Comment by p-e-w | Hacker News Reader

Comment by p-e-w | Hacker News Reader

Parent Full thread

> It has been shown many times that integration reduces bias

Strawman. Integration isn't the same thing as diversity, and doesn't automatically follow from it either. Different kinds of people living in the same locality doesn't mean they form an "integrated society". Almost every large city is a testament to that. The idea that city people are more empathetic and understanding than villagers because they live in a more diverse environment flies in the face of everyday experience.

> Moreover, your arguments taken to the logical conclusion have been the basis of some of history's worst atrocities

Appeal to consequences, and thus irrelevant to whether my arguments are correct or not. Not that I accept the ridiculous assertion that my arguments are in any way comparable to the ones you are linking them to.

> and I would advise you to do more research before seeding this particular thought pool.

Sorry, but I'm not swayed by such pretentious moralizing. In fact, the suppression of inquiry that you are attempting here seems much more in line with "this particular thought pool" than what I wrote.

> Strawman. Integration isn't the same thing as diversity, and doesn't automatically follow from it either. Different kinds of people living in the same locality doesn't mean they form an "integrated society". Almost every large city is a testament to that. The idea that city people are more empathetic and understanding than villagers because they live in a more diverse environment flies in the face of everyday experience.

You are calling for empiricism but mixing your arguments with anecdotes and "everyday experience". The empirical argument looks at questions like integrated communities as a form of diversity, and finds that it has socially beneficial consequences -- exactly what kind of diversity are you referring to which is distinguishable from integration? Do elaborate so that we can examine that in an empirical lens.

> Appeal to consequences, and thus irrelevant to whether my arguments are correct or not. Not that I accept the ridiculous assertion that my arguments are in any way comparable to the ones you are linking them to.

Let me draw the connection for you: do you agree or disagree that your initial line of reasoning would support the conclusion that ethnostates are morally neutral? When this reasoning is placed in a historical context, are you blind to the means by which ethnostates have historically arisen?

> You are calling for empiricism but mixing your arguments with anecdotes and "everyday experience"

Well… duh? It would have been distinctly anti-empirical to ignore those.

> exactly what kind of diversity are you referring to which is distinguishable from integration?

The kind you can find in most cities today. Where your neighbors come from many cultural backgrounds, but you don't even know their names, and they don't know yours.

Such societies are diverse statistically, but they are not integrated. In fact, it can reasonably be questioned whether they are "societies" at all.

> do you agree or disagree that your initial line of reasoning would support the conclusion that ethnostates are morally neutral?

"Ethnostates" are states dominated by a particular ethnic group. Which by definition includes states that are ethnically uniform. According to [1], many states existing today have essentially no ethnic diversity, including Japan, South Korea, Portugal, and Norway. Meanwhile, the most ethnically diverse countries include such luminaries as Liberia, Sierra Leone, and Ivory Coast.

In fact, if you look at the accompanying map[2], you will immediately notice a strong inverse correlation between diversity and living standard. Assuming the underlying analysis of diversity level is accurate, this map itself is a condemnation of the entire ideology surrounding diversity.

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

[2] https://commons.wikimedia.org/wiki/File:List_of_countries_ra...

I question your methodology, but let's take it at face value.

>In fact, if you look at the accompanying map[2], you will immediately notice a strong inverse correlation between diversity and living standard. Assuming the underlying analysis of diversity level is accurate, this map itself is a condemnation of the entire ideology surrounding diversity.

As a matter of fact, this correlation is much weaker than you portray it to be, and the number of outliers in this hypothesis call it deeply into question. Let's assume GDP (factoring in purchasing power parity) per capita as a proxy for standard of living. Consult this map and compare it to yours:

https://en.wikipedia.org/wiki/File:Map_of_countries_by_GDP_(...

You see many counter-examples to your thesis. Africa tends to have poor correlation between ethnic diversity and standard of living. Look also at East Asia: the more diverse Indonesia and Papua New Guinea are doing better economically than their homogeneous northern neighbor in the Philippines, in this region too is diversity poorly correlated with economic status. Also true of Europe -- just look around the map and you'll find counterexamples everywhere. Look at Switzerland -- it's at once one of the wealthiest and most diverse European countries according to your map.

Your example of Japan as an ethnostate is also an interesting case study. I agree that it resembles an ethnostate. However, it's also going through a demographic crisis which is in large part attributable to its status as an ethnostate: an increasingly wealthy and aging ethnically Japanese population is an impending disaster in the face of Japan's historical (and contemporary) strict policies on immigration and integration.

Speaking of demographic crisis, compare your map to this one:

https://upload.wikimedia.org/wikipedia/commons/b/b7/2017_wor...

