On the phenomenology of crowding and reliability, from Alon Levy:
People’s perception of utilization is always going to be skewed upward. There are always more people to witness crowding than emptiness. If half the trains run at 100% capacity, and half run at 20%, then five sixths of passengers will see 100% crowding rather than 20%; therefore, real utilization will be 60%, while perceived utilization will be 87%.
The same is true with frequency – more people are there to witness a train run late than to witness it run on time. In both cases, both the actual and the perceived numbers are important: the perceived numbers measure passenger experience, whereas the actual numbers measure system costs.
All very true of urban rapid transit. The exception, of course is the lonely empty bus doing Coverage work in a low-density suburb. Its emptiness is all too visible to the motorists all around it.
As for reliability, one thing I admire about real-time information displays is that they expand the audience that can notice when the service is on time. An electric display that says “Next bus: 17 minutes, 19 minutes, 22 minutes” can help broaden public understanding of how bad things really are beyond just the riders who’ll actually wait all that time.