Although the analysis is not volume-adjusted (borrowing one book from a library is treated the same as buying several from a bricks and mortar store), a couple of things stood out and were highlighted in the linked post: younger respondents (defined as 18 to 29) borrow books more often than any other age cohort, and seniors (defined as 65+) were much less likely to purchase books from online bookstores.
The other interesting thing, not specifically addressed by Forrester, was the use of multiple channels across all age groups. It appears that people don't just buy at retail, online or borrow books; they do all of these things.
To me, that suggests that offering content in multiple ways could well increase the chances of meeting and nurturing a reading habit. Useful data perhaps, for publishers thinking about limiting format choices.