That's alright, article-writer: no hard feelings!

I've written about this many times before, most extensively in this post that provides a pretty comprehensive comparison, IMHO, so I won't reiterate the pro and con arguments again. Apparently, similar arguments can be found in the book from which the survey is drawn (that I will buy and read in electronic form) though, according to the reviews, the author comes out a little more firmly - but also equivocally - on the side of paper than I. Of course, a balanced analysis doesn't make for such zappy headlines. Some people would love the world to be black and white, despite copious evidence of colours and shades all around them.

What I will observe, however, is that a survey result like this is not at all surprising, but the comparison is completely unfair and, largely, meaningless. A p-book is essentially a single technology (unless you count the difference between folios and scrolls), while there are thousands of different e-book technologies. It is ludicrous to compare an e-book read on a computer screen from an abomination like VitalSource, or a PDF file designed for paper, with the same thing read in native format on a new-ish Kindle, and very different again on a tablet or smartphone. Software can make a huge difference too: it is fair to say that a lot of it is still very badly designed, but there are decent e-readers available, and they are getting better all the time. This is not to mention issues like DRM that muddy the waters and get bundled into the same bucket. To treat this all as a single technology is dumb. It's like comparing walking with all other forms of transport at once (and, similarly, none will or should ever fully replace walking).

Given the enormous diversity and the infancy of the tools, It reminds me a lot of the spurious and nonsensical comparisons still sometimes being made between e-learning and p-learning. Both can be great and both can be ghastly. It all depends on how you do it. 

E-text (and e-learning) is getting better in leaps and bounds, with no sign of a change in rate of improvement, while p-text (and p-learning) has hardly improved in any significant way in a thousand years (I'm not counting manufacturing and distribution improvements and accept there have been  evolutionary tweaks like indexes and paperbacks). I doubt that paper books will ever die out - they do have unique advantages that we are still far from replicating on a commercial scale in other ways - but I think the eventual winner, assuming civilization as we know it doesn't collapse, is very clear indeed.