Wikipedia:No original research/Noticeboard

Welcome to the no original research noticeboard
This page is for requesting input on possible original research. Ask for advice here regarding material that might be original research or original synthesis. Include links to the relevant article(s). Make an attempt to familiarize yourself with the no original research policy before reporting issues here. You can also post here if you are unsure whether the content is considered original research. Sections older than 28 days archived by MiszaBot II. (For help, see Wikipedia:Purge) If you mention specific editors, please notify them. You may use {{subst:NORN-notice}} to do so.
Additional notes: "Original research" includes unpublished facts, arguments, speculation, and ideas; and any unpublished analysis or synthesis of published material that serves to advance a position. Such content is prohibited on Wikipedia. For volunteers wishing to mark a discussion resolved, use {{Resolved|Your reason here ~~~~}} at the top of the section.
To start a new request, enter a name (section header) for your request below:

SpaceX Starship costs

I wanted to add this section to the article on SpaceX Starship: In a update on their website in late October 2025, SpaceX stated that Starship is "SpaceX is self-funding representing over 90% of system costs" [sic], which combined with known public contracts (like the one for Starship HLS ) representing the remaining 10 or less percent would put the cost of the "core Starship system and supporting infrastructure" at the moment of the release at no less than $27 billion. But I'm not sure if that would be synthesis, or merely arithmetically combining two different sources (we know from one that it is 90% self-funded, and we know from the other what the remaining 10% translates to at the minimum). Thoughts? Hal Nordmann (talk) 18:43, 31 October 2025 (UTC)

It's OR because you don't know on what basis each of the sources computed their numbers. One number is from 2021 and one is from 2025, and you don't rule out that NASA or other agencies made additional contracts. There is a lot of uncertainty involved and your estimate could be quite silly and therefore IMO leans towards not legit even as a clearly stated estimate. 2A02:2455:8423:4800:4A09:B9E2:5AE8:60AE (talk) 19:53, 31 October 2025 (UTC)

The intent was to estabilish a lower bound for the costs (and clearly state so), as the SpaceX estimate is too vague for anything more. But if you say so... Hal Nordmann (talk) 21:01, 31 October 2025 (UTC)

You don't even know if the 90% claimed by SpaceX was not a bullshit PR claim. It's not in any kind of official document. Uncertainty upon uncertainty. I would not be comfortable at all relying on such an estimate.

I am not strict about OR if the wording makes clear that it is, but in this case it is a stretch. 2A02:2455:8423:4800:4A09:B9E2:5AE8:60AE (talk) 00:46, 1 November 2025 (UTC)

My gut feeling is that this goes beyond simple arithmetic because we aren't really sure if these figures are apples to apples or apples to oranges. Horse Eye's Back (talk) 01:27, 1 November 2025 (UTC)

References

Pedro I

"James I & VI" is a name used for the person who was both King James I of England and King James VI of Scotland. It is easy to find reliable sources with the name. The man known as Pedro I of Brazil was also known as Pedro IV of Portugal. Editors at his article wish to add the name "Pedro I & IV" to the infobox. I have asked for sources for the name on the talk page but none have been provided (in any language). I believe it is contrary to Wikipedia:No original research, Wikipedia:Manual of Style/Biography#Names and Wikipedia:Manual of Style/Lead section#Alternative names to use a name that is not found in reliable sources. DrKay (talk) 09:01, 2 November 2025 (UTC)

Highway distance

I have used the directions tool of Google Maps to find the distance along several highways in California. Is this original research? If not, do I need to cite Google Maps? — Preceding unsigned comment added by Latiasprismstar (talk • contribs) 21:31, 4 November 2025 (UTC)

It's OR. If no reliable source cares to mention these distances, why should our articles? EEng 00:15, 5 November 2025 (UTC)

Evie Magazine

Regarding the article Evie Magazine There are sources that describe it as conservative NYT Guardian

There are sources that say it is alt-right futurism opendemocracy

CNN makes the point that some call it alt-right:

"This was Evie Magazine, a publication and website founded in 2019 by married couple Brittany and Gabriel Hugoboom. The magazine has characterized itself as a “conservative Cosmo.” Some, pointing to its record of publishing conspiracy theories, vaccine misinformation and tradwife nostalgia, have characterized it as “alt right.” (Evie Magazine did not respond to requests for comment.)"

My opinion is that, in this case, we should write: "women's magazine described as conservative by some and alt-right by others", but other users insist on writing "conservative alt-right women's magazine". I think this is a case of SYNTH. Moreover, by definition, alt-right is an extremist movement that often disagrees with both left and conservative. I don't know what scholar would write "conservative alt-right".

https://en.wikipedia.org/w/index.php?title=Evie_Magazine&action=history — Preceding unsigned comment added by Deamonpen (talk • contribs) 18:34, 7 November 2025 (UTC)

I'll copy what I just wrote at Talk:Evie Magazine: The issue, or at least one of them, is that neither The Guardian nor The New York Times sources call Evie "conservative" in their own voice. The Guardian article avoids descriptors, but the overall point is that Evie is anti-feminist. The New York Times style guide specifically says to avoid using terms that might be seen as pejorative or extreme, so they tend not to use phrases like "alt-right". Instead, The New York Times article says that Evie publishes articles that promote positions that are fringe even within conservative circles. That is the NYT's way of saying that Evie is more than conservative. Woodroar (talk) 18:38, 7 November 2025 (UTC)

Reposted Guardian article to the right one that calls it conservative.

"Momfluencers and tradwives celebrate RFK Jr’s “Make America Healthy Again” policies while wearing breezy milkmaid dresses. Evie Magazine, a politically conservative version of Cosmo, appropriates the trending visuals of feminist magazines with headlines that decry body positivity and promote vaccine skepticism. As the New York Magazine writer Brock Colyar described young Republicans at a post-election night party: “Many are hot enough to be extras in the upcoming American Psycho remake."

The NYT says that there are articles that are fringe even for right:

"But readers who click past “hot girl” health trends and Adam Brody appreciation posts will find articles that promote positions that are fringe even within conservative circles."

But it recognizes a general trend towards conservative:

"Many of Evie’s writers have been affiliated with conservative institutions, and the website regularly publishes content that reflects today’s conservative positions, including opposition to abortion, transgender rights and vaccines, as well as support for the Trump family. People have labeled Evie “far-right,” which the Hugobooms find irritating; they repeatedly called it a “double standard,” arguing that outlets like Teen Vogue and Refinery29 aren’t always described as explicitly left. The couple, both of whom voted for President Donald J. Trump, said they felt that the way conservatives were portrayed in mainstream media was outdated."

The way it says "People have labeled Evie 'far-right'" suggest that the author of the NYT piece distances themselves from this position.

Another NYT piece

"She said she has used the apps Flo and 28, the last of which was founded by the creators of the conservative Evie Magazine and backed by the right-wing kingmaker Peter Thiel. Both are part of a fast-growing, multibillion-dollar market for women’s health technology."

NYT, again

"The Victoria’s Secret Fashion Show was on Wednesday. Some Christians were sharing an article online that had called for a return to “a hyper-feminine showcase of genetic beauty,” published in the conservative Evie Magazine."

I have changed the sources cited in the article to the two Guardian and NYT pieces that make no mention of Evie being anything other than conservative.

---Deamonpen (talk) 19:02, 7 November 2025 (UTC)

Background sections used for original research

Does Wikipedia permit any exemptions for its Background sections when it comes to sources and content having to be "directly related to the topic of the article" part of WP:OR? My recent edits to three articles to flag sources that do not refer to the article titles, read broadly, have been undone either on the basis of the content or sources being "factual" or because an editor did not believe the statements in question had to be directly related to the title subject if the information was "relevant and uncontroversial". Human to human, I understand editors' urge to want to provide more context, but is there merit to the points from the perspective of Wikipedia?

