A behavioral psychologist interested in animal behavior noticed that

We want evidence that undermines 'best way to keep dog from barking at strangers is to not punish it physically'. We need to get this out of something about 25 dogs that are known to bark at strangers.

A. Some of the owners of the dogs studied did not physically punish the dog when it barked at a stranger.
-> Means at least some barking was done by dogs not punished physically for barking. That said her conclusion is about physical punishment in general. They could have been physically punished previously, or for thing other than barking. So it's not particularly weakening.

B. Some of the dogs studied were never physically punished.
-> A much stronger version of A. Means that some dogs barking at strangers did so despite never being physically punished (not for barking, not for other things, not by a previous owner, etc). This actually calls into question her original sample she used to create her original conclusion since it runs contrary to her previous finding.

C. The owners of some of the dogs studied believe that a dog which barks at strangers is a good watchdog.
-> Interesting one. If some owners want this behavior they could reward it which would create an alternative source of barking other than physical punishment. That said it doesn't weaken the theory that the best way to *avoid* dogs barking at strangers is to avoid physical punishment.

D. Some of the dogs barked at people who were not strangers.
-> Since her conclusion is about what causes barking at strangers this is out of scope.

E. None of the dogs was disciplined by the method of a rolled-up newspaper.
-> This is a weaker version of B. Since her conclusion is based on dogs being physically punished at all this is again insufficient to weaken. They could have been physically punished in different ways than with a newspaper.