Encouraged by Hamlet Au’s article on how Japan is growing its number of SL residents, I thought I should do some statistics as well, just to keep everybody amused in what you can find out about Second Life.
Hamlet has this breakthrough announcement that “Japan Residents Surpass Europe”. What he really wishes to say is that Japan has more residents than any European country, but the title is misleading. As you can see on this spreadsheet, just the European Union has about 40% of all users of Second Life, compared to Japan’s 8%. If you add the rest of the European countries, it’s close to 50%, which is not surprising: Europe has 800 million inhabitants, compared to Japan’s 128 million or the USA’s 300 million. I would be very surprised indeed if there were more Japanese in SL than the whole of Europe put together; in fact, Japan has about a sixth of the population of Europe, and so those ratios make sense. There is also one American for each 2.66 Europeans, so the ratios between the US residents and the European ones are also close to what they should be.
Still, we have some anomalies. Brazil has come on the second place of the overall country ranking, and this is surprising coming from a country which is not yet overall as developed as the US, Europe, or Japan, although in large areas of Brazil there is the same level of development as on the Western world. In fact, Brazil has 186 million inhabitants — just 45% more than Japan — and in spite of a much lower average standard of living, they already have more residents than Japan in SL. Sure, in terms of sheer population, they ought to have 12% of all residents, and they “only” have 8%, but this figure is still amazing!
When figuring out that astonishing fact, I thought that these kinds of statistics would be most unfair for many countries. For instance, China has more inhabitants than Europe and the US put together, so they should theoretically have something close to 75% of the resident population, which is clearly not the case — they have less than 1% of all active users. Which shouldn’t really surprise us; China has much less Internet users, as a percentage of their population; so it’s unfair to judge them in terms of the overall population. In fact, a more fair figure would be comparing how many Internet users in China are using SL.
This encouraged me to get the data from Internet World Stats and make the following nice graph:
You can see now a different picture emerging from the numbers! First, of course, that China barely registers on the graph; still, from the top 19 countries in SL, you can see that the difference between the extremes is not staggering — or, put into other words, you can see that the average for every country is around 0.05% of the country’s population being in SL. China, of course, is way below that, and Mexico and Argentina don’t fare well either. Even Brazil is quite below that mark, while the Western world is pretty much consistent with that “average”, except for the Netherlands.
But when we consider a percentage of all Internet Users, things start to become interesting. Internet access is cheap these days, even for someone in Brazil or China — obviously, not for the poorest sector of these countries. But there is a considerable percentage of the population — 21% in Brazil, 12.3% in China — that can afford access to the Internet. Among those, however, the Chinese are not too interested in SL — while the Brazilians are crazy about it!!
So I believe that the red bar above shows a much more realistic figure of what the penetration of SL is in each country. I tried to get broadband figures for those countries, but, alas, it’s tricky to find consistent data. For instance, for the European Union, the “official” numbers are completely different from the ones published by Internet World Stats — when they publish anything at all about a specific country. This makes things look quite different than they are — since a country might have a very high percentage of Internet users, but only low broadband penetration, and thus the percentage of SL users suffer from that. Australia is a good example: 72% of the population has Internet access, but only 10% are on broadband (compared to the USA’s 23% or so) — or, in absolute numbers, 1.3 million broadband users vs. 69 millions in the US. This means that we should expect only 2000 or so Australian active residents. There are, however, 14,000! This means that the Australians are way more likely to connect to SL than their fellow Americans.
So we see that there are so many different factors that will make a specific country log in more or less to SL. The consulting firm Kzero has a graph showing the propensity of Western European countries to log in to SL. They explain that this comes mostly from the availability of national brands in SL besides cultural reasons. Similar trends related to cultural factors can explain why some of us spend so much time in SL.
Are cultural references so important? It’s hard to say without a serious study. However, I remember SL in 2004 (or the Internet in 1991) — if you didn’t speak English, you’d be out of it. It was expected that you had at least a good fluency in English to have some fun in SL. You might occasionally find someone who did speak some other language, but it was rare — 75% of all users came from the US, and you could safely add another 20% coming from the English-speaking world. The remaining 5% had good writing skills in English to make themselves fully understandable and usually had no issues, no matter what their country of origin.
What happens in 2007?
As you can see, English still dominates by far, thanks to the US, Canada, the UK, Ireland, Australia, New Zealand put together — as well as a myriad of other countries who have English as their primary language. However, 60% of the SL residents don’t speak English as their first language any more. In fact, one in ten only speaks Portuguese (!) and very likely no other language whatsoever.
This is striking difference from the early days of SL. Interestingly enough, the language spread in SL is quite different from what happens in the real world! According to the Wikipedia, the first ten languages in the world are, respectively, Mandarin Chinese, Spanish, English, Arabic, Hindi, Portuguese, Bengali, Russian, Japanese and German. Compare that to Second Life’s English, Portuguese, German, Japanese, French, Italian, Dutch, Spanish, Polish and Mandarin Chinese!
These statistics should, however, be taken with a pinch of salt and a critical eye. I did some corrections on the tables for the largest bilingual countries, most notably Canada or Switzerland, but I had to assume that the distribution of native speakers of each language in those countries is similar to their distribution in SL, which might not be the case. Another issue arises from countries like Belgium, where Dutch, French and German are official languages, but there is no official census, just an estimate that 59% speak the local variety of Dutch and 40% speak French. If these are correct, the above numbers for French are a bit too high and the ones for Dutch are a bit too low; my apologies to the Belgian Dutch-speakers.
And I’m sure I got a few of the countries wrong — apologies in advance if you found a blatant mistake, and rest assured that I didn’t intend to offend anyone, but mistakes happen!
Last but not least, many people might be counted as residing in a specific country, but speaking a completely different language — or speaking the same language in a certain country which is not their country of origin. It’s impossible to gather those statistics correctly. And, last but not least, there are true bilingual people in SL, and it would require to ask them what their “mother language” is. Even if we just counted the languages that are understood by most residents – to the effect that they’re fluent enough in them — I’m pretty sure that the dominant role of English would be much more clearer. In effect, I have no idea on how many SL residents do not speak English at all — although I’m pretty sure it’s not 60%!
I take full responsibility for any mistakes and misinterpretations of the above figures and graphs; feel free to correct me in the Comments section if I got something dramatically wrong.