Studies about international English proficiency

This appeared a few months ago in the `Johnson' blog at The Economist, but I only stumbled across it recently. It seems that a company called EF Education First, which teaches English along with other subjects, has published a large-scale international study of English language proficiency around the world. `Johnson' sums up the study and findings well:

"EVERYONE knows the stereotypes about foreigners speaking English: Scandinavians are shockingly fluent, while the Japanese lag despite years and billions of yen spent trying. Now a big new study confirms some of those stereotypes. But it holds some surprises as well.

EF Education First, an English-teaching company, compiled the biggest ever internationally comparable sample of English learners: some 2m people took identical tests online in 44 countries. The top five performers were Norway, the Netherlands, Denmark, Sweden and Finland. The bottom five were Panama, Colombia, Thailand, Turkey and Kazakhstan. Among regions, Latin America fared worst. (No African country had enough takers to make the lists’s threshold for the minimum number of participants.)"

While it's great to have such a large number of participants, the problem with this study's findings is similar to claims about proficiency based on TOEFL scores: the test takers are not likely to be representative of the population as a whole. In the case of the TOEFl tests, Japanese people will take it not quite on a lark, but certainly without serious preparation, just to gauge their English level, or even as a requirement for a job application. In countries that don't have such a test-philic culture, as much disposable income, or as easy access to the test, it's much more likely that people will wait until they feel confident before taking the test.

In the EF study, the test takers are also self-selected. As `Johnson' points out, "they were by definition connected to the internet and interested in testing their English; they will also be younger and more urban than the population at large." So, at least, the study may be comparing similar groups in different countries, but it's hard to know what to make of that.