My experience for what it is worth. Approaching to land at Manila in a 737-200 many years ago. Due ATC we were kept high (4000 I think) for noise abatement reasons and intercepted the localiser around 8 miles. That means the glide slope should have been on slope of three degrees at roughly 8 times 320 = 2560 ft. Sea level airport approx.
We were above 8/8th stratus cloud which was tops around 2000 ft and base about 1000 ft (from memory). We were cleared to descend when established on the localiser and had already dirtied up and reduced airspeed in preparation for a steep descent from 4000 where the glide slope indicator should have displayed full scale fly down. Wind was light and variable in the air and on the deck.
But it didn't and in fact displayed a slightly fly down indication of one quarter a dot. This obviously did not compute with the DME v height calculation. By the time we were around 2000 ft and going into cloud layer the glide slope indicator still showed very slightly fly down while the speed was accurate for the configuration of land flap, but the rate of descent was well over 1300 fpm and stable.
I guessed we had a false glide slope indication since in another life I did navaid flight calibration work and knew a bit about the symptoms although had never actually seen one that I knew of. I was about to go-around when we cleared cloud and had runway visual and obviously very high on slope. The ILS glide slope indicator still showed one quarter dot fly down....The T-VASIS was no help as it was all over the place and making no sense.
Manila was 11,000 ft length so we adjusted the visual approach to reduce rate of descent to land further along the runway. Landing safe with plenty of runway to spare.
Contacted ATC and said the ILS glide slope had a flyable false glide slope of around 5-6 degrees and the T-VASIS was dangerously useless and should be de-commissioned until fixed.
ATC replied they had a permanent instruction to have the T-VASIS on for all landings and that is why it was on. They added that they had serious earth tremors that day which may have knocked the T-VASIS boxes around but they would investigate it anyway and thanks for the info. But they still left the T-VASIS on!
Next day they arranged an ILS calibration test and found a flyable false glide slope around six degrees. Maintenance of the Manila navaids was a well known problem but it seemed we certainly had flown the six degree slope.
Several months earlier a China Airlines Boeing 707 from Taipei had crashed well short of the same runway on a visual approach. The 707 was a write-off. The captain swore he was on glide slope all the way, but that he needed idle thrust and speed brakes to keep the speed back to a manageable figure. The rate of descent must have been horrendous in a 707 in that configuration and idle thrust but he said he was on slope.
See this link:
AirDisaster.Com Accident Photo: China Airlines Boeing 707 B-1826
Asked why he didn't go around when it was obvious close in something was wrong, he told the investigation he was worried the noise of the engines spooling up would frighten the passengers
It would seem he had experienced a similar false ILS glide slope that we had a few months later but for some reason he pressed on regardless. An ethnic culture of real men don't go around, perhaps?
Although as far as I recall, the subject of the possibility of a flyable false glide slope was not mentioned in the accident report, I wouldn't be at all surprised that was what he was unknowingly following while ignoring other information that would have proved a glide slope problem. Eg DME v Height.