Talk:Religion in France/Archive 1

Religion in France
So the section contradicts itself, or there's no apparent consensus on religion distribution in France. -- Andersmusi[[tittles are da bomb| tittles are da bombcian ]]  VOTE  17:50, 24 October 2007 (UTC)

There isn't any consensus on religion in Franceof France|French reformed]] background. —Preceding unsigned comment added by 88.140.20.72 (talk) 12:48, 9 March 2008 (UTC)

If the United States CIA factbook is going to be listed as a source for the 2015 statistics, it should be noted that it has always been significantly different than French based polls and studies. — Preceding unsigned comment added by Rvawiki (talk • contribs) 20:15, 10 November 2015 (UTC)

It leaves one wondering "why the sudden shift in religious affiliation," when the main change is that a non-french source was used. — Preceding unsigned comment added by Rvawiki (talk • contribs) 20:17, 10 November 2015 (UTC)

Cleanup
This article needs a bit of a cleanup. Catholis who don´t believe in God ?! Sorry, but that´s laughable.14:29, 5 March 2008 (UTC)
 * Apparently it's how they responded to the poll. It doesn't make sense, but supposedly it's what they said. Jedibob5 (talk) 03:43, 7 January 2009 (UTC)
 * Yeah, this makes sense because catholicism is more a matter of tradition than a matter of faith. They may not believe in God, but people still go to churches for weddings so a not to offense the elderlies in a familly who are believers.
 * One can be culturally Christian but not necessarily religious. 86.154.119.85 (talk) 18:45, 27 July 2009 (UTC)
 * Of course it makes sense. Go to France, and you'll see. Real believers are very, very rare. Most people who declare themselves as Catholics only mean that historically, that's what their family was. It's a cultural thing. loulan (talk) 20:07, 19 September 2009 (UTC)

The problem in this article is that the person who wrote the Wiki interpreted the results of several polls and wrote his own personal interpretation, a clear case is the 2006 CSA poll, the wiki author mixed pears with apples to come to the 9% conclusion and there is no a single word about it in the poll documentation —Preceding unsigned comment added by 190.142.64.97 (talk) 15:20, 25 November 2009 (UTC)
 * I corrected the error about the 9%. The 9% are not from the poll, but from an estimation of the population of Muslim origin.The International Religious Freedom Report (2007) states it clearly. The poll says 4%. I also corrected the year of the poll. The US report writers got confused with the year of its publication in the French press (January 2007). The difference between the estiimated 8-10% (or "9%") and the 4% can explained by the people of Muslim origin not practising, not believing in God, or simply not identifying themselves as Muslims. Ideportal (talk) 15:35, 27 July 2011 (UTC)

Hey, I just noticed that the link for references 1 and 3 are broken. MOM4Evr (talk) 23:25, 20 April 2011 (UTC)

The article passes judgment
The article uses phrases such as "very controversial". Very controversial among whom? Which sources for that?

The article should especially distinguish controversy in, say, the US press from controversy inside France. David.Monniaux (talk) 12:13, 14 July 2010 (UTC)

The beginning Paragraph states that the overall irreligiousness of the French citizenry, is the what makes France one of the " more secular" countries in Europe. But 'secularity' has nothing to do with how much an entity lacks or stifles faith, but by the plurality and acceptance of the freedom to religion and worship, including its manifestations. USA is more secular than most of Europe. Secularism does not equate to atheism.

The article should use a more appropriate word, 'irreligious' or 'agnostic' for example in place of 'secular'. — Preceding unsigned comment added by 14.97.64.104 (talk) 16:33, 5 March 2015 (UTC)

Cults and Minor Religions
Are all of the listed religions, IE Mormonism, on the list of government watched sects? From what I understand, Jehovah's Witnesses are on the 'cultes associale' which is a sort of middle ground between an accepted religion ('culte' in French) and the non-accepted 'Sectes'. --IronMaidenRocks (talk) 20:15, 3 September 2011 (UTC)
 * The LDS church is not officially listed as cult in the parliamentary reports. It appears in some annual reports of the MIVILUDES (e.g. 2006), but much less than the Church of Scientology or Jehovah's Witnesses. --Europe22 (talk) 22:56, 3 September 2011 (UTC)

laïcité (or "freedom of conscience" ?
This statement is not exact ! "Laïcité" is not only freedom of conscience, it is the freedom of religion AND the separation beetween churches and state : politics do not intertfere in religion, and religion do not interfere with politics, as it is says after in the article (I'm french, so sorry for the bad english) — Preceding unsigned comment added by 194.153.110.6 (talk) 14:52, 24 August 2012 (UTC)