I'll pick two examples from the edits in question. In Grokipedia, the Background section talks about efforts to create an alternative to Wikipedia, with Conservapedia being given as an example. The cited source is from 2007, predating Grokipedia by many years. In Epic Games v. Google, the Background section says that the Epic Games' CEO has been outspoken in criticizing the 30% fee of digital storefronts. The cited article mentions stores like Steam and GOG, predates the title lawsuit by three years, and makes no mention of Google. Daisy Blue (talk) 17:50, 11 November 2025 (UTC)

I think that's an interesting question. I tend to avoid adding "background" that no sources on that article topic have mentioned as relevant. Once background information is established as relevant by sources on that article topic, it's reasonable to provide brief context for the reader which might entail the use of a source not on that article topic. I think original research comes into play when an editor decides for themselves that information is related even though sources on that article topic haven't made that connection. I don't know how the community as a whole views it though. Schazjmd (talk) 18:06, 11 November 2025 (UTC)

An article about a battle in the Second World War could contain a background section on the state of the war at that point, even if the sources used in that section didn't directly mention the battle specifically. Doing so wouldn't be OR. -- LCU ActivelyDisinterested «@» °∆t° 21:08, 12 November 2025 (UTC)

What guidelines would you reference to make that case though? A question for everyone, even if I'm replying to just one person. Daisy Blue (talk) 21:53, 12 November 2025 (UTC)

It's a good question. I personally try to always find a source linking any background information to the article subject. To my knowledge, this isn't codified as a guideline anywhere. As an aside, similar questions could be asked of "See also" sections, which tend to become a coat rack of barely related topics. Anne drew (talk · contribs) 22:13, 12 November 2025 (UTC)

It is original research, and I'd argue it's also POV editing. These sections are a result of editors adding coverage that's not given weight in the source. For the Grokipedia case, that's an editor coming to their own conclusion about Conservapedia and Grokipedia being comparable, based on their own POV. If something isn't covered in sources about the subject, then it doesn't belong in articles about the subject. Thebiguglyalien (talk) 🛸 04:09, 13 November 2025 (UTC)

Wikipedia articles are meant to be comprehensive so that they can be used and understood while offline. Background sections are generally needed to help with that approach, particularly around events, so the reader does understand how events did not just magically crystalize from nothing. We do want to make sure we don't go so far back or so far away from the topic to be excessive (like the common joke "Start at the beginning..." "Well, first the Earth was made..."). This type of reasoning is equivalent to the same processes we use to summary sources for a topic, the type of required "original research" to make the decisions as to the most relevant points, and thus appropriate. Further particularly for topic that involves a sequence of events that are separated by some time, you will get coverage of one step of the event which often will point back to their previous coverage of the earlier event, but rarely would cover the entire sequence in full. Obviously, it would be great if there were sources that did that, but as long as we reasonable can follow the line of logic from sources on the event sequence, and again applying common sense to what is appropriate, these background sections are fine. Masem (t) 13:32, 13 November 2025 (UTC)

The key is that any “line of logic” has to come from a source that directly presents that line and links A to B to C. Using our own logic is Original Research. The source does not have to be directly “about” the topic of the article… but it does need to explicitly state the line of logic we present. Blueboar (talk) 13:52, 13 November 2025 (UTC)

I don't believe that, when the line of logic is deductive, such as with the example I gave below (where a source says that X is a member of group Y and another source says that all members of group Y have property Z), we need to have the logic 'X has property Z' explicitly spelled out in a source used in the article X. But I also acknowledge that this type of sourcing is often used for more defeasible lines of logic, such as a source which states that X is a member of group Y, and another which states that 'many' members of group Y have property Z, while not explicitly mentioning X as one which does. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 14:53, 13 November 2025 (UTC)

In the case of Epic Games v. Google (which I am a primary contributor of), there is a ear series of events starting from Sweeney's comments on revenue share to the onset of the lawsuit. I know I have the sourcing that clearly links each set of consecutive events in thus chain (newer articles on a new event linking back to at least the previous event), just not sourcing that is specifically about the lawsuit/trial that doesn't dig back that far. I would definetly be more concerned if I brought up some event in the best that has no clear sourcing linkage, as that becomes a true OR problem, just like with the deductive logic issues above. Masem (t) 15:26, 13 November 2025 (UTC)

I have no doubt that sourcing exists. However, I don't think the article's current sourcing needs improvement, at least not based on this question. I outlined my logic below, but to summarize: there is no claim in the article, implicit or explicit, which isn't sourced in the article. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 15:32, 13 November 2025 (UTC)

• This is a situation I've encountered before. It actually pops up with some regularity on WP:FRINGE topics, where I edit frequently, and where it's usually done to push a pro-fringe POV.

Despite that, my thoughts are that it's not as cut-and-dry as some people believe. I've had conversations about the question of "does every source need to be on the topic of the article or on the topic of the claim?" and most editors seem to agree with the latter. That is my interpretation, as well. If the information is relevant, verifiable and WP:DUE, then I don't see the problem. Indeed, the best argument against doing this is arguing that it's undue, as the sources aren't about the topic. That can be a very valid complaint.

Despite that, the specific type of WP:OR looks to be WP:SYNTH. And calling these examples synthesis seems seems a bit of a stretch. After all, synthesis is when two or more claims are combined to produce a third claim that isn't verifiable in either source. The first claim in either case seems to be that the subject of the article exists. That's reliably sourced. The second claim, which is also reliably sourced, are the ones you gave about. So what is the third claim, in either of these cases?

In the first case, the third claim is that 'both Gockipedia and Conservapedia are WP clones that push a right-wing narrative'. Sources extant in both articles are explicit that each is a WP clone that pushes a right-wing narrative.

In the second case, the third claim is that the 30% fee is pivotal to the lawsuit. Well, that claim is actually reliably sourced elsewhere in the article.

In neither case can I find a claim that is not supported by reliable sources. In both cases, it appears to be just a summary of the sources, and summarizing sources is explicitly within our remit. Indeed, it's our very purpose here.

So here's the tl;dr. When encountering issues like this, we need to carefully consider what WP:SYNTH means, not just what it says, or what we remember it saying the last time we checked that page. Comparing that section and the essay WP:SYNTHNOT is a good idea. My personal advice is always to carefully think through what WP is explicitly saying, and try to separate that from what we're reading into it, as well. There are some fundamental tenets of logic which are inescapable, after all. If all X are members of group Y, and all members of group Y have property Z, then I don't think using a source that says "all members of group Y have property Z" is original research in an article about X. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 14:46, 13 November 2025 (UTC)

Explaining her 2005 edit that introduced the "directly related" part to WP:OR, SlimVirgin said:

today we have a new user trying to add an explanation of what the term "fair comment" means in an article saying that a court in New York judged that to call Lyndon LaRouche an anti-Semite was fair comment. The new editor is trying to argue that what fair comment means in law is not what it means in everyday language, and in so doing (whether he's right or wrong), he's trying to build a case, which is not allowed. I'm going to try to come up with a paragraph explaining why this is original research.

Like the policy wording itself, that challenges the idea that it's our job as editors to provide context or background explanations via sources that are not "directly related to the topic of the article". From looking at the prior discussions about Background sections, I also found an essay by Phenylalanine, who takes the opposite position, similar to Masem's and that of a number of other editors. The essay cites WP:AUDIENCE (previously at WP:PCR) to make the point, but that guideline does not support the idea, as it explains "providing context" merely as not using jargon, pointing out simple facts like the Ford Thunderbird being a car, or wikilinking. There is no hint at sources being within that scope.

Nevertheless, given that we do have these discussions and debates, I think it calls for even more clarity in WP:OR, like adding a footnote to say that a directly related source cannot be used as a proxy for using unrelated or indirectly related sources with respect to the article topic. Daisy Blue (talk) 22:32, 13 November 2025 (UTC)

A term like "fair comment" in the context of a legal case is a far cry from "jargon", there are exact terms of art that have legal history and meaning, and if we don't already have a blue link to explain the term, a law person explanation provided by an unrelated but reliable source is absolutely in line. It would be OR to try to interpret the term in another manner without the use of sourcing. On the other hand if we were talking cars and we had "the driver souped up his car", "soup" is clearly jargon that can be replaced without an explanatory source. Masem (t) 23:31, 13 November 2025 (UTC)

See my reply to Blueboar for how "directly related" doesn't mean the same thing as "explicitly mentions". ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 00:48, 14 November 2025 (UTC)

I believe it does, as further underscored in WP:NOR/EX, which provides an example similar to SlimVirgin's, containing what looks like helpful and valuable context classified as original research simply because the source does not mention the subject of the article:

Intelligence officers of the attacking forces claimed to have telephoned the residents of buildings where military assets were suspected of being stored, to warn them to leave before it was bombed.[Sourced to an article about the bombing] During the war, 90% of the telephone system was down.[Sourced to an article that does not mention the bombing] (Implied: Thus the attacking forces could not have warned 90% of the residents whose buildings were bombed.) Daisy Blue (talk) 11:17, 14 November 2025 (UTC)

