One theory would have to include the probability of a stall. A situation where moderate or severe turbulence had disconnected A/P and A/T, and the PF had instinctively reduced thrust, with or without correct airspeed indications. In case of inoperative stall warning, the resultant airframe vibration or buffeting was masked by turbulence. Since the airplane remained intact and impacted the water in a pitch-up, low bank and little sideslip attitude, it's conceivable that the airplane remained stalled during the entire descent. The crew subsequently may also have tried to re engage the automatics, but without first lowering the nose to effect stall recovery.
Such a stall event had played out on 22 Dec 1996 when an ABX DC-8-63F on a test flight encountered a full stall with severe buffet in night IMC without the benefit of a working stick shaker. Upon recognition of the stall, the crew had added thrust, but had failed to lower the nose.