Semi-protected edit request on 29 November 2016
In the "After Charlie hebdo shooting" part, it should be added that the incidents were reported respectively by the Education minister, Najat Vallaud-Belkacem, and a teacher regarding the "Allahu Akbar" incident, as said in the sources. 81.194.35.225 (talk) 14:42, 29 November 2016 (UTC)
 * Which sources state that? What section? -- Dane 2007  talk  06:21, 4 December 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 7 one external links on Religion in France. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20080222013645/http://www.ambafrance-us.org/atoz/secular.asp to http://www.ambafrance-us.org/atoz/secular.asp
 * Corrected formatting/usage for http://www.lefigaro.fr/election-presidentielle-2007/20070116.WWW000000233_premiere_etape_de_sarkozy_dans_la_france_eternelle.html
 * Corrected formatting/usage for http://www.femmesdavenir.segoleneparis.fr/node/33
 * Added archive https://web.archive.org/web/20130117013643/http://www.fgi-tbff.org/sites/default/files/elfinder/FGIImages/Research/fromresearchtopolicy/ipsos_mori_briefing_pack.pdf to http://www.fgi-tbff.org/sites/default/files/elfinder/FGIImages/Research/fromresearchtopolicy/ipsos_mori_briefing_pack.pdf
 * Added archive https://web.archive.org/web/20070304094059/http://a1692.g.akamai.net:80/f/1692/2042/1h/medias.lemonde.fr/medias/pdf_obj/sondage030416.pdf to http://a1692.g.akamai.net/f/1692/2042/1h/medias.lemonde.fr/medias/pdf_obj/sondage030416.pdf
 * Added archive https://web.archive.org/web/20121016062403/http://redcresearch.ie/wp-content/uploads/2012/08/RED-C-press-release-Religion-and-Atheism-25-7-12.pdf to http://redcresearch.ie/wp-content/uploads/2012/08/RED-C-press-release-Religion-and-Atheism-25-7-12.pdf
 * Added archive https://web.archive.org/web/20090806045719/http://www.quid.fr:80/2007/Religions/Les_Sectes_En_France/1 to http://www.quid.fr/2007/Religions/Les_Sectes_En_France/1

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 12:35, 30 November 2016 (UTC)