That's a SNYTH logic problem that was discussed above, as it is making a direct and not-necessarily-true inference from a disparate source. On the other hand, explaining what a legal term if the sources on the lawsuit where the term was used with a disparate source is not trying to create a novel conclusion but to help inform the reader. It is the same as if we blue linked the legal term to an article about it (eg we frequently blue link the term en banc without explanation when it comes up). As long as we do not try to twist that definition to interpret the decision in the case, absent any other source that actually does that. Basically, stuff like this falls under Wikipedia:Make technical articles understandable-type of providing necessary context to the reader as needed. Masem (t) 13:55, 14 November 2025 (UTC)

That guideline is similar to WP:AUDIENCE in that it deals with jargon, wording, summarizing the body for the lede, and formatting, but does not hint at allowing sources that aren't "directly related to the topic of the article". It tells us how to present information rather than how to source it. If anything, it too reminds us to avoid original research, whether by synthesis or not. Daisy Blue (talk) 18:09, 14 November 2025 (UTC)

I think the relevant part of the quote you provided above re; the 'fair comment' bit is ...and in so doing (whether he's right or wrong), he's trying to build a case, which is not allowed. (emphasis added)

The 'case' that user is trying to build is the claim that's not supported by the sources.

Masem was responding to that and your argument which referenced it. I don't read their comment as implying that the guideline he linked gives remit for us to include background information. Both of us have addressed that already with different arguments. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 21:05, 14 November 2025 (UTC)

The problem with that example is that a source which shows that 90% of the telephone system was down does not show that 90% of the telephone lines in that building were down. That building could have had complete telephone service at the time of the bombing, by dint of being part of the 10% that did have service. (It's worth noting that most apartment buildings have a single telephone connection point, so if one telephone in the building worked, then it's highly likely that all or at least most of them worked.)

It's introducing a third claim which is not supported by either source. Again, in both of the examples you give, that 'third claim' is, in fact, supported by sources, just not the same sources as are used to support the out-of-context meaning of each claim.

I would say that it's a poorly-written example. In fact, I'm fairly certain I have previously said so about that exact example. I don't think that implied claims are as cut-and-dry, because the implications aren't necessarily the same for each person, though I will acknowledge that that example provides a particularly bad instance of an implied claim. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 14:15, 14 November 2025 (UTC)

I am requesting wikipedia community guidance on a possible violation of No Original Research policy

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

hello. I am involved in a content dispute in which another editor has used personal speculation and personal statistical analysis to advance his position. I am requesting guidance on a possible case of No Original Research in a content dispute. First, I should say that the text is somewhat long in order to fully explain the problem. I apologize for that.

The issue is related to a sensitive data regarding the percentage of Afghanistan Ethnic Percentage Table , and we have discussed this for a long time and made some progress. But now the discussion has reached an stalemate with another editor because he refusing to accept the violation of NOR.

I should also say that all the information in Afghanistan Ethnic Percentage Table is base on the entire territory of Afghanistan and includes all 34 provinces of Afghanistan.

but what is the problem? To understand the problem, I invite you to look at pages 39 and 40 of this ABC News survey.Please look at this PDF. https://abcnews.go.com/images/PollingUnit/1083a1Afghanistan2009.pdf On pages 39 and 40, you will see the names of the 34 provinces of Afghanistan in which the survey was conducted. And the numbers in front of the provinces indicate the number of data collection centers and information. The year and time of the survey are also written in the top row of these statistics. For example, the years 2004. 2005. 2006. 2007 and 2009. Everything is written very clearly and understandable.

pages 39-40 explicitly shows incomplete and flawed geographical coverage for certain years:

Year 2004 = Information is available for 28 provinces. (Two provinces were not yet formed at that time)

Year 2005 = Information is available for 31 provinces.

Year 2006 = Information is available for 31 provinces.

Year 2007 = Information is available for all 34 provinces.

Year 2009 = Information is available for all 34 provinces.

As we can see,The source clearly states the fact of incomplete coverage. the data for 2004, 2005, 2006 are not from the entire territory of Afghanistan. The Original Research Violation: (another editor SdHb insists on keeping this geographically incomplete data in the main national table. to support this position, he have introduced his own novel statistical analysis, arguing that:

( these provinces that are not in the statistics together account for 2.8% of the Afghan population and it is not important and this is smaller than the sampling error (Margin of Error) of ±3.5%." Therefore, this incomplete and defect data still provides a reliable national representative )

The editor's argument confuses two distinct statistical concepts:

Sampling Error (Margin of Error)'= Uncertainty measures the size of a given sample. For example, in Year 2005 sample, it measured the uncertainty of 31 provinces.

Coverage Error: Occurs when parts of the population are systematically excluded from the sampling frame.

The Margin of error only measures the uncertainty due to sample size.

Margin of Error is fundamentally different from Coverage Error.

An example to help you understand: Suppose you want to measure the average temperature of the “whole” of a lake.

The “margin of error” says: “I measured 100 random points on the lake, so my measurement Uncertainty is ±1 degree.”

The “coverage error” says: But I completely omit all points to the north of the lake from the measurement.

the result is You have measured the average temperature of a part of the lake, not the whole lake, because the points to the north of the lake are not covered. So the claim that the data are for the whole lake is completely disproved.

The truth is that SdHb have no right or permission to ignore or downplay the lack of statistics for three different provinces. - The editor uses his false personal speculation and personal statistical analysis to advance his position. This is a case of NOR that is prohibited in wikipedia.

Question for the Community= Please tell me whether this problem violates No Original Research rules or not? thank you very much. Badakhshan ziba (talk) 00:42, 14 November 2025 (UTC)

This doesn't seem like OR on SdHb's part. If a poll claims to be tracking national statistics, it's not original research to say that. Without knowing the poll's methodology, it's impossible to say how they accounted for those missing provinces. As a compromise, possible limitations with the data could be mentioned in a footnote. Anne drew (talk · contribs) 02:00, 14 November 2025 (UTC)

Hello everybody, I'm deeply disappointed that we couldn't solve this without escalating to WP:NORN. That said, I want to clarify that my edits don't violate said policy. At no point have I introduced unpublished analysis or synthesis. I have relied strictly on what the ABC News survey reports state explicitly in their published methodology reports (https://abcnews.go.com/images/International/1026MethodologyNote.pdf, https://abcnews.go.com/images/Politics/998MethodologyNote.pdf), which is that these polls were designed and executed as nationally representative opinion surveys based on stratified random sampling proportional to population size. The sources themselves also explain that sampling points were allocated according to population distribution and that margins of sampling error already account for design effects and clustering.

Badakhshan ziba argues that because some provinces had no sampling points in 2004, 2005, and 2006 because of security and accessability problems, the results are illegitimate on a national level. This is a misinterpretation of what national representativeness means in survey methodology. Coverage of all first-level administrative units (in this case provinces) isn't a methodological prerequisite but proportionally distributed sampling is. This doesn't invalidate the national estimate unless the survey organisation itself says so. And ABC News doesn't ever state that the data for these years are invalid or non-national. On the contrary, they explicitly present the statistics as national findings ([1], [2]):

An ABC News poll in Afghanistan -- the first national survey there sponsored by a news organization ...

... ABC News has sponsored five national public opinion polls in Afghanistan ...

The claim that including these survey years would be "original research" by me is ridiculous. I didn't interpret raw data or calculate new statistics. I merely summarised what the survey documents already state, which is that the surveys reflect national opinion and that their uncertainty is reported through a margin of sampling error of ±3.5%. Anne drew said:

As a compromise, possible limitations with the data could be mentioned in a footnote.

That's exactly what I proposed, thank you for confirming that this is an acceptable compromise. Including these years in a chronological table, accompanied by a transparent footnote identifying the unsampled provinces and noting that they together represent maximum 2.8% of the population, just reflects verifiable, published data from an otherwise mutually agreed upon reliable source. No interpretation or synthesis beyond what the ABC methodology itself states has been introduced by me. For even more details you can follow the discussion on the talk page. Thank you. SdHb (talk) 09:27, 14 November 2025 (UTC)

@Anne drew Please look at the above post by @SdHb. This is in my opinion a clear violation of Wikipedia's rules on WP:NOR.

