One of the tendencies I see in reviews is that the reviewer focuses on the wrong stuff. For e.g. when reviewing the design document, they focus on the validity of the requirements and they may even stray into commenting on formatting and grammar in the document; when reviewing code, they focus on design; when reviewing test cases, again they focus on reviewing requirements; when reviewing deployment scripts, they focus on the choise of infrastructure. Basically, people tend to pay attention to things that are already past the decision stage while ignoring the matter at hand.