Better source
I found this source by Ifop (http://www.ifop.com/media/poll/religions_geo.pdf) and I added it in the pie chart. The sample was 91.559 but the data is from 2006. I found also other sourveys from Ifop (with also a very big sample (over 100.000) but I could find only data for Catholics and Muslims for respectively 2010 and 2009, I tried to merge that data with the 2006 ones for Protestants, Jews and Non religious but the total is 100.3. If you find newest data for other religions published by Ifop, please update the pie with new data. I leave here the links for Catholics in 2010 (64.4% http://www.ifop.com/media/pressdocument/238-1-document_file.pdf) and Muslims in 2009 (5.8% http://www.ifop.fr/media/pressdocument/48-1-document_file.pdf).--FrankCesco26 (talk) 14:35, 15 May 2017 (UTC)
 * I strongly contradict to exchange a 6 year old statistics by an 11 year old, though the sample base might be bigger. Main reason is that a survey from 2006 asking adults produces (in 2017) a gap of all under 30 who are not covered by this 2006 data. And membership data are rapidly shrinking. Other options might be newer PEW or EU data. BR Ulrich--Nillurcheier (talk) 07:06, 16 May 2017 (UTC)
 * stop adding back data from 2009 when we have data for 2016.--Wddan (talk) 16:45, 15 September 2017 (UTC)
 * Newer data is not better data. 2009 data is based off a very good and large sample of 93,000 people, instead of a very schematic and strange source based off a small sample. Muslims in France can't be ony 2%, as Buddhists! A study from IFOP dating 2016 (so the same year of your favourite source) say that Muslims in France are 5.6% calculated on a sample of 1029 people of muslim culture of wich 874 declared theirselves muslims. Link: http://www.lemonde.fr/religions/article/2016/09/18/une-enquete-de-l-ifop-offre-un-portrait-nouveau-des-musulmans-de-france_4999468_1653130.htmlFrankCesco26 (talk) 19:51, 15 September 2017 (UTC)
 * I found a new source, here it is: http://www.institutmontaigne.org/res/files/publications/a-french-islam-is-possible-report.pdf
 * Survey results are the most realistic ones I found, I will update the article with this new 2016 source, based off a sample of 15.000.FrankCesco26 (talk) 20:13, 15 September 2017 (UTC)
 * At page 13, this Institute Montaigne report says that "in the original sample of 15,459 people, over 47 % of those aged 15 and over identify as "Christian", 37 % "of no religion", 6 % declined to answer the question, and a little over 3 % follow a minority religion other than Islam"; these are very similar to Ipsos' (and others') findings, and strangely they are not what reported in the pie chart, which apparently portrays a sub-sample.--Wddan (talk) 21:02, 15 September 2017 (UTC)
 * this data has not been published by an official organ of any institute, is just a press article, and is attributed to a 2014 book by Yann Raizon du Cleuziou entitled Qui sont les cathos aujourd’hui ?, and further attributed to Ipsos.--Wddan (talk) 09:58, 16 September 2017 (UTC)
 * Well, 47% data include the non-respondent population, wich is excluded in the final data, and so you obtain 51.1%. Also, I found Ipsos source from 2017 saying that 53.8% percent of the population is catholic, based on a sample of 28.000, the double of the 2016 IFOP source. I am still searching the full data, in order to use a newer source that shows also the Christian denominations. So, the findings of the multi-nation survey very different from the majority of the other surveys, wich say that the Muslim population is somewhere beetween the 5% and 7.5%, and obviously not 2%. So they shouldn't be merged, as you wanted.
 * Also, question on religion for the 2017 source started in June 2016, and not in 2014, as you can see here: https://international.la-croix.com/news/the-sociology-of-french-catholics/4491
 * Also, in 2014 Yann Raizon described that according to him there are four types of catholic families, but the study was conducted in June 2016 and the results has been published in the early 2017.


 * I paste here the information of date and sample from the link above:


 * In implementing its survey of “involved Christians,” Ipsos took a representative sample of the metropolitan population aged over 18 years numbering 28,204 persons, and a sub-population of 15,174 persons characterizing themselves as Catholics practicing or not. This represents 53.8% of the population. This group can also be subdivided based on their mass attendance.


 * Another representative sample of 1,007 people from within this group representing Catholics regarded as “involved Catholics,” namely practicing Catholics (weekly, several times a month, major events and religious feasts), whether they described themselves as “involved” or not, and non-practicing Catholics who described themselves as “involved".


 * The survey was carried out in June 2016 using the quota method. The margin of error for a given percentage depends on the size of the sub-sample considered. Thus, for a population of 1,000, the margin of error was around three per cent, while it may be up to five percent for a population of 500 and seven percent for a population of 200.FrankCesco26 (talk) 10:44, 16 September 2017 (UTC)


 * I found in this page (https://www.ipsosglobaltrends.com/about/) that the source clearly say that the findings aren't rapresentative of the total population, but only of the working age population; in particular it says "the results should be viewed as representative of a more urban, affluent and ‘connected’ population." and so they shouldn't be used as main source of the country. In fact, in the page it say that the sample include only urban people from 16 to 64 and thus, the sample is not rapresentative of the entire countries' population.FrankCesco26 (talk) 12:52, 16 September 2017 (UTC)
 * The non-responding population of the Montaigne 2016 survey has to be properly represented, since it might be of any religion or none.
 * Regarding the type of population that Ipsos 2016 represents, the official page does not say that, indeed. It says "In established markets with a higher level of internet penetration (more than 60% online), the results can be taken as representative of the general working age population. However, in emerging markets where internet penetration is lower, the results should be viewed as representative of a more urban, affluent and ‘connected’ population." In other words, in developed countries the survey represents the general working-age population, while in developing countries it represents the urban popolation. France is among the first ones.--Wddan (talk) 15:28, 16 September 2017 (UTC)
 * And please stop removing the CSA 2006 data. The original location is still linked, though the link is dead. These figures were reported in previous versions of this article and in French Wikipedia as well.--Wddan (talk) 15:39, 16 September 2017 (UTC)