@SdHb The word "national" that I initially mentioned is not a media headline or media claim, but rather the full coverage of the 34 provinces of Afghanistan and the entire territory of Afghanistan. @SdHb You are using an article in ABC News site that claims that this was a "national poll". This is a media claim.

The problem is not about defining the word "national". The main problem here is that the poll data (ABC NEWS/BBC/ARD POLL) in 2004, 2005 and 2006 does not cover the entire territory of Afghanistan.

Incidentally, ABC itself has also fully confirmed that in those three years, the entire territory of Afghanistan was not covered by the survey.

The information mentioned in the Afghanistan Ethnic Percentage Table , is all based on information from the entire territory of Afghanistan (all 34 provinces of Afghanistan) and we are not allowed to include data( that does not cover the entire territory of Afghanistan) in this table with personal interpretations and opinions.

As Wikipedia editors, we must follow the policy of ""No Original Research". This means that we must base our edits on verifiable facts from the original source, not on claims, whether personal claims or unverified media claims.

We want to use the data and statistics of this source in Wikipedia ( (ABC NEWS/BBC/ARD POLL ).

In fact, this is the main source cited. So what does the main source tell us? It is clearly stated on pages 39 and 40 of the main source that the survey in the three years 2004, 2005 and 2006 does not have complete coverage (it does not cover all 34 provinces of Afghanistan and the entire territory of Afghanistan)

Note = I did not see anywhere in the main source that claimed that a survey was conducted in entire territory of Afghanistan in 2004, 2005, and 2006. If any person find anything about this, please show us. Badakhshan ziba (talk) 22:24, 14 November 2025 (UTC)

While the PDF explicitly identifies this as a national public opinion poll, I completely understand your objection: how can a survey be "national" if it excludes entire provinces? The reality is that polls often rely on statistical methods to correct for incomplete samples.

Regardless, it isn't our job to peer-review the methodology of reliable sources. If the source claims the data is national, we should describe it as such. Excluding data based on our personal analysis of their sampling is simply editorial overreach. Anne drew (talk · contribs) 22:59, 14 November 2025 (UTC)

@Anne drew I do not intend to completely delete this survey.

The statistics of Afghanistan Ethnic Percentage Table are based on the entire territory of Afghanistan and 34 provinces.

I have been discussing this table with @SdHb for three months.i think He is not impartial and neutral.

possibly He has been trying to find sources in these three months to increase the percentage of Pashtun statistics. (I will register a request in Wikipedia:Neutral point of view soon in this regard) Now, if we want to put incomplete and incorrect data of 2004,2005,2006 in the table, the number of Pashtuns will be artificially increased. And in my opinion this is the possible reason why @SdHb insists on including these incomplete statistics in the table.

Therefore, we cannot include the data of 2004, 2005, and 2006 in this table.Instead, the data of 2007 and 2009 can be included in this table. Instead, the data of 2007 and 2009 can be included in this table. This way everything will be perfectly fine.

Let's leave aside the discussion about the definition of the word national. My main objection is the lack of complete coverage of all 34 provinces of Afghanistan. Not the discussion about the word national. Please someone address my objection. Why should we include information that does not include 3-4 provinces of Afghanistan in the Ethnic Percentage Table of Afghanistan؟ Badakhshan ziba (talk) 23:37, 14 November 2025 (UTC)

@SdHb Again, you have made the wrong interpretations and personal opinions.

Let's compare what the source'says with what the editor @SdHb adds.

1. What the authoritative source (ABC NEWS/BBC/ARD POLL) states as fact:

Fact A: For 2004, 2005, and 2006, the poll was conducted in 28, 31, and 31 states, respectively, not all 34 states.(Source, pp. 39-40). Fact B: The poll has a margin of sampling error of ±3.5%.

The source reports these as two separate facts. The two item are completely unrelated with each other.The source never links this two facts together.It never states that fact B (the margin of error) makes fact A (incomplete coverage) statistically un important.

2. What the editor (@SdHb) adds is Original research:

The editor makes a Wikipedia:Synthesis. that is not found in the source:he calculate that the omitted provinces constitute 2.8% of the population. then he correlate and blend this 2.8% with a margin of error of 3.5%.then he draw a new conclusion and say: Therefore, the data are "reliable" and "representative of the entire territory of Afghanistan" despite the incomplete coverage.

The editor is not reporting the findings in the source; he is creating a new, derived conclusion to advance his own personal position.

3. The editor's entire argument is based on a fundamental statistical error.

He is confusing two different concepts: Sampling Error (Margin of Error): and Coverage Error:

I have already explained the difference between the two in detail this text , but (@SdHb) still insists on his wrong position and apparently does not understand the difference between the two.we cannot use "Margin of Error" to justify "Coverage Error".

They are different types of problems.

The editor @SdHb is doing original research to justify his mistakes. thank you. @Anne drew Badakhshan ziba (talk) 23:14, 14 November 2025 (UTC)

Okay, no problem.@Anne drew We have nothing to do with defining and interpreting the word national.

Let's talk about this text . This text clearly demonstrates @SdHb faulty reasoning and analysis. Do you agree that this is original research? if No why? Badakhshan ziba (talk) 23:42, 14 November 2025 (UTC)

No, I don't agree. As Xan747 correctly pointed out, WP:OR does not apply to talk pages. This isn't an original research issue - it's at best a WP:RS issue (if you're saying the source is questionable) or a WP:NPOV issue (if you think the source's content is being given undue weight). I also would caution you from questioning the motives of other editors; to assume good faith is a Wikipedia guideline, and continually violating it could result in sanctions. I'm not going to participate in this discussion any further - I've said my piece and I'm not a fan of repeating myself. Anne drew (talk · contribs) 16:22, 15 November 2025 (UTC)

@SdHb (and Badakhshan_ziba, Anne drew, and anyone else I might have missed):

Based on the different descriptions of the issue, I don't think this qualifies as original research. However, it does sound like an extrapolation, if you're essentially trying to apply a different data set to another set that has gaps. This could arguably be considered a personal opinion (possibly WP:NPOV) since most extrapolation relies a bit on assumptions that are sometimes up for debate. Why not just acknowledge the gaps in the data, explaining the (legitimate) reason for the missing data, and let it stand as-is?

Separately, I don't think the gaps in the data should be considered a "sampling error". There was no error made. It was a known "limitation" on collecting data from the excluded regions. It wasn't included in the original sampled population to begin with, so the probability was zero (0) that it would end up being represented in the final data.

If it seems I've misunderstood the reasons for data limitations/gaps and the difference of opinion on how to handle the data set, please let me know.

(EDIT: Sorry for all the edits I made to my reply. It's a long discussion and I intended to address it to the involved editors directly but made some typos/omissions. I probably still left someone out anyway.) BetsyRogers (talk) 22:48, 17 November 2025 (UTC)

@BetsyRogers thank you for your answer. maybe i have raised my issue in the wrong place.

@SdHb is insisting on something wrong that is not true at all. The truth is that in 2004, 2005 and 2006, no polls were conducted in 3-4 provinces of Afghanistan and this is quite clear. please see p 39-40

https://abcnews.go.com/images/PollingUnit/1083a1Afghanistan2009.pdf

But unfortunately, another editor (probably with the help of artificial intelligence or chatGPT) insists on hiding this issue by making wrong arguments. thank you.Badakhshan ziba (talk) 18:44, 18 November 2025 (UTC)

@Badakhshan ziba I'm aware that there is data missing from those years, and I haven't seen SdHb suggest otherwise, at least not here. But either way, could you clarify how you think the survey data from those years should be handled? Are you suggesting that the data shouldn't be shown at all, or do you think it be shown in a different way? Or something else? (You might have already clarified this somewhere else, but I only came across this discussion yesterday and it's a lot to read, so I might have missed it.) BetsyRogers (talk) 20:01, 18 November 2025 (UTC)

I am an involved editor. I would just like to add that WP:NOR does not apply to talk pages:

This policy does not apply to talk pages and other pages which evaluate article content and sources, such as deletion discussions or policy noticeboards.

So, even if SdHb's reasoning would be considered OR if used in an article, its putative use in a talk page to evaluate a source's reliability would not be a violation. Also, these ABC polls are particularly strong relative to other sources also in use because they disclose so much detail about their methodology that others don't, giving them relatively better compliance with WP:V, allowing readers the ability to judge their reliability for themselves instead of us doing that for them as the filer would have us do—which would be a potential violation of WP:POV, and what SdHb and I are attempting to avoid by including sources the filer would have us remove. Xan747 (talk) 16:00, 14 November 2025 (UTC)

@Xan747 I don't understand what you mean. Our discussion on the talk page reached a dead end and was not resolved. So I'm stating my point here. This issue is not resolved yet.

