this study has a grand total sample size of 170 (that's 10 speakers per language), it also measures syllable rate (i.e. syllables/second times diversity of syllables) as a proxy for information. this is pretty far from how people will intuitively conceptualize "information".
i would take the results too seriously.