The study multiplies the information density by the rate of speech, and you are correct, the Japanese rate of speech is the highest, but it did not compensate for the low information density. This is likely a result of the small number of syllables present in the language (~400 compared to ~7,900 in english).
Of course, as many people here are saying, the analysis is based on written text and reading out loud, not spoken conversations, which would probably require a different analysis.
Of course, as many people here are saying, the analysis is based on written text and reading out loud, not spoken conversations, which would probably require a different analysis.