And that I see no reason to keep content that violates Wikipedia rules in the table.

ME and no one else have not any intention to deleting the(ABC NEWS/BBC/ARD POLL). THE 2007 AND 2009 Data will remain in table. Badakhshan ziba (talk) 23:58, 14 November 2025 (UTC)

I would like to ask more of the Wikipedia community to give their opinion on this text Is it original research or not? If not, please explain why.

This is a discussion almost related to statistics and mathematics, which clearly contains incorrect analysis and speculation. Badakhshan ziba (talk) 00:21, 15 November 2025 (UTC)

No, it's not OR, because the WP text isn't claiming anything that ABC News doesn't claim. Bernanke's Crossbow (talk) 20:13, 17 November 2025 (UTC)

Can you answer a few follow-up questions to make sure I understand? (Apologies if this was already answered somewhere but this is the first time I'm reading this, as an uninvolved editor, and it's a lot of information to process.)

- Regarding the table you linked from your talk page, does this also appear in a specific article here? (If so could you add a link to that, for context?).

- When you said this is sensitive data, could you clarify what you mean by sensitive?

- In the pdf you shared showing data from ABC, for the years with gaps in the data are they just using a simple average/mean calculation (where data from each province is expressed as a portion of 100%)?

- On the table on your talk page, when a province has missing data and just has a "--" mark in it, would it be feasible to add a footnote symbol and add corresponding footnotes to the bottom of the table stating how many provinces were included that year? That way it's abundantly clear if people are only looking at the table (a lot of people skim over the text and just look at tables). For example if a cell has " -- " in it, you could add a footnote symbol like " -- ^a" (or b, c, d, etc) where "^a" would link to a footnote at the bottom of the table stating exactly how many provinces were included/omitted in that particular year. BetsyRogers (talk) 00:15, 18 November 2025 (UTC)

@BetsyRogers Thank you very much for your careful attention to this matter.

1-Yes, this data is intended for the table in the Ethnic groups in Afghanistan article, specifically in Ethnic groups in Afghanistan#:~:text=Estimated ethnic composition section. Unfortunately, @SdHb has changed the content of this article a lot.

Just compare the content of this article with January 2023 and January 2022.

2-Meaning of "Sensitive Data = I use the term “sensitive” because data on the ethnic composition of the Afghan population is a highly political and controversial issue in the country of Afghanistan. The percentages assigned to each group can have significant social and political consequences.

The Taliban terrorist group and its leaders from the Pashtun ethnic group, and of course there is a lot of evidence that various Pashtun governments in Afghanistan have always tried to drastically reduce the population of other ethnic groups or make them look small for the past 60 years. They are trying very hard to make themselves look above 50%. While most of their sources mention between 38 and 42 percent. Even a one to two percent increase or decrease in the population percentage is sensitive in Afghanistan.

Therefore, it is crucial that the data we provide is as accurate and methodologically sound as possible to avoid misrepresenting the country's demographic landscape.

3- the pdf and data from ABC = The data on the page 39 and 40 is not % percent or simple average .On pages 39 and 40, we see the names of the 34 provinces of Afghanistan in which the survey was conducted. And the numbers in front of the provinces indicate the number of data collection centers. Wherever the number 0 is placed in front of a province, it means there is no data collection center there.

4- There is dispute over the data in this table. This table shows the total percentage of different ethnic groups based on all 34 provinces of Afghanistan. Afghanistan ethnic group table

Adding a footnote, although transparent, does not change the fact that the number shown in the table is presented as representative of 34 provinces, when objectively it is derived from only 29 to 31 provinces.

Forexample It is like labeling a map as "France" but only examining 30 of the 34 regions.The presentation (title "Map of France") is still misleading.The solution is to either not present that map or to label it correctly as “Map of 30 Regions of France”.

Your answer is yes. If we enter the incomplete and defect data into the table, the data and precentage will change in a misleading way.

For example, some percent % will be artificially and incorrectly added to the Pashtun ethnicity. And again, your answer is yes. Many people do not pay attention to the footnotes at all and only look at the top and header of the table.

The "--" sign does not refer to the provinces of Afghanistan. It refers to the ethnic groups of Afghanistan. The "--" sign means that information about that ethnicity is not mentioned in the source.

I think it is correct to include only years with complete and verified geographical coverage in the table (such as 2007 and 2009) ABC for which all provinces of Afghanistan are fully covered.

In my table, the data for 2004, 2005, 2006 has been removed.

But @SdHb insists that we should present the incorrect and incomplete data of 2004, 2005, and 2006 (which lack 3-4 provinces) as representative of the ( 34 province) entire territory of Afghanistan in the table. thanks for your attention. Badakhshan ziba (talk) 20:43, 18 November 2025 (UTC)

@Anne drew The main problem we have and the main reason why this discussion is going on for over 3 months now (this isn't reduced to this particular case only) is that Badakhshan ziba still doesn't understand the difference between WP:VERIFIABILITY and WP:NOR. On WP the rule is very straightforward: if something is verifiable through WP:RS, it can be included. It's not our job to decide whether we personally believe the methodology of a source is scientifically perfect or whether we would have designed the survey differently. What matters is what the published, reliable source itself states, not what an editor think the source "should" have done.

The ABC News surveys we're discussing are professionally conducted opinion polls with publicly available methodology notes (this is mutually agreed upon). These documents (as I've shown in my comment above) explicitly describe the sampling strategy, allocation of interview points, margin of sampling error, and limitations in accessibility during some years. Most importantly, they clearly and consistently present the surveys from 2004 to 2009 as national public opinion polls. Since this comes directly from a reliable source, they are verifiable and therefore must be included per WP:V. Whether an editor personally agrees with the methodology or thinks (WP:NOR!) the coverage is incomplete is irrelevant. The only relevant question is: does the source itself present these results as national? Yes, it does. So the data belong in the article, plain and simple. An editor's personal opinion can't override the statements of a reliable source.

Badakhshan ziba keeps claiming that I'm speculating or doing original research. I'm not. None of my edits fall under WP:OR. I didn't generate new results (I quote WP:CALC here directly: "Routine calculations do not count as original research", and counting 1.2, 1.1 and 0.5% together is a routine calculation). I didn't interpret raw data. I didn't contradict the source. All I did was summarize what the ABC methodology reports already state, which is how the sampling was designed and how ABC News itself presents the surveys.

By contrast, what would be OR is insisting that certain years should be excluded because "true national coverage" supposedly requires sampling all 34 provinces. That's a new, editor-generated conclusion and not supported by any reliable source! The ABC methodology never claims that limited access to a few provinces invalidates the national estimates. If that were the case, the source would say so. But it doesn't. If ABC News calls their survey a national poll, then Wikipedia reports it as a national poll. No need for recalculating anything, for deciding whether the "northern lake region" analogy even makes sense, or for personally deciding how many provinces are "good enough". Removing these years based on personal methodological opinions would be the actual WP:SYNTH and WP:OR.

If the source acknowledges limited access (which it does), we we handle this with a footnote, as it is already consensus between Xan747, Anne drew and myself, instead of deleting data. This is fully in line with WP:NPOV with how we normally deal with methodological notes in reliable sources. Now please let's rest this case for good now. SdHb (talk) 11:04, 15 November 2025 (UTC) Edit: Just for the record: I've seen Badakhshan ziba's repeated claims of an alleged bias that I have. This violates WP:PA and WP:AGF, and this is not the first time this happens. Can anything be done against it? SdHb (talk) 11:10, 15 November 2025 (UTC)

@SdHb there is sufficient evidence that you are not neutral about the pashtun ethnic statistic.

the necessary evidence will be provided in this regard at the appropriate time at NPOV Noticeboard. Badakhshan ziba (talk) 19:22, 16 November 2025 (UTC)

We cannot use footnotes in table Afghanistan Ethnic Percentage Table

If we use incomplete and defect data, the numbers and information of ethnics in the table will change deeply.

Footnotes do not solve our problem. Basically, there should be no wrong numbers or percentages in the table. For example, is it correct to say: 2+3=6 and then write a footnote for it?

