"? Not really feeling the love though.
Away to cool off until the next poll shows up. RERTwiki (talk) 08:01, 14 April 2023 (UTC)
And finally: apropos more polls - if polling shows a notable change after this news, it argues for including the news as explanation. RERTwiki (talk) 08:25, 14 April 2023 (UTC)
Thats going down the interpretation rabbit hole though (and the polls have been trending down for Yes since December so would be difficult to directly ascribe to this event unless it’s a massive drop). I thought most folk had agreed that the article/table is here to present data to users and not to spoon feed them. Given that these types of items are not reported on other similar pages I really don’t think they should be here (not just the Murrell news but all the other events). Dunk the Lunk (talk) 07:56, 15 April 2023 (UTC)
No Notes Note
I've responded to this by removing generalities (which might need sourcing) from the intro, since most of the rest of the intro was free-standing, as far as I could see.
Any suggestions for places anywhere where an in-line citation might be helpful? RERTwiki (talk) 08:15, 14 April 2023 (UTC)
Not sure I completely understand what the notice was about. An in line citation, appears to be a link that is treated as a reference, and generates an entry in teh reference field. Would changing the link to data tables from a hyperlink to a hyperlink within a reference, its not as neat but would provide an answer to this suggestion. It would be a fair amount of work but is doable as links are already there, I can change a few over if folk would like to see it in operation Soosider3 (talk) 10:37, 16 April 2023 (UTC)
Just had a thought, some of the pollsters provide link to data tables but also to an analysis piece carried out by them, we could add the latter as an inline citation. This should answer the query asked of us while at same time giving readers wider access to reliable sources Soosider3 (talk) 10:43, 16 April 2023 (UTC)
As a taster, I have added inline citations for the last 2 Redfield Polls. My thinking would be that we add these citations where the Pollster has produced a commentary such as these (thereby avoiding less reliable sources)
Would welcome comments Soosider3 (talk) 10:58, 16 April 2023 (UTC)
I feel having the direct link to the tables (with the external link arrow) and numbered citation directly next to each other is possibly a little confusing for the reader. However (at the moment) I can’t think of a better way of doing it! Dunk the Lunk (talk) 17:12, 16 April 2023 (UTC)
Aye, its far from an elegant solution and as you say its not user friendly (having to hover over No. and click in link in pop up box)
Not the easiest documents to follow but my take from them is
We appear to be only UK article of this type with this Template Message, rather suspect it is a one size fits all message
1 The difference between the links we use to tables and inline citation appears to be no more than a bit of coding Cite error: There are <ref> tags on this page without content in them (see the help page). around hyperlink. The impact of using this coding in this type of article is to actually reduce the clarity and usability of the article
2 I think it maybe possible for us to remove the Template if we believe it does not apply to our article, we need to evidence this by consensus in the Talk page
Personally I do not think this requirement sits well with these sort of Articles and to implement it would diminish the article , that the links we use are actually better than what is suggested.
Therefore I propose that we seek that consensus on here, we currently have 3 contributors to this talk, could I ask that we indicate whether Agree or Disagree with proposal that This template should not apply to this article and should be removed Soosider3 (talk) 08:42, 17 April 2023 (UTC)
Agree, though Allodoon is a 4th, if that matters RERTwiki (talk) 16:56, 22 April 2023 (UTC)
Disagree the formatting should be changed to be references instead of inline external links. As per WP:ELMcMatter(talk)/(contrib) 17:29, 22 April 2023 (UTC)
What about WP:ELLIST? The second example table includes official website links in the table. Dunk the Lunk (talk) 17:41, 22 April 2023 (UTC)
None of these are the official links as this topic is not about an organization or person. The links in the table are where you are citing information from therefore should be citations not EL. Even then if you look at the examples the Official link should only be in the external link section not as inline EL. McMatter(talk)/(contrib) 17:43, 22 April 2023 (UTC)
The second table has them as a separate and clear indication that clicking on that will lead to an outside page, by listing them as webpage, not hidden as they currently are. McMatter(talk)/(contrib) 17:47, 22 April 2023 (UTC)
I am grateful that the person who placed the Template re EL has joined the conversation.
We could get into a long exchange of interpretation about what Guidance does or does not mean, but at essence we should be guided by first principles in producing an article that is useful, accessible, helpful to any reader, factual and provides reliable information. It is my belief that this article in its current form achieves all of that and does so in a manner that is quick and easy for a reader to engage with, any one with a modicum of sense will immediately be able comprehend what the article is about and glean a sense of understanding of what it is attempting to do. There are several Articles on polling in the UK that very successfully follow a similar format, in doing so have become a useful resource for many. Furthermore I am of the view that to follow your suggestion would seriously diminish the article, its usefulness would be compromised its accessibility and navigation ruined.
