'Anonymised' data can never be anonymous, says study

The headline is wrong. It depends on the size and range, but more than 3 attributes may not be used in anonymized studies anyway. This study talks about 15 attributes, which of course leads to trivial deanonymization.