The truth is that data that does not cover the entire territory of Afghanistan ( all 34 province of Afghanistan) should not be included in the table. All the information in this table is based on the entire territory of Afghanistan. And you are destroying the validity of this table by insisting on including incomplete and defect data.

The content in the chat gpt cannot justify and ignore these major flaws. Badakhshan ziba (talk) 19:39, 16 November 2025 (UTC)

"Is it correct to say: 2+3=6 and then write a footnote for it?" Yes, if people genuinely claim that, then here at Wikipedia it is. Thus our coverage of (e.g.) Mochizuki's "proof" of the abc conjecture.

Remember, our goal is not to truthfully describe the world, but to truthfully describe what other people think about the world. That is the point of WP:NOTTRUTH. I highly recommend that you read and internalize that essay. Bernanke's Crossbow (talk) 20:11, 17 November 2025 (UTC)

=====================

Please answer my question. this question . My question is more related to statistics science and mathematics.

Badakhshan ziba (talk) 20:33, 16 November 2025 (UTC)

My answer to you is: it‘s time to WP:LETITGO. SdHb (talk) 20:40, 16 November 2025 (UTC)

Okay. WP:LETITGO So I will remove the data that does not relate to the entire territory of Afghanistan from the table. I hope there is no objection to this.

What I am doing is based on pages 39 and 40 of the main source.

https://abcnews.go.com/images/PollingUnit/1083a1Afghanistan2009.pdf

Everything we say should be based on pages 39 and 40. Because these two pages are our main source in the table.

this table

Wikipedia's policies require us to base our content on verifiable facts, not claims. You prioritize a general claim on page 1 over the verifiable, methodological data** on pages 39-40. Your position is synthesizing the page 1 claim with the pages 39-40 data to create a new conclusion: "Data from 31 provinces can represent 34 provinces !? ! ! This synthesis is your original research. Badakhshan ziba (talk) 21:18, 16 November 2025 (UTC)

@Badakhshan ziba Yes I do mind if you unilaterally remove data from reliable sources from the table. Please do not. Xan747 (talk) 15:23, 17 November 2025 (UTC)

Hello @BetsyRogers, to keep everything clear and consise (losing track in these discussions has become a fairly common theme unfortunately), I'd like to answer your questions down here.

[D]oes th[e table] also appear in a specific article here?

Yes, you can find the table in the article Ethnic groups in Afghanistan in a slightly modified way which we agreed upon here. The live version does only contain reliable sources we discussed and agreed upon beforehand, whereas the other user's talk page table contains sources we haven't even discussed yet, let alone mutually agreed. The ABC survey years are one of several sources used in that table.

When you said this is sensitive data, could you clarify what you mean by sensitive?

Sensitive in this context refers to how strongly ethnic data in Afghanistan is politically charged (this is also a great read on the topic), not to any Wikipedia policy category. That sensitivity is exactly why sticking closely to what reliable sources themselves state is crucial.

In the pdf you shared showing data from ABC, for the years with gaps in the data are they just using a simple average/mean calculation (where data from each province is expressed as a portion of 100%)?

I'm not 100% sure if I understood that question, but I'll try to answer. The ABC methodology notes (e. g. [3], [4], [5]) explain that all years, the ones with accessibility limitations included, were conducted as national opinion surveys using stratified random sampling proportional to population size. Sampling points were allocated to provinces according to population, and then further distributed randomly to districts, villages, and households. They didn't calculate "province means" in the sense that each province accounts for 1/n and the missing provinces were calculated as "0*1/n". Instead, the survey results are based on the provinces that were actually surveyed, with sampling points weighted according to population. Since this is how the source itself presents the data, by WP:V we have to reflect that without re-evaluating the sampling strategy.

Why not just acknowledge the gaps in the data, explaining the (legitimate) reason for the missing data, and let it stand as-is?

That's exactly what I've done a couple of days ago. I fully support the use of footnotes. This has already been proposed by multiple editors (you, Anne drew, Xan747, Bernanke's Crossbow, myself). This way we avoid deleting reliably sourced content, and WP:NPOV is satisfied, because we present all verifiable data plus the methodological limitations that the source itself acknowledges. SdHb (talk) 09:26, 18 November 2025 (UTC)

I don't think Wikipedia editors need to parse out the distinctions between sampling errors and coverage errors or to, in general, understand statistical inference at a high level in order to determine whether this oft-cited in the above statement from ABC means that what we have at hand is a national poll, regardless of whether people in every province were sampled. This may be annoyingly simple-minded but if we agree that the ABC poll is RS, then how it describes the situation (it describes it as a national poll) is what we need to rely on. Novellasyes (talk) 19:20, 18 November 2025 (UTC)

Wait, so the debate is whether or not this should be referred to as a "national" poll? That's the main issue? (If so, then I'm officially lost on why this is being discussed here under a "no original research" heading). BetsyRogers (talk) 20:05, 18 November 2025 (UTC)

@BetsyRogers If only one could tell the filing user exactly that (that this isn‘t an issue of OR) and that repeating themselves twenty times over won‘t help their case … SdHb (talk) 21:22, 18 November 2025 (UTC)

Oh. I thought the issue was whether you had added some of your own statistical analysis to the table, separate from whatever was listed/stated in the pdf from the ABC report. BetsyRogers (talk) 22:28, 18 November 2025 (UTC)

@Novellasyes . @BetsyRogers What I am doing is based on pages 39 and 40 of the main source.

https://abcnews.go.com/images/PollingUnit/1083a1Afghanistan2009.pdf

i think Everything we say should be based on pages 39 and 40. Because these two pages are our main source in the table not page 1.

this table .

Of course, there may be material in this case that is related to Wikipedia:Neutral point of view/Noticeboard.

@Novellasyes Is our priority page 1 or pages 39 and 40? we use page 39-40 data in table. The page 1 and 39-40 contradict each other. Wikipedia's policies require us to base our content on verifiable facts, not claims.

@SdHb prioritize a general claim on page 1 over the verifiable, methodological data on pages 39-40. his position is synthesizing

the page 1 claim with the pages 39-40 data to create a new conclusion: "Data from 31 provinces can represent as 34 provinces ! !

i think This synthesis and original research. Badakhshan ziba (talk) 21:23, 18 November 2025 (UTC)

@BetsyRogers I think that due to the long length of the discussion, I may have misrepresented my point.

And some one may not have understood the point. If I am allowed by Wikipedia rules I suggest that this discussion be closed and after that i open a new discussion tomorrow or the day after that just only focused solely on the issue of ( NOR ). Badakhshan ziba (talk) 21:39, 18 November 2025 (UTC)

@SdHb The source's own methodology in page 39 -40 note verifies the data is from 29 to 31 provinces, yet you insist on placing it in a table cell that represents 34 provinces??? This is the heart of the problem.

please Tell me why you insist on using incomplete and incorrect data in the table? This completely questions your impartiality and neutrality.

Removing the 2004, 2005 and 2006 ABC survey data from the table does not create any problem with the data.

But adding this data artificially and incorrectly changes the percentage of Afghan ethnic groups in favor of the Pashtuns. Do you intend to increase the percentage of the Pashtuns in the table by removing data from 3 to 4 Afghan provinces? This also violates Neutral Point of View .. Badakhshan ziba (talk) 21:01, 18 November 2025 (UTC)

@Novellasyes @BetsyRogers I think combining and synthesizing the claim and content on page 1 with the content on pages 39 and 40 and creating a new, false and defect conclusion, is related to this link.

this link --  If a single source says "A" in one context, and "B" in another, without connecting them, and does not provide an argument of "therefore C", then "therefore C" cannot be used in any article

Badakhshan ziba (talk) 22:03, 18 November 2025 (UTC)

I'm still not completely understanding where the "original research" is. Could you link to one example (for example one specific year) in the table and specifically state in what way you think original research is being included? BetsyRogers (talk) 22:10, 18 November 2025 (UTC)

To clarify, can you link to the exact version of the table you're discussing (I think I've seen multiple versions) and explain exactly what we should be looking at to determine if there's any original research being included? Thank you. BetsyRogers (talk) 22:13, 18 November 2025 (UTC)

@BetsyRogers @Novellasyes according to Wikipedia rules,Synthesis of published material this link it is written that =

If a single source says "A" in one context, and "B" in another, without connecting them, and does not provide an argument of "therefore C", then "therefore C" cannot be used in any article.

In this case, a violation of the rules has occurred regarding to synthesis .

