The concept of Economic Man is that of someone who acts more or less perfectly in his own self-interest. For economists, it is a convenient approximation because it allows equations to be written and "thus" turn economics into a science. For anyone who isn't an economist, this is a perfect example of the maxim "all models are wrong, some are useful."
All you have to do is look around to see counterexamples in everyone, and the Wikipedia article above cites numerous criticisms of the idea. The "Elephant and Rider" model described in Switch produces a much more functional description of human behavior -- but it doesn't easily lend itself to mathematical modeling.
We get so attached to producing answers (being "deciders") that we want everything to act like science. We take useful observations and push them so hard in our desire to create science that we destroy any value they might have had. In our desperate need for answers, we take good information and turn it into half-truths -- which is why the core of Hard Facts, Dangerous Half-Truths And Total Nonsense: Profiting From Evidence-Based Management is dedicated to the exposure of half-truths, and to discovering which half is worthwhile (and to teaching the reader how to do the same).
The middle ground has the most promise: using science when we can, to produce helpful guidance, but not insisting that the results be a science that describes everything. If we can liberate ourselves from the deterministic mindset, we can benefit from what works best without demanding that it fit into our personal cosmology. What I find hopeful is that many of the leading books that I've been encountering seem to take this attitude -- there may even be a movement going on that is yet to be named, or whose name I have yet to encounter.