Home » Blog » Geekery » Counting International Connections on Facebook

Counting International Connections on Facebook

My friend Onnik Krikorian has become a Facebook evangelist. Onnik, a Brit of Armenian descent, living in Armenia, is the Global Voices editor for the Caucuses, which means he’s responsible for rounding up blogs from Armenia, Georgia, Azerbaijan as well as parts of Turkey and Russia. This task is seriously complicated by the long-term tensions in the region. Armenia and Azerbaijan are partisans in a “frozen” conflict – the Nagorno-Karabakh war, which lasted from 1988 – 1994, and remains largely unresolved.

It’s taken Onnik years to build up relationships with bloggers in Azerbaijan, relationships he needs to accurately cover the region. Azeri bloggers are often suspicious of his motives for connecting and wonder whether he’ll cover their thinking and writing fairly. But Onnik tells me that Facebook has emerged as a key space where Azeri and Armenians can interact. “There are no neutral spaces in the real world where we can get to know each other. Facebook provides that space online, and it’s allowing friendships to form that probably couldn’t happen in the physical world.” (Onnik documents some of the conversations taking place between Azeri and Armenian bloggers in a recent post on Global Voices.)

Picture 1
Graph from the front page of peace.facebook.com

Onnik was talking about his love of Facebook at an event hosted by the US Institute for Peace, where I and colleagues at George Washington University and Columbia were presenting research we’d carried out on the use of social media in conflict situations. Onnik’s hopes for Facebook as a platform for peace were echoed by Adam Conner of Facebook, who showed the company’s new site, Peace on Facebook. The site documents friendships formed between people usually separated by geography, religion or politics. Some of the statistics seem clearly like good news – 29,651 friendships between Indians and Pakistanis per day. Others are rather dispiriting – 974 Muslim/Jewish connections in the past 24 hours.

I’m a data junkie, and there’s little more frustrating to me than an incomplete data set. Basically, by showing us a very small portion of the nation to nation social graph, Facebook is hinting that the whole graph is available: not just how many friendships Indian Facebook users form with Pakistani users, but how many they form with Americans, Canadians, Chinese, other Indians, etc. Obviously, this is info I’m interested in – I’ve been building a critique that argues that usage of social networking tools to build connections between people in the same country vastly outpaces use of these tools to cross national, cultural and religious borders.

Without the whole data set, it’s hard to know whether these numbers are encouraging or not. Are 29,651 Indian/Pakistani connections a lot? Or very few, in proportion to how many connections Indians and Pakistanis make on Facebook in total? In other words, we’ve got the numerator, but not the denominator – if we had a picture of how many connections Indians and Pakistanis make per day, we might have a better sense for whether this is an encouraging or discouraging number.

I made a first pass at this question this morning, using data I was able to obtain online. Facebook tells us that the average user has 130 friends – a number that might be out of date, as the same statistics page lists “over 400 million users”, not the half billion currently being celebrated in the media. (Ideally, we’d like to know how many new friends are added per day so we can compare apples to apples, but you got to war with the data you have…)

We also need a sense for how many Facebook users there are per country. Here, we turn to Nick Burcher who publishes tables of Facebook users per country on a regular basis. Nick tells readers that the data is from Facebook, and the Guardian appears to trust his accounts enough to feature those stats on their technology blog. They are, alas, incomplete – Burcher published stats for the 30 countries with the largest number of Facebook users, and revealed a few more countries in the comments thread on the post.

Because we don’t have data for Pakistan, we can’t answer the India/Pakistan question. But we can offer some analysis for Israel/Palestine and Greece/Turkey.

Facebook for Peace tells us that there are 15,747 connections between Israelis and Palestinians for the past 24 hours. The term “connection” is not clearly defined on the site – it’s not clear whether a reciprocated friendship is 1 connection or 2 – because I’m going to count the number of Israeli friends and Palestinian friends, it makes sense to count a reciprocal friendship as two connections. (If Facebook is counting differently than I am, my numbers are going to be half what they should be.)

3,006,460 Israelis are Facebook users… a pretty remarkable number, as it represents 39.92% of the total population of the nation and roughly 57% of the country’s 5.3 million internet users. There are very few Palestinian internet users – 84,240, or 2.24% of the population… This mostly reflects how few Palestinians are online, as Facebook is used by 21% of Palestine’s 400,000 internet users.

