At the surface this is a fun visualization and it confirms some vague stereotypes. A closer inspection shows that the appearance of 'cultural regions' is primarily done by the use of a color scheme that is taken from beyond the data plotted. Note that India and Poland are next to each other in the data space. If you ran a clustering you would find the Indo-Polish cultural group to be very hard to split. That said, I suspect that it would be very hard to convince anyone that India is more like Poland that say Bangladesh, or conversely that Poland is more like India than say Hungary. This map is perhaps a fairly arbitrary reduction in dimensionality.

And in the 2017 version, Poland no longer borders India, but is caught with Chile and Argentine (South Poland?), while India now is neighbors with Italy. It's certainly not a very stable mapping; perhaps all it does is amplify the noise.

or, alternatively, Poland and India have seen large cultural evolutions over a short period of time.

It's not only Poland and India. And the attitudes the graph supposedly shows shouldn't change that much in 9 years.

There has been some serious changes there these last 9 years.

Yeah, seems very arbitrary. A multi dimensional clustering may be more useful then an attempt to flatten that with minimal distortion between the clusters would be more useful. But with too many dimensions it becomes nearly impossible....

As so often, these big survey projects lag considerably. At some point in time, these dimensions were most interesting, and it must have been very satisfying to see them explain such a big amount of variance in such a low dimensional space.

I recommend the time series video: https://www.youtube.com/watch?v=ABWYOcru7js

Indeed it seems less a good explanation nowadays, but this is another take away, I guess.

Agreed. Maybe colouring the dots as a function of per capita GDP would've been easy, nice, and more interesting

The word choice for the scales is unfortunate. I would've labeled them "religious to mundane" and "locally to globally concerned." Mostly I have a problem with any definition of "traditional" which has Japan at the least traditional country.

How on earth are these metrics defined even somewhat objectively? How is it justified that countries such as China, India, and Russia occupy only one point despite being stunningly diverse polities? Why are all 'traditional' values rolled into one category? Why are 'survival' and 'self-expression' considered opposite extremes?

In this case it's entirely factor analysis of responses to a particular survey, the World Values Survey [1], which gets periodically administered across a number of countries. The survey data and factor analysis can be done in a fairly "objective" way, in the sense that you do have answers that vary between countries, and you can analyze trends there. Whether the factors pulled out of these particular sets of questions can accurately be labeled with fairly strong terms like "traditionalism" is maybe more controversial. You also have the general issues around self-response data.

[1] https://en.wikipedia.org/wiki/World_Values_Survey

These are well tested social science concepts. Its just hard to communicate in a two word label..

When you and your family and your neighbors, village, tribe are worried about where your next meal is coming from, you don’t have much energy for self-expression. See also Maslow’s hierarchy.

China and Russia have similar politics in that they both have long authoritarian histories, as well as Communist histories. I don’t know much about India, but I’d guess that the average Indian and the average Chinese person have similar concerns: how to get enough money to move their family up from poverty or near poverty and into the middle class.

This is based on a certain positivist view of culture, that tries to quantify on certain cultural dimensions. This doesn't look behind coded appearances. As a specialist for Turkey I would say that it is much more secular-rationalist than Germany. But would never voice it as such, which is a form of respect. For sure in quantitative interviews one would not be able to account for that. On top of that this map's cluster is unscientific because it clusters things that are not comparable (provokes a type error). Confucian: religion, English speaking: language, South Asia: continent. I am a PhD student in intercultural management.

> As a specialist for Turkey I would say that it is much more secular-rationalist than Germany.

Wow, that's surprising, can you expand on that?

I had a Turkish girlfriend for a few years, it felt like I was living in Turkey. (For example, we couldn't hold hands in the street–in Australia!–because Turkish people would see and her family would find out. Being married makes a female suddenly off-limits to male harrassment, but public affection with a boyfriend seems shameful and quite the opposite.) I learnt that e.g. changing your religion in Turkey can get you killed. Then the Islamists took over and it felt like things were getting medieval quickly. But I haven't heard news from there for 5+ years.

