While I question that the tests in the document can be used to classify the languages into effectiveness, the summary of the article suggests to agree with what you are disagreeing with. From the article:

In general, the differences between languages tend to be smaller than the typical differences due to different programmers within the same language.

Which to me means that while some languages may be more suitable than others for certain tasks, their differences are smaller than the differences between programmers. Hence, programmers carry a larger weight on the appropriate choice of programming language than the language itself. Note that the world isn't black and white, and that the features of the language do play a role (as I indicated in another post). It just means the don't play the major role in the decision.