There were no aerodynamic icing effects documented during the Colgan investigation. Colgan was related to icing for two reasons: first, the incorrect reference speed was bugged; it should have been 20 knots higher based on the icing reference switch being selected in conjunction with the operattion of the ice protection system. Second, it is possible that the crew incorrectly perceived the shaker and pusher as an ice contaminated tailplane stall.
But actual ice accretion did not have a role. Alpine Flyer is correct. There have been no Dash 8 accidents due to inflight icing (they respond to inadequate ground deicing, particularly on the upper stabilizer surface, about as well as any airplane...which is to say very poorly). The ATR continues to have difficulties.