New issues
The most recent edits by user FrankCesco26 have produced the following issues:
 * The box for atheism was unexplainably removed from the table;
 * The section about Buddhism was unexplainably moved to the bottom of the list ignoring the alphabetical order.

--Wddan (talk) 09:53, 9 October 2017 (UTC)
 * I removed the box for atheism since only Ipsos Global trends uses it and separing it from other non-religious people wasn't useful for the comparision.
 * Also, I moved Buddhism in the bottom becouse it is a secondary religion in France. If you want to move it on the top you can.
 * Regarding the problems produced by your edits, where are all of the other useful information regarding other surveys? It seems that you have removed about 20 sources with no reasons.
 * --FrankCesco26 (talk) 11:58, 9 October 2017 (UTC)
 * The reason, as explained in edit summary, was that the sources were old and often unreliable. Articles should not be simply amasses of statistics; data need selection and contextualisation. Some data are less reliable than others and some others are outdated and, when they are not relevant, they should be expunged. Sorry, but my editing style is completely different from yours, in that you keep stuffing articles with raw, randomly collected, often old data.--Wddan (talk) 08:59, 10 October 2017 (UTC)
 * The only problem is that removing half of the article only becouse one user think that some sources aren't relevant doesn't make the article better.--FrankCesco26 (talk) 13:42, 10 October 2017 (UTC)
 * Define what better means. To me, "better" does not mean an amass of decontextualised data, often taken from minor agencies or think tanks.--Wddan (talk) 07:56, 11 October 2017 (UTC)
 * PS: You keep making grammar and form errors in your English. For instance, "becouse" in proper English is written "because".--Wddan (talk) 07:58, 11 October 2017 (UTC)
 * Data coming from "minor" (according to whom?) agencies doesn't mean that is not reliable and should be completely removed. Also, they could be easily contextualised, and not hastily removed.-FrankCesco26 (talk) 13:00, 11 October 2017 (UTC)


 * And, in addition to the edit summary, I add: do not make again wholesale reversals, since my version not only contains edits to the pie chart but also a variety of other, minor, improvements.--Wddan (talk) 14:50, 22 October 2017 (UTC)