At 3,090,700 Palestinian and Israeli Facebook users, we should see almost 402 million friendships involving an Israeli or a Palestinian. If we extrapolate from 15,747 friendships a day to 5.7 million a year, we’re looking at Israeli/Palestinian friendships representing 1.43% of friendships in the Israeli/Palestinian space… with all sorts of caveats. (The biggest is that the use of a year-long interval to calculate total friendships is totally arbitrary and probably not supportable. If you’ve got better data or a suggestion for a better estimation method, please don’t hesitate to speak up.)

We get very different results from looking at Greece and Turkey. 2,838,700 Greeks are Facebook members (25.11% of the national population), while 22,552,540 Turks (31.08% of the population) are. That’s roughly 3.3 billion friendships projected, and our year-long approximation finds us just over 4 million Greek/Turkish connections. That suggests that only 0.12% of friendships in the pool are Turkish/Greek friendships.

What explains the disparity between these numbers? While there’s certainly a long history of tension between Greece and Turkey, the last major military confrontation between the nations ended in 1922. Israel and Palestine, on the other hand, are involved with an active conflict and Israel’s recent incursion into Gaza ended a few months ago. What gives?

It’s possible that the numerous efforts designed to build friendship between Israeli and Palestinian youth are having an impact, much as Onnik’s work in Armenia and Azerbaijan is showing positive results. But there’s another possibility – 20% of the Israeli population are Arab citizen of Israel, and the majority of this set is of Palestinian origin. It’s certainly possible that the high percentage of Israeli/Palestinian friendship includes a large set of friendships between people of Palestinian origin in Israel and Palestinians… indeed, given the difficulty for both populations in meeting in physical space, we’d expect to see increased use of the internet as a meeting space to compensate for the difficulties of meeting in the physical world. This could be a factor in explaining India/Pakistan friendships as well, as well as Albanian/Serbian friendships, as the emergence of new nations through partition and conflict left groups united by cultures, separated by borders.

My goal in this post isn’t to belittle the power of Facebook for providing a border-transcending space where friendships can be built – Onnik’s story makes it clear that Facebook is a real and powerful tool for good, at least in the Armenian/Azeri space. But I continue to think that we overestimate how many of our online contacts cross borders and underestimate how often these tools are used to reinforce local friendships. I’d invite friends at Facebook to correct my numbers or my math… and mention that we could do a much better job of answering these questions if Facebook would release a data set that shows us all the cross-national connections made on the service.


Ross Perez has created some great interactive maps that visualize the adoption of Facebook around the world, using Burcher’s data – worth your time.

19 thoughts on “Counting International Connections on Facebook”

  1. Thank you, Ethan, for an interesting post. I enjoy Facebook. One of its primary uses for me is keeping in touch with people I met at different points in my life, who are spread all over the world. This is to say, I really appreciate the fact that Facebook allows this kind of cross-national communication. At the same time, I think we should be very careful celebrating Facebook as a peace-building platform. I think there is a very long way to go from this communication enabling function that Facebook allows to actual meaningful communication that promotes mutual understanding and tolerance. At the end of the day, Facebook is a platform. It can be used to promote something big, positive, and complex as peace, trust, and tolerance. It can be also used for much more mundane and even negative things within or across nations.

    I think what your post is getting at is only the tip of the iceberg – these cross national links are rather marginal in the grand scheme of things. I would argue that an even more important question is the quality of those links. What does it mean that an Indian has a Pakistani on their friends list on Facebook or what does it mean that a Palestinian has an Israeli friend? I am not saying that this is absolutely meaningless or in any way negative, but I am also not sure that the fact that there is a link on Facebook necessarily has any influence on promoting peace or even says much about the nature of the relationship between these people.

    I think that counting the links can only get us so far; i believe that surveying people, who are connected about what those links mean to them can shed more light on the peace-building-enabling qualities of Facebook. This could be a really interesting project.

    Just my two cents :)

  2. there are too many caveats to compare one’s international network vs domestic counterpart, e.g., one just has much less chance to meet people abroad, and language issue, etc.

    i would suggest a more useful parameter for comparison. e.g. palestanian-israel connection vs palestanian egyption or palestanian iranian.
    or turkey greece vs turkey macedonia or turkey syrian

    such comparisons makes the base population of palestine or turkey irrelavent, but that of the counter-party still matters. so the ratios above should probably be normalized by dividing the israeli and iranian base facebook users respectively.