I would point to the very start of the page you referred to WP:EL there are 2 boxs right at the head, that state that EL are permissible,(in certain circumstances) can be useful and most importantly that 'Common Sense' be used in the application of this guidance.
Lastly I would note that those that have so far expressed a view on the topic have done so in the positive, not a single article editor has spoken against the proposal, these are editors that are amongst the most active in maintaining the article. In respect to them I ask you exercise common sense and withdraw your template. Soosider3 (talk) 08:23, 23 April 2023 (UTC)
The policy is in place because the links within an article should only link to other articles on Wikipedia. This makes a standard across all articles and what readers expect when reading articles on Wikipedia. The only allowance for use of external links is when it is very clear that it is an external link. The common sense approach here is to bring this page into alignment with every other article on Wikipedia and in compliance for ease of the reader. I don’t care about what the primary maintainers prefer, this is an encyclopedia for the reader. McMatter(talk)/(contrib) 13:23, 23 April 2023 (UTC)
These are guidelines , not policies and as such are there for the guidance of the wise not as a rigid uncompromising set of rules.
I would remind you of the purpose of Wikipedias Purpose and its "just-the-facts style"
Look forward to the discussion raised Soosider3 (talk) 14:46, 23 April 2023 (UTC)
I think I now understand what's being asked, and I think it's absurd. This article is exactly in-line with all Wikipedia polling pages I use. That's because this is a sensible and highly usable way to structure these tables. We are not out of line here, it is this change request which is out of line. At the moment, it seems like all of the active editors agree.
Could you give an example of an alternative format, and explain in what way it improves the article for the reader? RERTwiki (talk) 16:29, 26 April 2023 (UTC)
If you haven't already you should pop over to the discussion that was opened,Wikipedia talk:WikiProject Politics#Opinion Polls lists it appears the matter is far from as clear cut as originally stated. Think we are heading to a position where if not removed by administrator I will remove it myself as being irrational in terms of access and readability. 212.84.177.108 (talk) 03:32, 27 April 2023 (UTC)
Ooops, must remember and log in before commenting Soosider3 (talk) 03:33, 27 April 2023 (UTC)
Redfield & Wilton Strategies -
Redfield and Wilton are now producing a monthly tracking poll, with a full spread of 'Scottish' polling, I have linked to their Data Tables but notice that as well producing this they also do an analysis/commentary on their poll, this seems to produce good information, graphics and graphs and am thinking this may be something that would be useful or of interest to our readers. However am not sure how we would present that information in our article and also have some slight reservations about how we might maintain a high standard for inclusion of this sort of information.
If other editors think its worthy of inclusion there are a few ways we might record this
1 Add it as a Note - not easiest or clearest method but has advantage of being quick and easy
2 Introduce a new column by separating Pollster and client, then making link against client ( at present R&W are both).
3 Introduce a section for Trackers with this sort of commentary
This is to my mind really good information presented in a clear and useful manner, would really like to include it but worry about boundary for what is or is not included. Soosider3 (talk) 08:35, 4 May 2023 (UTC)
I agree that it’s a good presentation, and quite thorough in reviewing key questions in Scottish politics.
However, other than as entries in our various tables, it’s quite hard to see how to include it here. Maybe there is scope for other wiki pages.
Another aspect is that it is somewhat invidious to single out one polling organisation and highlight their material. It would be like we were advertising them.
Lastly, we know that individual firms exhibit clear biasses, probably caused by methodology for the most part. Again that makes singling out one quite risky.
So: agreed it’s good stuff, not keen on doing much more with it until we gave a really good scheme worked out.
On the plus side, I see no time constraints, so we can chew it over for however long it takes. RERTwiki (talk) 20:36, 17 May 2023 (UTC)
My apologies for not responding sooner ( I forgot to subscribe to the article)
I take your points and share some of your reservations, at present it is only pollster producing information of this calibre so yes would be highlighting there practice but that may not be a bad thing. My greater concern is that once we open this channel we have to look at gate keeping to maintain the standard, I for one don't have the stamina for what that might become. Agree lets let it lay, there is no hurry. Soosider3 (talk) 19:06, 24 May 2023 (UTC)
YouGov MRP April 10 - May 21
Could we have a discussion about this work from Yougov, had been hunting for data tables today but discovered that it is not actually a poll but rather an MRP, yougov describe it themselves as a "New Model" sample size 3,500 but fieldwork ran from April 10 to May 21
Key questions for me is does it have a place in this article? and if so where? Soosider3 (talk) 19:00, 24 May 2023 (UTC)
Pretty clearly nothing to do with opinion polling on Scottish Independence? Let's just stick to the page mission. People are welcome to start other pages if they have the energy/stamina. RERTwiki (talk) 17:46, 25 May 2023 (UTC)
Colouring of Timing table
I'm minded to colour correctly the cells in the timing table. The only shading is green (for YES) even though many of the entries ought to be red for NO, including some of the entries currently coloured green. I guess it will be a fair amount of work.
