mostly the dispute between reviewers is about how interesting the topic is, not about whether the evidence is convincing

Your observation of little conflict about whether the evidence is convincing could be explained by a consensus about whether it is convincing, but it could also be explained as low priority. That is my experience in math.