 * The article can't be keeped as it is in this moment. You merged adjusted data with not-adjusted data and this version is not supported by the source itself. In particular, the source doesn't cite the 6.6% of unanswered people, but 6%. Also, Muslims and Jews aren't reported in the approximative raw data, and they are reported only in the detailed adjusted data. You said that the Muslims have a lower share in the detailed data rather than the raw data, well, they aren't cited in the raw data.
 * You should also know that in the approximation of data, numbers become very imprecise and you can't calculate the raw Muslim percentage by simply doing a subtraction.
 * On balance, don't restore a version of the article that isn't even supported by the source itself.--FrankCesco26 (talk) 17:53, 22 October 2017 (UTC)
 * The source (Institut Montaigne) at p. 13 says that in the original sample 47% are Christians, 37% irreligious, 6% unanswered, 3% minorities other than Islam. The remaining 7% is logically inferred that represents Islam. I have fixed the whole picture accordingly. Also, recently-established consensus says that the unanswered datum must be included. In our case the source itself provides full data including the datum for the non-respondents, so we must include it.--Wddan (talk) 13:26, 23 October 2017 (UTC)
 * Wait, the general consensous said that shouldn't be excluded, not that they must be included at all costs. In this case, the final data of the source is preferred since it's more precise (note the commas) and could have been weighted. So the higher share of Muslims in the raw data could be becouse of the approximation on the nearest number (removing the commas) or the weighting, since there could be an oversample of Muslim for statistical matters (the main survey goal was to give demographics informations about the attitudes of the french muslims). Also, I want to remember you that the number of Muslims isn't cited in the raw data section, and has been calculated by doing a subtraction, so it's only speculation.
 * So, all things considered, unanswered people shouldn't generally be excluded if the source leaves them, but if the source removes them (in this case likely to apply a weight, in many SPSS programs unanswered sample is removed by default), adjusted data must be used, since it's more precise. I will wait a bit before reusing adjusted and weighted data. I would also know the opinions of, and . --FrankCesco26 (talk) 19:13, 23 October 2017 (UTC)
 * Consensus was in favour of including the unanswered datum when the source provides it, and our source provides it . In my opinion the datum for unanswered should always be included, for a number of more complex reasonings; people might have a variety of motivations for not answering. For instance, what if they belonged to "religions" which teach non-identification (even with doctrines), such as Buddhist viraga for cessating the craving for attachment which only produces woe? What if those 6% who chose not to identify were the theorised 5 million Buddhists? We simply may not know what the unanswered datum actually represents.--Wddan (talk) 19:31, 23 October 2017 (UTC)
 * That is your opinion, but generally people not answering are simple people that don't want to tell their personal information, for example they call you to the telephone to do a survey, but you don't want to tell strangers your personal information so you don't answer and you will be counted as unanswered. Why should be as complicate? Using approximative, not weighted, raw data and construing the datum for the Muslims only becouse it's cited a very approximative percentage of not-answered people is excessive. The source provides a clear and very precise data that could be also weighted, it makes no sense to leave the article as it is, it's speculation. Also, in every cases, it's completely impossible that the Unanswered people is composed by only Buddhists, these are absurd considerations, since there are only beetween 1 and 4 millions Buddhists in Europe. Also, the consensous wanted that unanswered people shouldn't be excluded if the source includes it. --FrankCesco26 (talk) 19:53, 23 October 2017 (UTC)
 * For the sake of clarity and brain gymnastics, my considerations about Buddhists who may not want to identify is a hypothesis meant to illustrate that there might be a million reasons why people choose not to identify (and, by the way, it is not absurd that Buddhists may choose not to identify; just study a little more about the content of what you write about), and these reasons that we may not know establish the reason why we should not exclude the unanswered datum (as established by consensus, by the way) . Also, the datum for Islam is not simply construed but logically inferred and implied in the list of data provided by the source. They provide the data for all other categories, they say, other than Islam, implying that Islam is represented by the rest.--Wddan (talk) 20:01, 23 October 2017 (UTC)
 * The source doesn't say the datum for Islam and you can't construe it, since there is an other field for "Undecided" in the questionary, so it's everything except logical. Clarified why you can calculate the datum for Islam, you should notice that the raw data is not weighted and data must be weighed if you pretend to have statistical value, and the detailed data is weighed. Also, you are playing with the general consensus to do what you want in this specific case .--FrankCesco26 (talk) 20:17, 23 October 2017 (UTC)
 * The datum for the undecided is clearly given, so there's no confusion with Muslims. I don't understand what you are writing about. As I wrote above the source provides data for "Christianity" 47%, "non-religious" 37%, "unanswered" 6%, and "minor religions other than Islam " 3%, so it is implied that the rest is Islam. As for the weighting, the source does not provide information about any weighting, even concerning the adjusted data.--Wddan (talk) 20:23, 23 October 2017 (UTC)
 * I mean that if you saw the questionary, there is also "Undecided" that in the weighted data is comprised by the 0.4% of the population but it's unknown in the raw data. This excludes you from doing a simple subtraction.--FrankCesco26 (talk) 20:39, 23 October 2017 (UTC)
 * , and, since you are users that are interested in the articles regarding religion demographics, I would know your opinion about this matter.--FrankCesco26 (talk) 15:22, 24 October 2017 (UTC)

