In an act that I'm sure was designed to throw me into a fit of rage, Markus Knauff and Jelica Nejasmic published An Efficiency Comparison of Document Preparation Systems Used in Academic Research and Development in PLOS ONE. The paper purports to compare the efficiency of Word versus LaTeX in academic document preparation.
Most everyone here knows my feelings about Word but that doesn't mean I'm not open to a reasoned analysis of the relative merits of Word and LaTeX in producing technical documents. Sadly, this paper is not that analysis. Claus Wilke has an excellent post on the problems with the paper. You should definitely read it.
The thrust of Wilke's post is that the design of the experiment discussed in the paper is very weak and does not support the rather strong conclusions that the authors reach. Most of the problems stem from the authors' failure to distinguish between layout and content. For example, one of the tests involved reproducing exactly a page of continuous text (including title, author, and address lines). But, of course, LaTeX users don't work that way. They download a style sheet from the journal they're submitting the paper to and let the style sheet worry about the layout.
The most infuriating thing about the paper is their explanation for—what they consider—an anomaly: the fact that LaTeX users express less frustration and enjoy using their system more than Word users. They posit that the poor deluded fools who feel that way must be suffering from cognitive dissonance. Because, after all, it couldn't be that LaTeX actually was better for it's users.
Their other conclusion is that journals—at least non-mathematical journals—should stop accepting LaTeX and insist on Word or PDF. This even though they admit that LaTeX is easier for the journals to integrate into their work flow and produces a better product. I find their reason for this conclusion particularly offensive. Since their study concludes that it takes longer to prepare a document with LaTeX than it does with Word, researchers should be forced, forced, to use a document preparation system other than the one they'd like because then they'll have more time to do research and improve the common weal. As if researchers were slaves. It's a disgraceful argument.
UPDATE: There's an interesting discussion of the paper on the Org-Emacs list. In particular, see the post by Christophe Pouzat. He obtained the original data and reanalyzed it. His findings are very revealing.