Automation and workload.
While flight deck automation in modern aircraft has reduced some aspects of pilot workload in approaches, particularly in the very lowest visibilities, it has certainly not reduced the need for good cockpit management. Rather the reverse - indeed, the 2013 US report of the PARC/CAST Flight Deck Automation WG emphasised the importance of flight deck task/workload management as a factor in overall flight path management.
In many accident and incident reports, 20/20 hindsight makes it easy to see that the crew were inadequately prepared for the situation in which they found themselves. Very frequently, they had performed a briefing, but had to some extent been just "going through the motions".
When doing the briefing, the crew perceived that they were "in a safe place" as far as threats were concerned. The weather was fine, the aircraft was in good shape, and they were familiar with the destination. But later the situation changed, and things did not go as expected either on board or in the surrounding environment. Then one or more of the following occurred.
- The pilot flying (usually the Captain) became absorbed in dealing with new and immediate problems while also handling the aircraft. In the process he lost sight of "the big picture" which showed approaching danger.
- The pilot monitoring, (the First Officer in the vast majority of cases) also became over-involved in helping the PF's attempts to fix the short-term issues. The PM may have detected the developing hazardous situation but was unable either to divert the attention of the pilot flying to it, or assume personal control.
- The weather actually encountered meant that visible cues during the latter stages of the approach were far less than anticipated.
- When a go-around was called for and attempted, inadequate preparation led to loss of control.
This sort of inadequate preparation often results from the crew starting the descent apparently under the impression that because their current flight situation contained few if any safety threats, this benign state of affairs would continue all the way to the destination.
So while in many of these cases the CVR did not record proper briefings, on the contrary it often did show "non-pertinent" discussion of other matters, sometimes work-related and sometimes not. The subsequent accident report shows that the crew then became overwhelmed by changed circumstances, sometimes triggered by their own actions or inactions.
A good example of this might be the 1995 AAL B757 Cali accident, where according to an excellent analysis by David Simmon, "There was no information on the CVR either of an approach briefing or that the descent checklist was performed. There was confusion about the aircraft's position and an absence of dialogue about planned speeds, crossing altitudes, radio tuning and management of displays...... The CVR transcript showed that there was concern about whether there was time to retrieve the approach chart and that there was a sound similar to rustling pages."