Well, this goes back to my earlier thought about BeDuhn. His project of comparing the NWT vis-a-vis other translations wrt to bias was a very subjective enterprise. He selected which features to compare and his determination of bias was entirely dependent on his own opinion of how certain texts, particularly very controversial ones, are supposed to be understood. His interpretations collectively become the rubric against which bias in modern translations are to be measured. A different person with a different background and training would come up with quite a different stance on the same texts. Since there is room for disagreement and doubt across scholarship wrt many of these controversial passages, I hardly find BeDuhn's interpretations the most reliable standard for assessing bias, particularly in cases where I don't agree with BeDuhn's own assessment. That is why I feel it is more objective to assess bias in the production of the NWT by tracing its unique or uncommon renderings to models already attested in the literature some time prior to the translation work. This shows that the translation choices made in the NWT reflect particular beliefs and interpretations already adopted by the Watchtower Society; they represent a theological Tendenz peculiar to Jehovah's Witnesses.
I was also disappointed by BeDuhn's discussion of specific passages, which struck me as rather superficial. So for instance he says that adding "[other]" in brackets in Colossians 1:15-20 is "accurate" and "implied in the Greek" and thus does not represent an example of bias (p. 87). I hardly think it's as clear as that. His discussion mainly follows the Watchtower justification of the NWT rendering but doesn't address various grammatical issues posed by the text, such as: (1) the logical force of causal hoti (i.e. it is Jesus' role in the creation of "all things" that makes him the "firstborn of all creation", not his own presumed creation), (2) the scope of pas "all" (which would include Jesus in v. 15 but then suddenly exclude him in v. 16-18, such that panta in v. 16-18 does not complete a set that already includes Jesus), and so forth. And BeDuhn supports the NWT's rendering by referring to other texts whether "other" is implied (Luke 11:42, 21:29) without noting the important ways in which these differ from the wording of the hymn in Colossians. So for instance, in Luke 11:42 mint and rue are included in the set of "garden herbs," pan occurs to complete the set. But in Colossians 1:15, if we would treat "firstborn of all creation" as a partitive (as the NWT does, which motivates the adding of "[other]"), pasés would include Jesus within the set of "all creation" and then suddenly, inexplicably, panta excludes him from the set in v. 16-18. Panta in v. 16-18 does not complete a set that already includes Jesus; rather, it has to denote a new set that excludes him. At least those are problems I see in my own reading of the text.