@SdHb are combining two separate parts of the ABC source.

Part A (from page 1, the article): This is a national survey" (claim) ! ! ( just national in 2007 and 2009)

Part B (from pages 39-40, the methodology): The factual data showing the survey of 2004 .2005.2006 was conducted in only 29-31 province out of 34 provinces.

Part C (then he synthesizing and combining these two parts to create and advance a new conclusion by his own false statistical analysis: "Data from 30 provinces can represent as 34 provinces ! !

And that no argument was provided in the ABC poll for this Contradiction between the page one and pages 39 and 40.

As a result, according to this Wikipedia rule, data from 2004, 2005, and 2006 must be removed.

in this link we can find table this link

Please look at the period 2004 - 2021. Click on the expand column.

The rows related to 2006 ABC, 2005 ABC, and 2004 ABC should be deleted according to a violation of the synthesis and wikipedia rules. Thank you for your attention.

( If a single source says "A" in one context, and "B" in another, without connecting them, and does not provide an argument of "therefore C", then "therefore C" cannot be used in any article.) this link

======= Badakhshan ziba (talk) 23:06, 18 November 2025 (UTC)

according to the given explanation, please tell me whether there was a violation of the rules related to synthesis and original research or not? Thank you. Badakhshan ziba (talk) 23:20, 18 November 2025 (UTC)

@Badakhshan ziba, No. Based on the information you've provided here (in your intial post and subsequent clarifications) I don't see any violation related to original research, synthesis, or even neutral POV. BetsyRogers (talk) 23:40, 18 November 2025 (UTC)

Okay

What part of „it‘s not me who says this is a national poll but the very source itself“ is OR exactly? As I have pointed out earlier, if anything, it‘s you who is doing the OR here, because you’re insisting that "true national coverage" supposedly requires sampling all 34 provinces. ABC News itself states it‘s national, so it‘s verifiable. It’s not me who does the conclusion. End of the discussion. @BetsyRogers I would like you to stop giving them a platform to repeat unreasonable claims for the 1000th time. By now it‘s obvious that they are alone with their opinion, as 6 separate users are sharing a consensus that this case in fact isn‘t OR by me. Thank you. SdHb (talk) 23:20, 18 November 2025 (UTC)

Done. ✅️ BetsyRogers (talk) 23:41, 18 November 2025 (UTC)

@SdHb You still claim, citing page one, that this is a national poll and should be included in the table, but you still ignored pages 39 and 40. Why? WHY ?

You are deliberately ignoring pages 39 and 40 and trying to advance your position.

The source saysA ("national survey") on page 1.

- The same source says B ("conducted in 31 provinces") on pages 39-40.

- You are synthesizing A and B to create C: "Therefore, the data from 30 provinces can be presented as representing 34 provinces in the national table."

Hey @SdHb, to put it simply, we cannot synthesizing and combining the article on page 1 with pages 39 and 40 and then claim that the source data represents 34 provinces and the whole of Afghanistan. A+B=C Badakhshan ziba (talk) 23:47, 18 November 2025 (UTC)

@BetsyRogers,@Novellasyes

I apologize, I forgot to mention page 38. We got the data for the Ethnicity table from page 38, not pages 39 and 40.

So the ABC News source itself is in violation.

This is also Wikipedia's rule regarding sources abc

( If a single source says "A" in one context, and "B" in another, without connecting them, and does not provide an argument of "therefore C", then "therefore C" cannot be used in any article)

PART A = Page 1

PART B = Page 39-40

PART therefore C = Page 38

so in this result again we should omit abc data. https://abcnews.go.com/images/PollingUnit/1083a1Afghanistan2009.pdf Badakhshan ziba (talk) 23:59, 18 November 2025 (UTC)

if Assuming that @SdHb has not violated any rules (it is quite clear that the information on pages 39 and 40 is deliberately ignored),

but the source itself is incomplete and contains errors, and violates Wikipedia rules.

this law

This law is mostly about the source, not about people or editors . ( Although individuals can sometimes be subject to this law )

Does this source apply and relate to this law in this case? if yes please say yes or if no. please say that why not?

==

In my opinion, there is no doubt that the source violated this law and at least the years 2004, 2005, and 2006 can be removed from this Table according to this law.

It's late here where I am now need to sleep. I hope you will give an impartial and careful opinion on this matter. Thanks alot. Badakhshan ziba (talk) 00:24, 19 November 2025 (UTC)

I'm not going to talk about @SdHb anymore.

I suggest closing this very long discussion and then opening a new discussion just only about the source itself so that people can give poinion about the source itself and comment on it.

This is a very important and sensitive issue. I hope at least 4 or 5 people will give their opinion on this.

@BetsyRogers, can I close this discussion? Badakhshan ziba (talk) 00:54, 19 November 2025 (UTC)

Badakhshan ziba: You seem to be under the mistaken impression that if you just find the right rule that the ABC source violates, then we'll be forced to omit that source from the article. Wikipedia does not work that way.

Wikipedia has neither laws nor rules, only guidelines. (It is, at best, a common law or customary law system.) Even if you find a WP:ESSAY directly on point, editors will always be able to cite WP:IAR. One of the central principles of Wikipedia is to include as many sources as possible, but with caveats as necessary. AFAICT the only exceptions are sources that are entirely faked (to be clear: this does not apply to the ABC News source, since even you don't dispute that the didn't fake their data for 31 of the provinces), or published by crackpots (again, not ABC News).

Your argument against the ABC News source seems to be that the content of the source is an unjustified extrapolation, that Afghanistanis care a lot about these numbers, and that we should be cautious in any claims to avoid inciting political/ethnic conflict in Afghanistan. Wikipedia doesn't care about any of those things. Wikipedia is fine with other people's extrapolations, Wikipedia is not swayed by controversy, and Wikipedia includes information even if it makes people violently uncomfortable.

IMHO this discussion is snowballing towards a conclusion that the source should stay. Creating a second discussion would waste editor time and not produce much of value. I recommend that you make your peace with the fact that Wikipedia is going to include the source and the data from it. You can stop wasting your time here and find other areas of the encyclopedia to improve.

It's already been 2 decades since the ABC News poll. In another decade, nobody's going to care what a 30-year-old poll says in one line of a >40-row table. Bernanke's Crossbow (talk) 01:22, 19 November 2025 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Japanese mythological genealogical trees

I would like input on some mythological genealogical trees that were made a while ago. Some are pages, some are templates.

Template:Three generations of Hyuga
Template:Eight generations of Izumo
Template:Generations of Jimmu
Family tree of Japanese deities
Family tree of Japanese monarchs

I have two primary concerns with these:

1. They often present myth as fact. This is particularly evident in the last two which show how the emperors of Japan are decended from the gods. While this is the official claim, there is obviously no proof this is true. I think having an "official genealogy as published by the Imperial household" would be fine, because then we're just stating what they're claiming, but the current trees are questionable.

2. There is no single source for the genealogy of Shinto deities, meaning any attempt to make a single genealogy requires synthesis. There are two main primary sources, the Kojiki and the Nihon shoki (as well as several other minor primary texts) which have conflicting versions of the same myths and it is common for Shinto scholars to synthesize the information between them, but a Wikipedia editor should not be the one doing it. All these templates use several sources, often citing both the Kojiki and the Nihon shoki directly. An example would be Ōyamatsumi in the Three Generations of Hyuga template, who is said in his article to have different genealogies in the different texts. The same can be said of Takamimusubi who is in four of these five trees.

My gut is that we could have genealogies that were created by scholars, but that we shouldn't cobble together sources in order to create our own genealogies.

We could potentially have some minor genealogies "according to the Kojiki" as it's the most straightforward, but the Nihon shoki also exists and it would be misbalanced to ignore it, and it contains several conflicting stories, so someone would need to make a decision on what story was "right".

I personally feel it might be better to do away with the templates and handle it like it was for Ōyamatsumi where it lists his relations as they are the two major primary texts.

Please let me know what you think.— Preceding unsigned comment added by Erynamrod (talk • contribs) 19:02, 14 November 2025 (UTC)

Emily Riehl

It is a documented fact that mathematician Emily Riehl has performed in a band named "Unstraight", and has also published mathematical research on "unstraightening". I think that the conjunction of these two facts does not lead readers to any inappropriate conclusions about how that coincidence of nomenclature may have come about, and is not of great significance to the biography but adds interest to it and is appropriate to mention for that reason. In fact, this exact conjunction of facts was the hook for a DYK entry that ran in 2018.