Sure, first immigrant communities are always different. Depending on their context of emigration (from where, how did they emigrate) and also the surrounding society. It can be a part of their cultural identity to do as one thinks a Turk would do. Nevertheless that is exactly a part where I would think that a simplification to compare with a global rational cultural dimension is impossible. Marriage is in Turkey much more important, but most Turks would not counter you with Islam, but with a rational discourse, e.g. a woman can get pregnant, she has to secure her hypothetical children. If she gets pregnant, the family would look like they couldn't take care, and so on. It is another story if this is really something that the family should decide. On the other hand if you would visit Turkey, which is heavily urbanized, and especially Istanbul, which is like 20% of Turkey, you would see that a lot of things are not allowed, but done. Because, well, it is not that much not allowed. I would not take those numbers for granted, but it is a funny indicator in the Durex report, where Turkey has a mean number of sexual partners of 5.3 and France just 3.8 [1]. I would also be very careful when to blame Islamists. Turkey is not a religious country. There are a lot of countries who want to construct here a clash of cultures. And apostasy is not something that gets you killed.

[1] http://www.durexnetwork.org/SiteCollectionDocuments/The%20Fa...

Thanks. But you didn't say anything to persuade me that

> As a specialist for Turkey I would say that it is much more secular-rationalist than Germany.

And reading https://en.wikipedia.org/wiki/Religion_in_Turkey doesn't at all give the impression that

> Turkey is not a religious country.

Neither does https://en.wikipedia.org/wiki/Apostasy_in_Islam#Turkey which says

"there are several formal and informal mechanisms in place that make it hard for citizens not to be Muslim. Non-Muslims, especially non-religious people, are discriminated against in a variety of ways. Article 216 of the penal code outlaws insulting religious belief, a de facto blasphemy law obstructing citizens from expressing irreligious views, or views critical of religions."

Brazil would have to neighbor all colors for this map to be true, including protestant Europe. For instance, Uruguay is in a place that seems correct (Uruguay was once a Brazilian province, BTW, and still has close ties).

The question is whether those polled in Brazil are more or less traditional than those polled in Uruguay

Uruguay is a very secular place, should be further up on the graph.

Does anyone know what the dashed red line that roughly encloses the top-left quadrant of the map refers to?

The original publication that the diagram was produced for is available[1] (for purchase) from Cambridge University Press.

[1] - https://www.cambridge.org/core/books/modernization-cultural-...

Ex-communist states. Written there on the map

Thank you; I saw that label but was unclear whether it referred to the red-coloured area, or the area within the (also red) dashed line.

In that case I imagine that the red-coloured area is former Soviet states (?).

The red are is probably the ortodox christian region.

That'd make sense too, and would match the 2017 version of the diagram further down the page.

Azerbaijan and Georgia seem to have very different religious profiles though, which could be a counterpoint.

It's a shame the diagram doesn't have a key associated with it (at least on Wikipedia) :)

The fact that all of the regions on this map have such odd and arbitrary shapes--along with the fact that they don't correspond to obvious clusters in the underlying scatter plot--does not inspire confidence.

This is interesting to consider in the context of Zeihan's "The Accidental Superpower"[1] which says that - geography is destiny, and - the U.S. won the geography lottery.

The idea that geography has some input to the rest of the cultures values, as shown in TFA, sounds plausible.

Whether Zeihan actually has the right of it isn't clear, however.

[1] https://www.amazon.com/Accidental-Superpower-Generation-Amer...

Very interesting, I would love to see level-headed discussion on this, although it might be politically fraught. I'm curious how these two axes were decided, if anyone knows.

Also, it makes me wonder about cultural relativism, since it seems there's this implicit sense of "progress" by the way the map was constructed. It lines up with my values, but is that my Western bias?

It is stats 101, or maybe 201: simply the first 2 principal components in a PCA.

That is acceptable as they accounted for 70% of the explained variance, as explained in the wikipedia entry.

If you want, you can also use the first 3 PC and do a 3d plot on (x,y,z) - it is rarely used as most people have problem visualizing groups with more than 2 dimensions. It could separate countries like Japan, which end up in an interesting position.

To address your second question, if there was any "cultural relativism", it would have been in the question asked.

But when you take the answers to whatever questions, do a PCA and find the PC that explain a large part of the variance, you abstract away any cultural relativism.

You could critique the name used for the axis (it depends on the correlation structure), but not the axis themselves (they comes from mathematical results)

Interesting concept, but seems quite subjective and Eurocentric. Out of the 7 labels given, 3 are Europe-focused (Catholic Europe, Protestant Europe, English speaking). If a Muslim made this map, for example, would it make sense for them to split the world into Sunni, Shia, and Arabic speaking?

