>In a 2001 study, two scholars in Canada asked 70 people to read “The Demon Lover,” a short story by Elizabeth Bowen. One group read it in a traditional linear-text format; they’d read a passage and click the word next to move ahead. A second group read a version in which they had to click on highlighted words in the text to move ahead. It took the hypertext readers longer to read the document, and they were seven times more likely to say they found it confusing.

In other words, if you mangle a short-story and push it into a different format from the one it was written in, it's harder to read. That has nothing to do with hypertext.

>She found that comprehension declined as the number of links increased—whether or not people clicked on them. After all, whenever a link appears, your brain has to at least make the choice not to click, which is itself distracting.

That's more a question of wordiness than of hypertext. Whether and where to link is a very important decision in modern composition, probably more important than word choice.

Basically, these studies are assuming that all content is created equal, and that's far from the case. I would say it's more a question of poor web design shattering focus - that said, much of the web is designed with the intent of shattering focus.