After a not-logged-in editor tried to remove even the band name because it "could be misconstrued as something to do with sexual orientation" (duh, the subject also served as a board member of a notable LGBTQ association), established editor User:Cagliost backed up the other editor in the same removal until confronted with source overkill for the band name, and has since repeatedly tried to remove the mention of mathematical unstraightening, next to the band name, claiming (per the topic of this noticeboard) that this conjunction is intended to lead the reader to an unsourced conclusion and therefore forbidden by WP:SYNTH. In support of this position, Cagliost jumped to an unsourced conclusion (that the band was named after the mathematics, something I have no knowledge of and no intention to imply) and moved this unsourced conclusion into a footnote. When I reverted, saying all this, Cagliost removed the material altogether for the fourth time and demanded a third opinion.

So here we are: can I have a third opinion about whether it is permissible to mention mathematical unstraightening next to the band name Unstraight (as in this old version and the DYK hook), with the only intended connection between the two being that this is an intriguing coincidence of nomenclature that could plausibly have many different explanations and makes a connection between two facets of her life? Or is this going too far in leading readers to some particular conclusion about what this coincidence of nomenclature actually means? —David Eppstein (talk) 08:13, 18 November 2025 (UTC)

• Dare I mention her contributions to homotopy theory? EEng 23:55, 21 November 2025 (UTC)

Fuzzy concept

The page Fuzzy concept can be kindly described as a rummage sale of WP:OR nonsense. But it is in fact just an alphabet soup of meaningless items. I do not want to just wipe it out by myself, and certainly do not see how it can be easily fixed. Should it just be AFDed to avoid embarrassment for Wikipedia? If someone places the AFD flag I will support it. Or perhaps redirect to Fuzzy logic or Fuzzy set ? Thanks Yesterday, all my dreams... (talk) 07:34, 24 November 2025 (UTC)

US Commitee reports

I recently came across new articles on the FBI Richmond Catholic memo investigation and its Arctic Frost investigation, as well as older articles that have had sections about these investigations added to them (e.g. Paul Abbate) that heavily use, to verify factual claims, various reports and memos of Chuck Grassley's Senate Judiciary Committee which has been raking through Biden-era FBI investigations for arguably partisan reasons. My initial thought was that these are primary sources and over-using them risks OR. But on reflection I wonder if these are not just primary but also hyper-partisan, akin to press releases by politicians. Should they therefore be treated as opinion pieces that need attribution, rather than as primary documents? (Neutral eyes on these articles also likely beneficial.) BobFromBrockley (talk) 03:54, 25 November 2025 (UTC)

Congressional committee reports are official government documents, not press releases. WP:PRIMARY allows primary sources for straightforward facts - "the investigation issued 197 subpoenas" doesn't need a secondary source to explain what it means.

Calling Senate Judiciary oversight "arguably partisan" could apply to literally any congressional investigation of the opposing party. Should we discount all Democratic committee reports from 2017-2021 too?

The actual articles use tier-1 sources (CNN, AP, NBC, PBS, Washington Post, Axios) as their backbone. Committee reports fill in specific details those outlets didn't cover. The DOJ IG's April 2024 report independently corroborates the Catholic memo findings.

If you think specific claims are unsupported, please take it to the article talk pages. That's where the content disputes belong not here with vague concerns about entire source categories. Bladerunner24 (talk) 21:30, 25 November 2025 (UTC)

My question is about sources like these:

[6]

[7]

- BobFromBrockley (talk) 04:27, 27 November 2025 (UTC)

Those are two different things:

The Virginia AG press release documents an official action - 20 state AGs sent a letter. Primary source for the fact the letter was sent.

The Grassley/Johnson release publishes actual FBI documents obtained through oversight. The underlying FBI documents are primary sources; the press release is just how they were published. Same as FOIA releases.

Both are used with attribution ("According to Grassley," etc.). And the articles use CNN, NBC, PBS, AP, DOJ IG as their backbone.

If you have concerns about specific claims, please take it to the article talk pages. Bladerunner24 (talk) 05:47, 27 November 2025 (UTC)

Whether OR or not would depend how the sources were used. While partisan, they do contain a significant concentration of basic facts, and the editor has recently started to trim any OR down by using secondary sources, so for most of what they were doing I don't see this as an OR issue. Ⰻⱁⰲⰰⱀⱏ (ⰳⰾ) 08:54, 27 November 2025 (UTC)

"DOJ IG" is another primary source, if I understand right: it's the Inspector General report on FBI and DOJ actions in the 2016 election. I appreciate the editor has trimmed down their use and started adding attribution and primary sources, so these specific articles are in a far better place (they certainly didn't use CNN etc as "their backbone" when I raised the question!). However, my question, which perhaps I didn't articulate well, was a more general one. Something like: To what extent can we treat press releases by elected officials, especially partisan press releases on contentious topics, as usable primary sources for facts? I feel that the response to my raising this (find secondary sources, attribute) shows that the consensus seems to be that these are poor sources for facts, unless the facts are banal. Is that right? BobFromBrockley (talk) 10:18, 27 November 2025 (UTC)

There might be a policy on this specific issue, but my not within my knowledge. I would merely advise the editor to use discretion, taking into account fact manipulability, likelihood of omissions and in this case especially weasel words, excessive interpretation with insufficient supporting evidence, etc. So pretty much the same as most news articles. The facts concentration just happens to be denser in these government press releases, which might give an inexperienced encyclopedist a false impression of neutrality? This isn't an ordinary new editor, though. They seem to have a decent grasp on what they're using and how they're using it, so I wouldn't worry too much about them getting carried away now that they have been made aware of the relevant policies and are complying. It is also worth noting that almost all of the sources being objected to contain a mix of primary and secondary text. Some are almost completely primary, others secondary. Much easier to handle on a case-by-case basis at the article's talk page than in a general discussion. Ⰻⱁⰲⰰⱀⱏ (ⰳⰾ) 10:32, 27 November 2025 (UTC)

Village Pump discussion: FRINGESUBJECTS vs SYNTH

I've opened a discussion at Wikipedia:Village_pump_(policy)#Fringe_theories_and_synthesis_policies_-_contradiction?. Fences&Windows 21:24, 25 November 2025 (UTC)

Template:Empires

Looks like Template:Empires became an inflated OR-ish hodgepodge of various entities not widely considered empires, but which are persistently added by some editors. Mostly those entities were localized kingdoms or realms. In the past, I removed a few, but over time they have been re-added. My particular concerns include Kingdom of Armenia, Dʿmt, Calakmul, Grand Duchy of Lithuania, Chagatai Khanate and some others. An overall look suggests major purging. Any recommendations? Brandmeister ^talk 09:05, 26 November 2025 (UTC)

I agree with your perception that unless an entity is widely recognized as an empire (or having been one), it shouldn't be in that template. Then the question becomes, what to do about that. There are two different things you could end up having knock-down dragouts about. Those two things are (1) is it true (as you and I think it is) that to be in that template, the entity has to be widely regarded as an empire and (2) even if everyone agrees with that, then I think you'd have to proceed to identify the entities currently in the template that don't qualify. Then you would probably get a knock-down dragout for each of those entities. So if there are ten contested ones, that would be a lot of talk page going-back-and-forth about whether RSes have or haven't said it is an empire. What you might consider doing is opening up a discussion on Template talk:Empires about this, and direct folks here to have the conversation about (1). Then if you get a consensus here about (1) -- the idea that something has to be identified by RS as an empire or else it shouldn't be in the template, you can proceed to the work on (2). Novellasyes (talk) 18:20, 26 November 2025 (UTC)

I see. Might take more time to resolve, though... Brandmeister ^talk 10:58, 27 November 2025 (UTC)

If you achieve consensus on what has to be true of an entity to be listed in the template (namely some relevant RSes need to have described it as an empire), then I would assume that you would then proceed to remove from the template the ones that as far as you are aware, don't make the grade. But then I wouldn't be surprised if people who want a particular entity to be listed in the template would then look for relevant RSes that have described it as an empire. But maybe not! It's possible that once you get agreement on the main principle, it will be easy to remove the non-empire entities from the template because people who would otherwise want a particular entity listed in the template won't try to make the claim that such-and-such entity has been described by relevant RSes as an empire. Novellasyes (talk) 16:16, 27 November 2025 (UTC)