  3. Hi Ethan,

    Thanks very much for this piece, and not putting me down too much. Well, not at all, actually. :)

    However, just to say that I recognize the limitations you mention, but for now at least it’s a useful tool. For example, even as a British citizen I can not visit Azerbaijan, nor they Armenia, or Armenians Azerbaijan either.

    The only place we can meet is third countries such as Georgia. We also can’t ring each other as either telephone lines are blocked or monitored. But, with Facebook I can keep in contact combined with other tools such as Skype and email.

    Facebook, though, really can give an insight into someone — how they think, who they are etc. And I can really say that I’ve made some very good friends across the border as a result.

    Indeed, there are cases where I got to know people mainly through interaction on Facebook before finally meeting them in person. Strange as it might sound, it was if we knew each other all along.

    As a journalist too, it can also build up trust, and at a time when Armenian and Azerbaijani journalists do not communicate with each other to check facts, get opinions from the other side, I can… and do. And so do they.

    Back to Peace.Facebook.com, yes, not an incredibly useful dataset for you, but it is for me when the general perception is that Armenians and Azerbaijanis should and do not openly communicate with each other.

    But, the growing number of Armenian-Azerbaijani connections on Facebook is changing that perception and encouraging others to do the same. This is an unprecedented development.

    Yes, the numbers are in the minority, and Facebook is also used for perpetuating the conflict, but even this was something unthinkable a little a year and a half ago. What we now need to do is to monitor that trend even if only in terms of numbers.

    Now, of course, my elation is as this communication is in its infancy. In that regard I therefore appreciate your more critical look, and also concerns such as those voiced by Jillian, because one day they might be of significance and importance here too.

    For now, though, all I can say is this. If in June 2008 I knew only about 3 Azerbaijanis after meeting them in Georgia, I am now in near daily contact with dozens and dozens more thanks to Facebook.

    Basically, while it might be difficult for an outside to imagine, even the very fact of an Azerbaijani commenting on the Facebook page of an Armenian, or an Armenian commenting on an Azerbaijanis, is quite a revolution.

    This is particularly true when you consider the general environment in the region is still one of perpetuating ethnic differences, conflict and hatred for the “enemy” often at a national level, but also in the mainstream media.

    Ultimately, I think the general consensus at the USIP event, and it was one I also voiced, is that Facebook like many other things is simply a tool and it depends on how you use it. A bit like everything else, really — from the pen and paper to the printing press or personal computer.

    For me and many others who desperately wanted to cross the closed border but couldn’t, however, the tool has opened up new avenues for communication previously unavailable.

  4. Oh, and btw: I don’t doubt your point about Facebook being used more for local connections than foreign ones. However, for me that is not the point. Instead, I can make some very valuable personal and professional connections that would otherwise be totally impossible.

    So, basically, in the context of Armenian-Azerbaijani relations it is a very valuable tool indeed for those who want to use it for that purpose. My only concern would be one I also raised at USIP. That is, personal security and privacy issues which might arise in the future.

  5. Pingback: Global Voices in English » Armenia-Azerbaijan: Peace on Facebook

  6. Another btw: Okay, so I’m not interested in as much data as you are, for now at least, although I’m sure there’s plenty of interest (just that we’re not quite there yet in this region). However, that huge dip in India-Pakistan relations for several days in the peace.facebook.com is intriguing.

    Not sure if there’s any particular reason for it, but if there was, then even that graph is incredibly interesting. Any ideas?

  7. What does it mean that an Indian has a Pakistani on their friends list on Facebook or what does it mean that a Palestinian has an Israeli friend?

    Dima, I think every conflict has it’s own dimensions, but it might sound strange to you, but the very idea that an Armenian could have an Azerbaijani on their Facebook page, or an Azerbaijani could have an Armenian one, is still unthinkable for most in both countries.

    And, to be honest, I never could have imagined that the work myself and another Azerbaijani blogger did to do so openly and change that perception could have led to more. Yes, still by far a minority even among that minority that use Facebook, but still unprecedented nonetheless.

