I think the study is trying to link two quite unrelated aspects through a narrow lens of anonymized "international" friendship links. What constitutes international? Is any link outside of your own country considered international?
To me, this chart seems to highlight countries that have relatively open borders with its neighbors, or countries where immigrants as percentage of population is high because of economic opportunities and such. Consider Saudi Arabia or Oman, which gets a lot of immigrants from all over the Muslim world, which seems to be one of the most "international" countries. This inference seems counter to the fact that many of those countries are not as open to international businesses as say, Singapore or Hong Kong. Nor are they poor countries per se.
I couldn't find anything in the study that shows how friendship is defined within in anonymized data set:
> Facebook provided data on every friendship formed in 2011 in every country in the world at the national aggregate level. These data set included a total of 57,457,192,520 friendships.
So there's a few problems with language here...one of them being...what does "every friendship formed in 2011" mean? The friendship was actually made in 2011? Or that the friendship record existed as of the 2011 snapshot of the database? I'd have to imagine the former but if it's just the latter...then "friendships made by all users in one given year" is massively different than "all friendships of all users".
Second, in 2011, FB's interface was different than it is today. Maybe they've always had this data in the backend...but were users able to have multiple locations listed as part of their user profile in 2011? My hazy recollection is that you were able to list your location...and after the Dec. 2011 roll out of Timeline [1], then you started being much more granular about your life and location, because Facebook wants to be the yearbook of your life or something.
If we assume that an international friendship is made by two users who have locations in two different countries...OK...well, anecdotally speaking, I went to a Midwest college..we didn't have FB at the time but I imagine if we did, most/many students would have put the university as their current location/network. Even though the number of international friends I had substantially increased, if the data isn't captured at a granular level, it just looks like I made friends with a whole bunch more midwesterners.
I upvoted this. Just because we need more reminders of how awful published research can be...and yet, I'll be the first to admit that upon seeing the headline, I thought: "Oh yeah, that makes sense, my rich friend Bob is the biggest xenophobe I know", mixed in with "Fucking facebook and their ice-bucket-dumping morons" type of faux-highbrow sentiment.
Just like headlines are a standardized way of trumpeting a claim, it'd be nice to have a semi-standardized format for quantifying the scope and details of methodology.
edit: The more interesting part was the first "micro" study, in which people consented to handing over their FB network. But reading the actual details gives me additional doubt, though I can't say I'm very knowledgeable about what's been done so far to correct for MTurk quality issues in academic studies:
> We recruited 1069 individuals from Amazon's Mechanical Turk who lived in the United States to participate in the study in exchange for $1.00....
> [in the conclusion]... Second, our sample in Study 1 consisted of MTurk users which are not representative of the general US population. While MTurk users tend to be as good—if not better—than typical undergraduate samples (Buhrmester, Kwang, & Gosling, 2011), future work should endeavor to recruit a nationally representative sample to further corroborate the results.
edit2: I'm kind of interested in the type of person who, over MTurk or not, would hand over their Facebook network for a dollar? I mean I'm sure lots of people would do that, but still...
And the study they reference in terms of claiming that MTurk users are more representative than the undergrads that studies usually pick from...The study was done in 2011. I know MTurk has been around for awhile, but has much changed about its demographic? For example, I'd be very suspicious of a study based on Uber that depended on driver analysis from even as late as 2014, given the kind of turnover that has happened.
I'm curious how they took into consideration the fact that recent immigrants can often fall into the lower social classes while having many connections to family and friends in their home country. Edit: It seems like they didn't even look at the social class of the people they interviewed. They just focused on the GDP of their country which doesn't make any sense imo.
This seems like really bad research. Correlating international friendships to a countries GDP would only make sense if every person in the world was on facebook.
Yea, India with a 1.2 billion people, and one of the lowest GDP per capita in the world, is at the same "less international" level as the US. There is no correlation. They could've drawn other conclusions out of this, like something about the attitudes about the people living in that country.
Because it might as well be relative, some one flipping burgers in Kentucky probably earns more per week than a family in Bangladesh does in a month (if not a year, considering their poverty line is 2$ a day) but they won't be considered high social class in the US.
From the study: "[I]t is possible that migration flows could be a potential mechanism behind this effect. ... However, we do note that we found evidence against a migration explanation in Study 2."
