The author of Foreign Dispatches explains this as follows:
For Arabic, Vietnamese, Turkish, Polish and Romanian to figure in the top 5 without their users even numbering amongst the top 10 best represented groups online says very clearly that the intensity of the online interest in the topic of sex is several orders of magnitude larger than elsewhere.This is a good example of faulty reasoning. To see why this is the case, imagine that M out of 100 queries submitted by males are on the topic of "sex", while F out of 100 queries submitted by females are on the same topic. If males are more likely to search on this topic than females, then countries that are more male-dominated (i.e., in which women don't use computers as much) will appear to be more "sex"-interested. For example, if M=5 and F=1, then a country where 60% of Internet users are male, will have an observed interest of 3.4, whereas if 90% of Internet users are male, then the corresponding interest will be 4.6.
Similarly, young people are perhaps more interested in sex than older ones. With similar reasoning, we can see that in countries where old people use the Internet, the apparent interest in sex will be lower than in ones where Internet usage is limited to young people.
Furthermore, the cost of accessing the Internet is not similar across countries. In the First World, nearly everyone can be online, but in poorer countries Internet access is limited to the more affluent sections of society, i.e., working-age males.
And of course, "sex" is an English word, so someone who queries on "sex" must have at least a basic knowledge of English. Once again, the group of people who know some English is not the same as that which does not, biasing the perceived interest in the subject. Not to mention that one's level of English knowledge will dictate their use of the query term "sex" or not: poor English speakers may search for "sex", while more knowledgeable speakers may use one of the more specific terms for what they are searching for.
To summarize, the intensity of querying on a topic (and "sex" is just a convenient example) is a function not only of interest in the topic, but also of the characteristics of the querying population.