As far as I can tell, they're defining "religious" as people claiming to be religious, and "non-religious" as people claiming no religious affiliation. So they're assuming that everyone means the same thing by the word "religion", and that these people's unspoken definition is a useful concept.
Atheists like to talk about the harm done by "religion". But this harm might be done by deeper habits of thinking that will persist in people who identify in surveys as "non-religious".
Atheists like to talk about the harm done by "religion". But this harm might be done by deeper habits of thinking that will persist in people who identify in surveys as "non-religious".