Regarding the weight on data
There has been a discussion about using raw data or final data, and consulting the questionary I found that final data hasn't been only adjusted but had four weights, reported below: So, these weights are fondamental to make a survey data reliable, and only the data in the pie chart on page 13 has the weights. Raw data is only a report on how respondents answered and is not nationally representative.--FrankCesco26 (talk) 19:02, 25 October 2017 (UTC)
 * sociodemographics criteria: age and sex
 * socio-professional criteria: work
 * geographic criteria: region, size of the municipality, proportion of immigrants
 * civic criteria: nationality.
 * There was a consensus for using data which contain the non-respondents/unanswered, and only the raw data contain the datum for them. You are violating consensus (and 3RR), again.--Wddan (talk) 19:05, 25 October 2017 (UTC)
 * Also you broke the 3RR, but we are discussing now so I won't report anything. I am not talking about unanswered or consensus in other articles here, but about weight, and it's foundamental to make data reliable. The unanswered were already removed by the source, so it's not a problem: they did that in order to apply the weight.--FrankCesco26 (talk) 19:11, 25 October 2017 (UTC)
 * The discussion in the WikiProject:Religion was primarily driven by this very article, and the consensus was for keeping data which contain the unanswered datum. In this case they are the raw data, with the 6% unanswered. You are going against consensus and imposing your own revision of the article.--Wddan (talk) 19:17, 25 October 2017 (UTC)
 * Yes, but this case is not similar, in this case the final data also has a weight, so it's more reliable. I also had consensous for this revision.--FrankCesco26 (talk) 19:19, 25 October 2017 (UTC)
 * I can't find a single word that gave support to your version.--Wddan (talk) 19:21, 25 October 2017 (UTC)
 * reviewed my revision of the article and agreed with my version.--FrankCesco26 (talk) 19:45, 25 October 2017 (UTC)

Protection edit request
I agree with the protection of the article, but I ask that the protection be made to this version which contained undisputed data.
 * Consensus was established here about providing data containing the unanswered datum when available;
 * In this case it is available in the raw data set;
 * The version with the raw data set is that linked above;
 * User FrankCesco26 insisted on keeping his preferred version with the adjusted data which exclude the unanswered datum, against the consensus.--Wddan (talk) 21:27, 25 October 2017 (UTC)
 * Red information icon with gradient background.svg Not done: I have blocked the two parties that were edit warring on this article. One was blocked for much longer than the other due to his/her long history of edit warring and previous blocks. Now we may be able to consider unprotecting the article. What do you think ? &mdash; Martin (MSGJ · talk) 06:58, 26 October 2017 (UTC)
 * Sounds like peace and tranquility may now return. Did not realize it would be so simple. We can always reblock if it doesn't work out. Interestingly, my purpose in protection was to afford them the opportunity to work things out so they could constructively edit and avoid blocking. "All we are saying, is give peace a chance". -- the 1960'sDlohcierekim (talk) 11:34, 26 October 2017 (UTC)
 * PS Reinstated prior semi protection. Dlohcierekim (talk) 13:05, 26 October 2017 (UTC)

Modifications to French Revolution Period
The section on the French Revolution links to Dechristianization of France, but it needs some substantial expansion on the historical developments and legal status on religion during the French Revolution. Best of all, there is an image of the Cult of the Supreme Being with a brief caption, but its significance is never explained (nor its rival Cult of Reason). The expansion won't be too lengthy nor repetitive, but I'd like to add a little more historical context and information to this page for the French Revolution period. Squabdusky (talk) 19:28, 1 May 2018 (UTC)Squabdusky

May 2018, youth's religion pie chart removal
What is the criterion by which the pie chart representing religion among the youth was removed from the article, given that the article itself has a lot of space to accommodate its presence? The sources are institutional (two academic institutions and the European Social Survey), thus the data are to be considered of high quality.--Wddan (talk) 19:37, 7 May 2018 (UTC)

While it's true that the St. Mary's University source seems to be a respected source there is no need to be highlighted by pie chart, there are other reliable sources as well; no need to select one source that does not even present the whole population and create another pie chart for it.--desmay (talk) 19:18, 8 May 2018 (UTC)

