It's good food for thought. MPG doesn't tell the whole story when you consider the costs of materials, manufacturing, transportation and disposal/recycling ... but the article has plenty of flaws in its own analysis.
Here's a pretty good discussion about it, showing both solid research and faulty assumptions .
http://www.thecarconnection.com/Auto_News/Green_Car_News/Prius_Versus_HUMMER_Exploding_the_Myth.S196.A12220 .html?pg=1
One of their more problematic concepts is the idea of "social energy expenditures", such as worker commuting costs and transportation expenses.
Sounds like they're looking at the "big picture" ... until you realize that source and supply manufacturers make parts and materials for lots of applications, not just cars. And employee commuting costs might include stops for breakfast, shopping, banking or dropping their kids off at school, not just the actual trip to work. What portion of these "social energy" costs should be applied to a "dust to dust" vehicle cost?