Interesting that Cyprus is in South Asia according to the 2017 map. Clearly, the hand-painted clusters are not super insightful or thoughtful.

They should have at least flipped it left-right, and it would look like a badly drawn world map.

But then Western Europe would not be in the top right quadrant, where it belongs.

Those belong to the Emperor... embalmed ones... those that tremble as if they were mad...

“Out of Western world countries, the United States is among the most conservative (as one of the most downwards-located countries), together with highly conservative Catholic countries such as Ireland and Poland.”

Abortion is now legal in Ireland[1a] – it is illegal in Poland still[1b]. (Yes, abortion was only recently legalised in Ireland but that's because it required a constitutional amendment to change the law and those sorts of votes only come round once a generation, attitudes had long since shifted.)

in 2015 Ireland became the first country in the world to approve same-sex marriage by popular vote[2a] – Poland does not legally recognize same-sex unions[2b], either in the form of marriage or civil unions.

Ireland ranks 3rd in the world in the Human Development Index[3], 6th in the world in the Democracy Index[4], and 15th in the Press Freedom Index[5] – Poland ranks (32nd, 57th, and 59th) and the USA ranks (15th, 25th, and 48th)

The view that Ireland is a highly conservative Catholic[6] country is, I would argue, a very outdated view. It's a stereotype that needs to wither on the vine, along with other crude stereotypes of the place. Of course there remain conservative pockets, but the same could be said for any place.

My point being, if the chart (and article) are that off base about a country I know about how can I trust it is correct about other countries I am less familiar with? It all seems hopelessly reductive, almost a stereotyped distortion of reality.

[1a] https://www2.hse.ie/abortion/

[1b] https://foreignpolicy.com/2019/01/08/poland-is-trying-to-mak...

[2a] https://en.wikipedia.org/wiki/Same-sex_marriage_in_the_Repub...

[2b] https://en.wikipedia.org/wiki/Recognition_of_same-sex_unions...

[3] https://en.wikipedia.org/wiki/Human_Development_Index

[4] https://en.wikipedia.org/wiki/Democracy_Index

[5] https://en.wikipedia.org/wiki/Press_Freedom_Index

[6] https://www.cso.ie/en/releasesandpublications/ep/p-cp8iter/p...

( Though 78% self-report as Catholic only 35% attend Church on a weekly basis which I would argue is a truer measure of religiosity. https://en.wikipedia.org/wiki/Religion_in_the_Republic_of_Ir... )

The grouping is arbitrary made by some human or algorithm trying to make a pretty map.

As said in the article, the locations on the graph are based on the answers to the world values survey [0]. I have no idea about how exact they are or biases they might include but by that it seems you guys answer the same even though the playing field of society has changed.

Maybe we will see Ireland start trending towards the upper side if it is redone in another 10 years time?

[0]: https://en.m.wikipedia.org/wiki/World_Values_Survey

I don't think Quebec is going to like that ;)

Where is Israel on the map?

The fact that Europe and it’s colonies (with around a tenth of the worlds population) fill up half of the map hints at some Eurocentrism. Shoving all of Asia and all of Africa into tiny equally sized corners is quite strange.

Are you familiar with the methodology used?

> The fact that Europe and it’s colonies (with around a tenth of the worlds population) fill up half of the map hints at some Eurocentrism

It reads to me like "results of a 2d plot using matrix multiplication do not match my biases, it hints at some Eurocentrism"

But to answe the question, No, it just means there is more variance there in whatever cultural values where measured, then analysed by PCA.

I will try to give you similar counter intuitive examples.

Likewise, Africa has more genetic variance than the rest of the world, even when adjusted by population. It does not mean anything, except that there are more separate groups who did not intermingle much there.

If I remember correctly, Papua New Guinea has the most genetic distance to the rest of the world. It does not mean anything either, except that it's far away to the rest of the world, so had fewer contacts.

> Shoving all of Asia and all of Africa into tiny equally sized corners is quite strange.

No, it means that both 1st and 2nd PC had close values. Maybe if you look at the answers to the questions themselves, there would be little variance.

What is interesting is that we can make cultural clusters while we wouldn't have expected any to remain after PCA.

It shows geography impacts the diffusion of cultural values, which is fascinating - and more so for the exceptions, like Uruguay.