Sample size
1000 cases is quite standard for a nationwide sample. That gives percentages like 40 percent = X, 10 percent = Y and 2% = Z. in this case IPSOS did not use subsamples, with higher margins of error. So its subsample of Y's and Z's are too small to be useful. However the overall full sample percentages are valid. In terms of public opinion polls around the world, "many surveys usually aim for at least 1000 cases" according to Rjensen (talk) 13:45, 2 March 2019 (UTC)

line chart
I removed the line chart in the demographics section because I think a chart should not combine data from different sources with different methodologies and because it duplicates the info in the table above (presumably since it has no references). The table is better because it does have references. If a chart is considered desirable, a horizontal or stacked bar chart would be better assuming we have good data. --Erp (talk) 17:04, 27 March 2019 (UTC)

biased and racist article, because it claims that religion is the common hypernym of religion itself, and of its actual hypernym: metaphysical worldview
Not all metaphysical worldviews are religion or religious.

A religion requires anti-physics. Antiphysical cosmogony and in theism (but not necessarily in deism) a continuation (but not necessarily continuous) violation of physics during the lifespan of the universe.

Even in non-personocratic / non-personocentric religions (religions without all-powerful and cosmocausal person-gods); if a religion is "well nurished" which means to follow its principles, which also include state(s) of mind, individual personhood and persons are supposedly benefited in a non-rigorously explainable manner. Some behavioral habits and attitudes are scientifically proven beneficial for psychological and cultural cohesion reasons. Religious people claim that they are benefited for reasons outside physics; except if they are solely culturally religious and non-aware atheists. Few aware atheists self-define themselves as religious for political and cultural reasons (you cannot force people attribute specific definitions to nouns; it works for the majority of the people, but few people react and maintain their own definitions; it happens in other cohorts as well; some mixed blacks define themselves as purely black and some aracialists definite themselves as "other" when they are forced by some US universities, not because they are of some rare cohort but simply (as the French state does) none has the right to force you attribute oneself to a race, especially for non-immediate genetic studies accepted by the individual; its a philosophical reason of self-definition; the separation of genetics from the civil rights. Few people do the same with "metaphysical worldview", they don't accept the common terminology or the right to force an individual to answer form a state of power (state, university, business, etc). (I purposely diverged here, because some articles on self-definition have not yet been created.)

Few statisticians (which usually are atheists according to their data) give the excuse: "I didn't know that not all metaphysical worldviews are religions or religious" but they never officially recorded that claim.

Few people have no metaphysical worldview at all; they are nones on the question of metaphysical worldview which is a hypernym of religion. One might be non-religious but might have some metaphysical worldview. Metaphysics is a field of study focusing on questions. The approach isn't defined; thus may be religious, scientific, spiritual, or a combination of approaches. Most metaphysics is based on layperson reasoning with the usage of academic terms. Mathematical reasoning isn't necessarily intuitive as layperson reasoning. Layperson reasoning is called common sense, which is beneficial for everyday life, but usually not a good tool for metaphysics, because the thinker is trapped by the emotional and wished interpretations of the mammalian (specifically human) limbic system. On the other hand, there's no rule in metaphysics which denies the sometimes abolute usage of physics. Read the strictly scientific metaphysicalists: Sean M. Carroll, Steven Weinberg, Max Tegmark.

from hypernym to hyponym
personal belief (including political and other biases) > metaphysical worldview > same-tiered hyponyms: religion(s), agnosticism (= openness to non scientific ∕ non rigorous metaphysical worldviews), atheism, irreligion (usually religious indifference)

nowadays erroneous metaphysical hypernymic semantic cladistics as presented in surveys
religious beliefs (ultimate hypernym ∕ erroneous) > same-tiered hyponyms: the names of the religions, atheism, agnosticism, irreligion

In many surveys they merge the "openness to non scientific and non rigorous metaphysics which is agnosticism" with the "rejection of any precosmic personhood, personocratic cosmogony and violation of physics by some person during the lifespan of the universe which is a colloquial definition of atheism". Some statisticians merge the non-religious personal beliefs, because they are discriminating against non religious people. Faithism is a form of racism. They have no excuse to systematically merge opinions above 2%.

In France there is a culture of aracialism (= not to confuse genetics with the social definition of the French citizen), which is great, because even if you are a racialist (not necessarily a racist) and you want to self-differentiate, the state has no business to push France towards official racial fragmentation. In everyday life common nouns are used for culture and genetical race, but the French state's main goal is to unite France, and not to allow racism under the excuse of some art, food, etc linked to specific genetical percentages.

