Again with the confusing 'scientists' and 'papers'. 2/3 of the *papers* did not take a position, presumably because what they were researching did not actually cast any light on the question, not it would be absurd for them to 'take a position'. Of the papers in the study whose subjects actually implied one conclusion or other about AGW, 97% implied the conclusion that AGW is occurring.
The study authors were just being properly careful in explaining that they took a large corpus of papers which *might possibly* imply one or the other conclusion about AGW, then found the ones which *actually did*, and compared how many of those implied one conclusion and how many implied the other. The fact that it happens to be 1/3 of the papers they looked at which fit into this group is not particularly interesting, but if that information hadn't been included, _someone_ would've complained about it.