Interesting leak on the author of the study's gameplan prior to conducting the study. Just food for thought. Pay special attention to "Phase 3" of the plan, keeping in mind that 12,272 total papers were evaluated:
So the gameplan included a plan to "process 6000 neutral papers" looking for excerpted text which can be construed as an endorsement, "converting many of them to endorsement papers."
Yep, sounds unbiased to me. I would also add that the leak reveals that Cook et al started with the ideological goal of solidifying a consensus, not with objectively evaluating the papers for endorsement or refutation.