Hopefully this isn't controversial. Will be a week or so before I get round to it, I'll be away for a while. RERTwiki (talk) 17:51, 25 May 2023 (UTC)
I had actually considered that when amending the table to what I thought was a better layout but could not get my head around it. Eventually only did 'lead' for clear return between referendum and no referendum as I was not sure how useful it would be to have either individual leads per (timescales) or to amalgamate lead across different timescales, at one time had considered removing lead column.
Coming back to it after sometime I am not sure that the existing lead colour serves any real purpose and in fact may be creating a false impression, A bit concerned that lots of coloured boxes may just create a more confusing picture than a cleaner layout.
Perhaps best if we just go simple on it and remove existing coloured boxes and possibly even the lead column. 212.84.177.108 (talk) 03:25, 26 May 2023 (UTC)
We can pilot some schemes here in talk. RERTwiki (talk) 08:43, 26 May 2023 (UTC)
Meanwhile I've fixed what seem to me to be the apparent errors in the existing colouring. RERTwiki (talk) 08:56, 26 May 2023 (UTC)
On reflection, not clear if the data justifies an entry in the lead column for those polls. Green is quite definitely wrong, but it might be more logical to just remove the entry in the lead column for those polls. I'll leave that to others. RERTwiki (talk) 09:03, 26 May 2023 (UTC)
Had a we stab at using highlight colours, it really wasn't as difficult as I first feared so can be easily reverted if required.
However it did through up some queries that I am not sure how to progress with
For polls where more than one timeframe then it might make sense to highlight the lead in each timeframe, not sure we could add totals together to get collective leader (not sure that would make sense)
For polls asking only one timeframe, or only one response in more than one timeframe then lead column entry could work but it would be clunky.
There might also be an issue that data collection and questions asked have evolved over time, initially only Panelbase using a particular model, but they haven't polled on this for over a year. More recent polling does seem to be more likely to use a couple of timeframes.
Take point re colour vision defects, but thought of revamping whole article is daunting. Soosider3 (talk) 14:44, 26 May 2023 (UTC)
On a more general point I’m not sure if the Red/Green colours meet wiki guidelines more generally? On mobile so can’t check easily at the mo but red/green are a very poor choice for colour blind users, plus I guess it could be argued that there is an element of bias in the colours themselves (green: positive, red: negative?). Dunk the Lunk (talk) 14:22, 26 May 2023 (UTC)
Been having a look at this and found a very useful tool online that allows you to take a screen grab and then simulate how that might look to readers with various types of Defects in their Colour vision
As someone who has such a defect this is of particular interest to me, about a third of males have some defect in colour vision but tends to be fairly rare in females, Red/Green is fairly rare as is complete loss of colour vision
The key seems to be in the hue of colour used and having a secondary indicator (shading or cell edging)
Tried above app and the colours/shades used in bulk of article seem to stand up fairly well however bolder/prime colours used in lead column appear to be more problematic especially the red. Soosider3 (talk) 08:28, 27 May 2023 (UTC)
Polling Org discrepancies
(first, sorry to be a little slow with the graphs. Have been super busy.)
I'm just moved to rant about the discrepancies between polling organisations results.
If you want to know whether Scots would vote for independence, the simplest thing is to ask what the polling organisation is. If it's FindOutNow or IPSOS, the answer is yes. If it's Redfield or Yougov the answer is very likely no. The discrepancies in the 'de-facto referendum' section are laughable.
IPSOS have 15 polls ALL below an average WHICH INCLUDES THEIR OWN POLL. That's 15 straight 'tails' if it was a fair toss-up. Utterly implausible: the coin isn't fair.
These polls are very clearly, a statistical slam-dunk, not measuring the same thing.
IMHO the BPC has not done a good job in writing rules that oblige polling organisations to conduct polls in a way which produces consistent results in a highly charged atmosphere like the independence debate. At the very least, the differences should be readily explicable, beyond mumbling 'methodology'. I might write to the BPC and complain, with a view to getting this fixed in the long term.
Or not. Interested in any views on the matter. RERTwiki (talk) 10:59, 21 June 2023 (UTC)
No worries about updating the graphs-you do a great job.
I do know where you’re coming from regarding the differences between polling companies. Personally I’m extremely dubious about the Find Out Now polls (and noticing a trend that this company seems to be increasingly used/exploited by explicitly Nationalist organisations/funders so potentially introducing a very real bias in rolling averages) - I would be very interested to see the graphs without them ;). The Ipsos polls are stranger one as they are a well known reliable company but their results seem to be totally out of step with the other mainstream polling companies.
