The top data table, comparing AMP vs. non-AMP traffic, that makes up the reason for this investigation seems flawed. They did correctly isolate organic traffic

The top data table, comparing AMP vs. non-AMP traffic, that makes up the reason for this investigation seems flawed. They did correctly isolate organic traffic to blog posts, however the "Trafico NO AMP" group might contain desktop traffic which, in my experience, has a lower bounce rate.