There was a danger that skeptics and opponents would misread those likelihood ratio tests as rejections of an entire class of models, which of course they were not.

Change background