I don't have a direct source, but I've heard that Bombardier has basically said that the Dash-8 is not susceptable to icing induced tail-stalls.
Tail plane stall accidents have been occurring for decades, with several aircraft types showing particular susceptibility to the problem. DeHavilland Canada and Transport Canada got bit pretty bad by tail plane stall on the Twin Otter. There were a number of accidents and incidents after flight in icing where the tail stalled after selecting full flap. The problem was addressed on the Twin Otter, if I recall correctly, by a Flight Manual limitation against landing with full flap after flight in icing conditions.
Having learned from the Twin Otter experience, Transport Canada became the first major aviation authority to require certification flight tests to verify no susceptibility to tail plane stall. The test, as flown in Canada, requires simulated ice shapes to be installed, to create the worst case condition that would be found in FAR 25 Appendix C icing conditions. The aircraft is flown at VFE for each flap angle at forward CG, and extremely aggressive nose down pitch inputs are made, targeting 0.5g or lower. Typical tests use close to full nose down elevator, and the lowest load factor hits 0.2 or 0.3g. All Dash-8 models would have passed these tail plane stall tests.
Of course, we must keep in mind that certification tests only address conditions within the FAR 25 Appendix C icing envelope. There is good data to show that some parts of the world frequently have icing conditions that are much more severe than this (the area south of the Great Lakes is one of these areas). There are discussions to create an updated icing envelope that covers a larger percentage of the conditions that actually exist.
Based on all the info I have received, some of which is not yet public, I am certain that this aircraft did not experience a tail plane stall, and that the icing conditions were not a direct cause of the accident, except possibly as a distraction.