His point is not contradicted by the data. He's questioning the paper's claim that people self-identify using corporate brands, saying instead that nonreligious

His point is not contradicted by the data. He's questioning the paper's claim that people self-identify using corporate brands, saying instead that nonreligious people stick with brands that have worked for them in the past whereas religious people believe whatever the label tells them. Obviously, he's not expressing that claim as true. He's using it to point out that, given imagination and a sufficiently broad brush, this set of data can be interpreted to cast aspersions on any group you wish.