This age group includes college students, but do we know whether the surveys filtered out textbooks when polling physical vs. ebook preferences? I didn't see that mentioned in the article, and it would have a significant impact on the results.