Check this one out, it is the classic in book on causal inference: https://www.amazon.com/Experimental-Quasi-Experimental-Desig...

My impression this is going to be something other than a pure scientific experiment (and I don't mean that in a bad way).

Medicine and social sciences have a large body of research about things like the fact that you can't do a pure experiment with a control group if there's serious risks.

http://en.wikipedia.org/wiki/Design_of_quasi-experiments

The seminal work (which I slogged through once upon a time): http://www.amazon.ca/Experimental-Quasi-Experimental-Designs...

There are plenty of experimental designs other than double-blind that have been demonstrated to prove causality. Or actually "quasi-experimental" but still.

I had this book inflicted upon me in a past life: http://www.amazon.com/Experimental-Quasi-Experimental-Design...

Social sciences have well-developed methodologies for dealing with fallacies. http://www.amazon.com/Experimental-Quasi-Experimental-Design... is the bible, and is mostly taught to doctoral students. But chances are Mr. Collins has never heard of it, since he's in business faculty. But he might have, and might actually have a solid research design. Just that its hard to tell from the book.

The practice of management in the real world is full of imperfections, because its about people, and people are imperfect.