As I can't see the full article, I'll respond to the quote.
It seems the number of public domain vs. copyrighted works is likely skewed by the fact that any one can (and does) publish PD works. For popular authors (Dickens, Austen, etc) there can be dozens, if not hundreds, of editions available.
If filtered simply for in-print/out-of-print, and eliminating multiple editions, I think the gap would close significantly. I still think PD would be a significant portion, but not such an overwhelming majority.
At least that's my gut reaction. Someone with the time and data crunching inclinations could prove me wrong.