However, all we can do here though is to include the polls that are to the agreed specification. It’s up to others beyond this article to interpret them! Dunk the Lunk (talk) 17:26, 23 June 2023 (UTC)
Largely agree with you about polling company seeming to have a methodology bias. However it does seem to be well understood, I rather think that this perhaps brought into focus by polling being so relatively close for such a long time, that every little shift is focused on so intently. Secondly I suspect that the comparatively low number of polls on this topic further sharpens focus on even small variations.
The tab you did in Graphical Summary highlights this very clearly, although Ipsos is 5% one way, Survation and Panelbase on the money, Savanta 1% the other way and Yougov 2%, however when you factor in the number of polls carried out then perhaps if there is a bias it favours the No. To be clear I am not looking for an argument, rather I am hopefully demonstrating that there are many things outwith our control or influence.
I have found findoutnow polls to be eyebrow raising, there latest seems to be broadly consistent with there previous efforts, although not weighing for probability to vote and excluding 16/17 year olds makes for a 7% jump. Looking at the 2 tables the client published on their website I find myself scratching my head. Samples of 5000+ become representative sample sizes of just over 1000 and are different in both tables as are the results for Yes/No.
When I first saw there latest poll I did not post as the numbers 'jarred' I tweeted them and got response I recorded in Notes field, perhaps BPC needs to encourage an expectation that likelihood to vote is a desirable component (irrespective of client preferences) that may have a cost implication to client but it does seem fairly central to polling VI
There is a system of using Notes to highlight significant changes from established norms, perhaps we just need to use these Soosider3 (talk) 10:42, 26 June 2023 (UTC)
Importantly, check out the direct response from Find Out Now at the end (date 25th Jan 2023). However, 5000>1000 does suggest something atypical is going on with their methodology. The comparison with the results of 2021 Scottish parliament election is also interesting. Dunk the Lunk (talk) 16:14, 26 June 2023 (UTC)
Thanks for that, it certainly answers some of my original queries however has now replaced them with others, they don't weigh but sample a very large number to get sufficient across the different groupings. Is this compliant with BPC code? Is it directly comparable with other more usual polling methods? does it make any difference to capturing data on Wikipedia as we can demonstrate that their is methodology bias from most pollsters, is Findoutnow just another example of this. I'm away for my tea Soosider3 (talk) 16:33, 26 June 2023 (UTC)
Yeah it looks odd and as I said earlier I’m pretty sceptical about their results… But as far as I’m concerned they are BPC members and AIUI their polls meet BPC guidelines. So unless we hear differently in terms of this article should be included. As to whether they are directly comparable that’s a different issue but beyond the terms of reference for this article as I think we’ve agreed to avoid to much interpretation here!! The sooner they get listed on the Polling Org Vs Mean the better though so everyone can see (Sorry RERTwiki!!). Dunk the Lunk (talk) 17:24, 26 June 2023 (UTC)
I've been told by a person who has used FON that the reason they have a large sample which yields a smaller representative sample is that they start with a sample which is roughly demographically representative, then discard responses at random so that the final sample is correctly representative. RERTwiki (talk) 16:31, 27 June 2023 (UTC)
Another peculiarity with FON is that it is a gambling (postcode lottery) website. Well not quite gambling, since the stake is zero: they sell ads and polling services on the site to fund prizes. So, on the face of it they are selecting for a specific psychological profile.
More troubling to me, the panel is self-selecting, with registered users about 2M in the UK, guessing ~180K in Scotland. Further, having tried this myself, I had no difficulty creating multiple accounts with multiple email addresses. There might be checks I'm not aware of, but again, on the face of it, I can volunteer almost arbitrarily often to be on the panel.
I just don't think this is fit for purpose for a highly contentious issue like independence.
However, I have to agree they are a BPC member, and their polls should be listed here. The only proper avenue is to persuade the BPC to do something to correct the current farcical situation where different organisations flatly contradict eachother, and organisations cherry-pick their pollsters to get the result they want.
Having said which, we might consider whether this has gone far enough that separating polls by methodology might be more enlightening to the readers than slapping together these contradictory polls in charts and tables. RERTwiki (talk) 16:44, 27 June 2023 (UTC)
While broadly recognising that different pollsters have methodology bias, generally that is very small, we stray into very dangerous grounds when we start to assert motivations, would be do so for other clients such as Scotsman or Times - No, nor should we, but the argument could be made.
Sorry but introducing some mechanism to separate by methodology is a none starter, given the comparatively low numbers of polls it makes even less sense.
The article works well as it is and we should allow readers to exercise their own judgement, after all we already highlight significant issues re sample size etc and highlight lead etc. Soosider3 (talk) 05:40, 30 June 2023 (UTC)