But Study 2 appears to have actually looked at the relationship between GDP per capita (on a country-level basis) and number of international friends per capita (again on a country-level basis). GDP per capita between countries seems like a poor proxy for social class between people within the same society. Even if people in rich countries have fewer international friends than do people in poor countries, does that really indicate my rich neighbor is likely to have fewer international friends than does my poor neighbor?
Edit: looks like zopppo already pointed out this issue before I refreshed.
[+] [-] badusername|10 years ago|reply
To me, this chart seems to highlight countries that have relatively open borders with its neighbors, or countries where immigrants as percentage of population is high because of economic opportunities and such. Consider Saudi Arabia or Oman, which gets a lot of immigrants from all over the Muslim world, which seems to be one of the most "international" countries. This inference seems counter to the fact that many of those countries are not as open to international businesses as say, Singapore or Hong Kong. Nor are they poor countries per se.
[+] [-] danso|10 years ago|reply
> Facebook provided data on every friendship formed in 2011 in every country in the world at the national aggregate level. These data set included a total of 57,457,192,520 friendships.
So there's a few problems with language here...one of them being...what does "every friendship formed in 2011" mean? The friendship was actually made in 2011? Or that the friendship record existed as of the 2011 snapshot of the database? I'd have to imagine the former but if it's just the latter...then "friendships made by all users in one given year" is massively different than "all friendships of all users".
Second, in 2011, FB's interface was different than it is today. Maybe they've always had this data in the backend...but were users able to have multiple locations listed as part of their user profile in 2011? My hazy recollection is that you were able to list your location...and after the Dec. 2011 roll out of Timeline [1], then you started being much more granular about your life and location, because Facebook wants to be the yearbook of your life or something.
If we assume that an international friendship is made by two users who have locations in two different countries...OK...well, anecdotally speaking, I went to a Midwest college..we didn't have FB at the time but I imagine if we did, most/many students would have put the university as their current location/network. Even though the number of international friends I had substantially increased, if the data isn't captured at a granular level, it just looks like I made friends with a whole bunch more midwesterners.
[1] https://www.facebook.com/notes/facebook/timeline-now-availab...
[+] [-] danso|10 years ago|reply
Just like headlines are a standardized way of trumpeting a claim, it'd be nice to have a semi-standardized format for quantifying the scope and details of methodology.
edit: The more interesting part was the first "micro" study, in which people consented to handing over their FB network. But reading the actual details gives me additional doubt, though I can't say I'm very knowledgeable about what's been done so far to correct for MTurk quality issues in academic studies:
http://www.sciencedirect.com/science/article/pii/S0191886915...
> We recruited 1069 individuals from Amazon's Mechanical Turk who lived in the United States to participate in the study in exchange for $1.00....
> [in the conclusion]... Second, our sample in Study 1 consisted of MTurk users which are not representative of the general US population. While MTurk users tend to be as good—if not better—than typical undergraduate samples (Buhrmester, Kwang, & Gosling, 2011), future work should endeavor to recruit a nationally representative sample to further corroborate the results.
edit2: I'm kind of interested in the type of person who, over MTurk or not, would hand over their Facebook network for a dollar? I mean I'm sure lots of people would do that, but still...
And the study they reference in terms of claiming that MTurk users are more representative than the undergrads that studies usually pick from...The study was done in 2011. I know MTurk has been around for awhile, but has much changed about its demographic? For example, I'd be very suspicious of a study based on Uber that depended on driver analysis from even as late as 2014, given the kind of turnover that has happened.
[+] [-] jonaf|10 years ago|reply
[+] [-] cortesoft|10 years ago|reply
[+] [-] zopppo|10 years ago|reply
[+] [-] werber|10 years ago|reply
[+] [-] stretchwithme|10 years ago|reply
[+] [-] winter_blue|10 years ago|reply
[+] [-] manigandham|10 years ago|reply
Why not just say people from richer countries have fewer international friends then?
Seems like poorly thought out "research".
[+] [-] dogma1138|10 years ago|reply
[+] [-] bpodgursky|10 years ago|reply
[+] [-] andersriutta|10 years ago|reply
But Study 2 appears to have actually looked at the relationship between GDP per capita (on a country-level basis) and number of international friends per capita (again on a country-level basis). GDP per capita between countries seems like a poor proxy for social class between people within the same society. Even if people in rich countries have fewer international friends than do people in poor countries, does that really indicate my rich neighbor is likely to have fewer international friends than does my poor neighbor?
Edit: looks like zopppo already pointed out this issue before I refreshed.