Some French statisticians confuse aracialism with the suppression of non-religious people. They analytically report (not officially for the French government) all religious percentages, but many times merge all the non-religious people together. They claim that specifics about religion are worth mentioning. They claim that specifics about: 1. atheism∕non-personocracy at the cosmological level (the notion of personhood isn't precosmic, nor cosmocausal, neither a physics-violating agent), 2. agnosticism ∕ openness to the possibility of the violation of physics at least at one of the stages of the universe (during its lifespan), 3. indifference or simply non-adherence (usually mentioned as irreligion).

Some prefer the word faith as the common hypernym of religion and non-religion (faith as a noun includes religious definitions, and non-religious people don't prefer it).

Some uneducated agostics confuse the term agnosticism with agnostic atheism or agnostic theism. Agnostics aren't less intelligent nor less educated, but most people don't study the official definitions, and especially agnosticism being the most open to metaphysical interpretations; is also open to misinterpretions. If you are atheist but unsure about the metalogical provability that personhood isn't more fundamental than the universe, you might be an agnostic atheist (because it is an ambiguous term, not all accept the exact same definition of that two-worded / bilectic noun).

Some people don't accept the name of their category. Some atheists don't care so much about their non-personocratic worldviews at the cosmological level. They usually don't deny non-personocracy at the cosmological level, but their main focus is to differentiate from other atheists who accept some phenomena which supposedly violate physics. Watch on YouTube Robert Sapolsky professor at Stanford University elaborating on spiritual atheists. In many surveys many atheists are also spiritualists. It doesn't matter if we like the data; it's a fact. Also some agnostics select the wrong box on the surveys due to semantic error. If you are certain about your atheism, but uncertain about the probability of the non-personocentric and non-antiphysics of the universe, and you are presented only with two relative options: atheism and agnosticism, under these specific conditions you should select atheism, because you are sure about its validity, but uncertain of its provability.

Mistakes always happen, and some subjects have a distorted definition of the terminology or don't feel it represents them. We cannot blame the statisticians for everything. Also the statisticians have the right to merge ratios below some value. The maximum permissible merger is 2%, and of course we can be as inclusive and analytical as we want below the threshold of 2% (no excuse to merge non-personocentrists/atheists with the cohort of the open to the possibility of personocracy and anti-physics at the cosmological level/agnostics).

Few people make mistakes or don't accept the jargon. That's no excuse to merge non-religious opinions especially beyond the 2%.

Here two main things were mentioned (and two extras)
 * 1) religion isn't the hypernym of itself (social bias: there are the "group A" people, and the "anti-As" instead of the A, B, C, D ... peoples. The group A is the socially acceptable, and anything different is merged as non-A. Superior vs non-superior. Same as the title which is not the ultimate hypernym, vs opposite to the title. Correct vs non-correct. Strictly focusing on group A, and defining group B, C, D as anti-A. Analyzing methodically all the components of precious A, but merging as junk anything opposed to the biased and non inclusive group A which is also or title of correctness being the only possible sane option and template of normal thought); instead faith or metalogical worldview (the noun faith is a hypernym of metaphysical worldview being more generic, and it includes religious definitions something most non-religious don't appreciate)
 * 2) if we separately mention the religions, we shouldn't then merge non-personocracy at the cosmological level with the openness to the possibility of personocracy at the cosmological level (atheism vs agnosticism) and also with the religious indifference
 * 3) there's no excuse of ideological mergers beyond the 2% because all humans are of equal value; also the non-religious aren't in battle against the religious urged to present a large number
 * 4) some individuals make mistakes during the survey or simply don't accept the terminology; the fact that some people don't cooperate properly isn't an excuse to merge all or some of the non-religious.
 * 5) more people cooperate when are respected by better planned surveys
 * 6) those who don't cooperate aren't that many to provide excuse for an even bigger distortion of specifics by the statisticians themselves — Preceding unsigned comment added by 2A02:587:411F:700:2D7F:C691:25A0:A9F1 (talk)