I didn't check the Perl Best Practices book, but I would suspect that it says nothing about that. Because depending on the situation one or the other might be better.

Usually a ref is better for really big arrays or hashes or when you want to get a specific array or hash and not a copy. In most other cases the regular array/hash is just fine as it allows for simpler syntax

My personal practice is to always return refs because it lets me concentrate on more important things than trying to remember which functions return refs and which return arrays/hashes. Always returning arrays/hashes would have this same benefit, of course, but doesn't allow for returning multiple array/hash results.

And then there's the option of using wantarray to return refs in scalar context and arrays/hashes in list context. In theory, it seems like it could provide the best of both worlds by letting the caller get the result in whatever form they want, but it's not a habit I've cultivated. And it also doesn't work for returning multiple array/hash results.

In general if you get into the whole map, grep, etc frame of mind then returning lists is better than refs. Returning refs is marginally more efficient and better fits how some other people work. That also allows you to keep your style more similar as you move between Perl and other languages. The whole wantarray and return a ref or a single element seems like a good idea, but my experience with it has not been very good.

So any choice you make can be defended.

Note that the referenced link discusses prototyping without officially frowning upon it, thus, at least by implication, endorsing it.

It also invokes all of its example subroutines with a  & sigil, another frowned-upon practice. I only spent a couple of minutes reviewing the material, but I got the feeling it is a mixture of good and bad advice, with no way for the novice to tell which is which.