I don't have the links handy right now, but I believe two criticisms were:

-- For some reason the participants all wore loose, oversized t-shirts, which in at least one case made it more difficult to draw the weapon. If they had been wearing more common attire (shirts that fit, tucked-into pants, etc.) the results may have been different ...

-- While the test subject was kept unaware of the test, the "assailant" not only obviously knew about it, but also knew there was an "armed" person in the room and where they were sitting, and after hitting his primary target immediately started shooting at the test subject ...

Ah! Yes. Links:

Link 1

Link 2

Link 3

And many, many more in the Googleverse ...

---

Actually, it's a shame this experiment was so shoddily constructed -- a well-executed experiment might have yielded some interesting and relevant results. While the test subjects would have undoubtedly performed better, I have heard that unless you've actually been in a real shooting/being shot at situation, you're never really prepared regardless of how much training and practice you may have had. It would be interesting to see if that's actually true or not ...