    As an example, a recent U.S. State Department project to bring Armenians and Azerbaijanis together failed to get them both openly seen to be in “friendship” on Facebook. In part this is because many projects designed to bring more open individuals from either side get the wrong people who don’t believe in that.

    Now, Facebook is an excellent way to select potential participants for such projects — not just in terms of what people post, but also in what connections are listed. For example, if either side already has contact with a few people otherwise considered to be “the enemy.”

    And not just that. I think that someone who has expanded their FB connections outside their region just within their own country also has more potential for cultural understanding and tolerance. Yes, an assumption, but so far in this limited context, shown to be true.

    Incidentally, and as a sign of how Facebook has allowed me to reach out and to make many worthwhile personal and professional connections in Azerbaijan, unable to do so through traditional methods, NGOs working in the area of peace-building and cross-border initiatives now approach me and literally beg me for suggestions.

    So, in the case of Armenia-Azerbaijan connections, for now at least, yes, it does make a difference…

  8. Pingback: Armenia & the South Caucasus | The Caucasian Knot » Facebook: Connecting Armenians & Azerbaijanis online

  9. Friendship on facebook is such a loose term, I’m sure many of the friendships forged on the site are superficial,& have limited access to each other’s profiles. I think the amount of activity & correspondence between the two opposing parties should be examined rather than the number of friendships are made.

  10. Pingback: pligg.com

  11. hfm, sure, and one can say the same about most ‘friendships’ in the real world too. Meanwhile, I have some very real personal and professional relationships with people over the border.

    Another interesting point, however, is that society in countries such as Armenia and Azerbaijan is based on clans in all walks of life. This concerns government and opposition linked groups, but also civil society and networks of friends.

    So, while there is also a danger of creating ‘online clans’ it also allows individuals to break an otherwise very unpleasant and unproductive system in place. Indeed, this is also why a number of organizations are looking to new and social media to bring in more transparency into a lot of areas of life.

  12. Pingback: Global Voices in English » Armenia-Azerbaijan: Overcoming negative stereotypes in the South Caucasus

  13. Pingback: Armenia-Azerbaijan: Overcoming negative stereotypes in the South Caucasus :: Elites TV

  14. I’m not sure that Facebook “friendships” are the best indicator of online cultural interaction. I have a United Nations of more than a thousand Facebook “friends” but I only communicate with about 50 of them on a regular basis.

    I am less interested in the critique that “the Internet is less cosmopolitan than we thought it would be” because I don’t personally know anyone who thought it would lead us to form more friendships with other nationals than our own. But what does interest me is how our online communication and interactions compare with other media. This comment, for example, is a rather homogenous interaction. We’re both college educated, white men who work in the same field and read the same magazines and journals. But if I add up all my meaningful online interactions throughout the day, how does that compare to my communication on the phone, in letters, on TV, on the radio, in person?

    I think it would be fascinating for you to keep a “communications diary” for 30 days to compare your international and cross-cultural communication across across different media. I’d be very interested to see what you find out.

  15. David, it’s a great challenge – I need to a) think through the mechanics of doing it, b) get a larger sample set to make the same commitment and c) persuade you to do it as well.

    I suspect my diary would look like this:
    – face to face interactions are extremely homophilous, given where I live. The main dimension on which there’s any diversity is level of education, and even that’s pretty constant in most interactions
    – Like you, I’ve got a UN of Facebook friends and Twitter contacts. But I suspect my interactions are with a much tighter set, and that those contacts are significantly less global
    – Email is my primary space for interpersonal contact. THere are good tools that exist to document who you correspond with – MailTrends, for instance – http://lifehacker.com/379328/analyze-your-email-usage-with-mail-trends – maybe an experiment could start here, comparing behavior and try to incorporate other SNS behavior over time?

    I suspect my news consumption is also embarrasingly domestic. More than the social question you’re asking, this is the problem I’m hoping to document. I’d be interested in shaping an experiment that looked at how social networks brought news to us and what news we chose to pay attention to…

  16. Pingback: #pdlt : L’internet divise-t-il ou rassemble-t-il ? | traffic-internet.net

  17. Pingback: « FutureEmergency.com

  18. Pingback: …My heart’s in Accra » Shortcuts in the social graph

  19. Pingback: Armenia & the South Caucasus | The Caucasian Knot » Social Networks in Armenia-Azerbaijan peacebuilding & cross-border communication

Comments are closed.