RAT 5, sorry if the elaborative answer did not meet your requirements. I belong to the ‘old school’ of training, which with the luxury of time and resource, aimed to teach the big picture, instil flexible thought, and lay the foundations of airmanship.
I sympathise with the training problems resulting from commercial constraints. The growing trend for simplicity and cost cutting has much to answer for, and unless the current training programs or operations can alleviate the training shortfalls, the simple dilemmas faced by your students will one day be opportunities for their accidents.
So forgive the minor rant over training – you, to days instructors, still have to cope.
Many safety problems appear to originate from lack of knowledge (historic or situation assessment) or the inability to associate this knowledge with a course of action in a situation (cognitive flexibility theory if you want the research). Some of the reasons (the principle reasons) for these shortfalls is oversimplification (time and resource constraint) and lack of variety when explaining situations – how to associate knowledge to situations other than in which it was taught – joined up thinking.
Now apply the issues above to the current discussion on PAPI. I would argue that while it’s safe to say ‘follow the PAPI’ for all visual approaches, this must be qualified by the potential problems (albeit small probability) of distortion, non alignment, or failure. Pilots must be taught that situations are not clear cut A or B; aviation involves complexity, quick and simple rarely exists, and often there is no one perfect solution so they have to be aware of alternative solutions and a flexible response.
I accept that A or B is satisfactory for the first lesson, but after that the subject must be developed and elaborated.
If you only have one lesson? Then the industry is in a very poor state.
Inexperienced pilots must be mentored by their Captains providing an expansion of knowledge during line operations. This is still feasible, but again commercial constraint and perhaps a greater risk, that of the new breed of Captains also lacking depth of training, really will begin to hurt the industry.
All this and it started with something simple as PAPI.
"what shall I do?" Where the aids don’t agree is an ideal situation for the student to learn about deciding themselves – ‘old school’ (needs some background knowledge). The training hints are:-
At longer range (before normal descent point) follow the GS. This can be associated with the need to be aware of safe altitudes during the approach, avoidance of black hole descent illusion, etc – debrief.
At closer range use the PAPI for visual approaches. However, if the GS / PAPI are still in error, follow the higher beam – safety first. If the error is not promulgated on the charts (giving a reasoned answer as to which one to follow), file a safety report because the beams should be close enough not to show such a disparity.