I see some correlations there.

The standard metric to compare to would probably be the Human Development Index[1]. Which paints roughly, but not quite, the same picture as GDP.

But regardless of which development metric you choose, I'm sure you agree that the (inverse) correlation between development and diversity is obvious. Calculating the Spearman coefficient for the two datasets would be an interesting exercise, which I regret I don't have time for at the moment.

This leaves four possibilities:

1. High development reduces diversity (very unlikely).

2. High diversity impairs development.

3. High diversity and low development have a common underlying cause.

4. The correlation is coincidental.

Even in Africa, the correlation is apparent. The countries bordering the Mediterranean all have low diversity and high development, compared to almost all other African countries.

[1] https://commons.wikimedia.org/wiki/File:Countries_by_Human_D...

>Even in Africa, the correlation is apparent. The countries bordering the Mediterranean all have low diversity and high development, compared to almost all other African countries.

Okay, let's examine this assumption, comparing your first map (diversity) with the map you reference in this comment (HDI). First, Africa:

Consider southern Africa. South Africa is the most ethnically diverse country in this region; it also has the highest HDI. The two more homogeneous countries north of it, Botswana and Zimbabwe, are not outliers in their neighborhood when it comes to the HDI score. Let's look to Western Africa: along the Ivory Coast we see a number of relatively high-diversity countries, but their diversity scores relative to each other are not well-correlated with their HDI scores. In the Arab World, Northern Africa through to the Middle East, we also see poor correlation between diversity and HDI: Algeria, Libya, and Egypt are relatively wealthy and homogeneous, but Saudi Arabia is wealthier than any of them and has higher diversity; Yemen is both one of the poorest countries and most homogeneous in the region.

In Middle Asia, compare China to Kazakhstan: both are high HDI (Kazakhstan scores higher), but China is more homogeneous and Kazakhstan is more diverse. In the Mediterranean, the correlation is somewhat stronger but still has numerous counterpoints; Bosnia and Herzegovina is the most diverse country in the region but it's HDI scores comparably to its neighbors to the south and east. And, again, moving slightly north, Switzerland is the highest scoring country on the HDI in Europe and is also among the most diverse.

Your methodology is very weak and your conclusions make no sense.

Follow-up after an afternoon of reading: I dug into the source of the data you presented here in references [1] and [2], which is mainly from James Fearon "Ethnic and Cultural Diversity by Country" (2003). In this he actually explicitly cautions against making the kind of analysis and correlation that you posit here:

> First, it cannot be assumed, without argument, that ethnic distinctions are wholly exogenous to other political, economic, and social variables of interest. For example, poor economic performance could exacerbate distributional struggles, causing people to see and act along lines of ethnic division that were formerly considered unimportant. By contrast, robust economic growth might lead to the downplaying of ethnic divisions and a greater emphasis on national identity. If Botswana seems more ethnically homogeneous than Somalia does at this point, it may be that this is in part a result rather than a cause of economic growth.

The paper also goes into great detail on the problems inherent in trying to categorize ethnic groups; these two factors combined are why I object to your methodology here and I would recommend that you read the paper for yourself. In other words, any conclusions drawn or inferences made based on this data without controlling for innumerable complicating factors is likely to be a case of garbage in, garbage out.

Of course, ethnic diversity is just one of many forms of diversity, which further removes this data from your argument.

The passage you quoted is the standard "cover your ass" disclaimer that you can find in almost any social science publication today, where the authors basically state that their results become invalid the moment they are used to draw potentially controversial conclusions.

I've long learned to just ignore such statements. Either the data can be analyzed, or it's worthless. But which of the two is the case doesn't depend on the conclusion being drawn. In the Soviet Union, it was common for scientific articles to emphasize how the authors used "Marxist-Leninist science" to achieve their breakthroughs. I view those disclaimers as a modern equivalent to this phenomenon, unsurprising considering that scientific freedom during the past 20 years has been lower than at almost any point since the 18th century or so.

The passage I quoted is one paragraph out of seven pages spent addressing the challenges in defining and analyzing ethnic groups and drawing conclusions from that data, not to mention numerous caveats and counter-examples provided throughout the text intermixed with his methodology and data analyses, or the extra text re-enforcing these problems in the conclusion -- not some brief cover-your-ass passage but rather one quote from a lengthy and detailed academic analysis of the problems inherent in this data. It would be obvious had you read the paper -- but you selectively pick and choose whatever information and "data" that supports your racist conclusions and don't examine it any further in case it would turn out not to, a trend seen in every source you have cited in this discussion.

Reply on news.ycombinator.com