I believe 3272 was a wing stall followed by a tail stall. If you notice when they were about 3,000 feet and the airspeed was back up to 180 knots but the excessive nose down pitch caused a tail stall. They were clean when this happened and the EMB 120 stall speed is 117 knots clean without ice.
No engineering data is available to determine whether (or not) the EM2 is susceptible to tailplane stalls; the accident report doesn't say word-one about a tailplane stall and the tailplane stall recovery technique is
the exact wrong thing to do in the event of a low-airspeed, icing-induced
roll upset...which is what happens in transport-category turboprop airplanes without powered flight controls.
(You must push.)
I won't speak to the certification issues, since I'm not quite qualified to talk about them, but I can tell you about roll upsets in the EM2 and similar transport-category turbopropeller airplanes.
Comair 3272 is similar (indeed, very similar) to the Eagle ATR that crashed in Roselawn. Aircraft with non-powered flight controls are susceptible to icing-induced roll upsets if ice accumulates forward of the ailerons, causing the airflow to separate from them. This problem happens at high angles of attack (for reasons that should be obvious). Both the Eagle ATR and Comair 3272 were operating in non-trivial amounts of ice at low airspeeds; both aircraft lost roll control, and it was, as one might say, "off to the races."
The accident airplane here decelerated through 150 KIAS, flaps up, and roll control of the aircraft was lost during a left bank due to asymmetrical ice accumulation forward of the ailerons.
The NTSB found:
Had the pilots of Comair flight 3272 been aware of the specific airspeed, configuration, and icing circumstances of the six previous EMB-120 icing- related events and of the information contained in operational bulletin 120- 002/96 and revision 43 to the EMB-120 airplane flight manual, it is possible that they would have operated the airplane more conservatively with regard to airspeed and flap configuration or activated the deicing boots when they knew they were in icing conditions.
Findings, page 177.
The EMB-120 actually has a long history of various handling problems going slow in the ice; Comair had a few of these incidents (accidents, really, as in one case severe damage was done to the horizontal tail during the recovery from a high speed dive following a roll upset). ALPA's submission to the NTSB on this accident:
Beginning in 1989, the documented history of EMB-120 ice induced roll upsets began with an event in Klamath Falls, Oregon. Over the next six years, the aircraft experienced five additional ice induced roll upsets. A majority of the events appear to have occurred at approximately 150 to 160 knots with high roll angle excursions and significant drag increases being experienced.
Incidentally, after all of this, it happened to Comair again—this time, they lived to tell the tale, but severe damage was done to the tail after the crew recovered from the roll upset and during the recovery of the ensuing dive. (Their EADIs tipped over and tumbled, too. Not bueno)
The net result of all of this was that a bunch of changes were made to the Brasilia, including an admonition to activate wing and tail de-icing boots at the first sign of ice accumulation, the importance of doing at least 170 knots, flaps up, in icing conditions, and the installation of a low-speed icing condition warning alarm (colloquially, the Comair box). The SWS was also modified to take an icing input and activate the system sooner in icing conditions (the specific details escape my small brain at the moment).
tl;dr: roll upset, must push nose down to get control again.
