I took a test. Between a big city with highrises and a suburban, divided by river. Both equally big in size, with same utilities, zones etc. I deleted all transport routes (Before this bus and metro had most passengers). Then i layed down one bus route, train route and metro route, with all stopping at same spot. The bus had to take a road around the river, while train and metro crossed over.
With the policy on and off, it didnt matter. Bus passengers was decreasing while train and metro actually rose by 10-15%. More cars was on the roads, less people walking, those who used to take bus would rather drive.
Verdict: The policy seem to work for everything. The only reason i would believe buses are preffered is because they got more stops then the train, and are usually more connected. I think Cims calculate travel distance same as they do with cars. So if a taking a bus is faster than taking train or the car, they will preffer the bus. But i think the nr. 1 priority is walking though.
This is my observation.
FYI, i tested this with fast time loop, while shuffling the policies on and off, (with transport mode) i started testing 1 train 1 bus, 1 metro. then after some times loops i doubled the train and metro, and tripled the bus. There result was as written above. Tested this for about 30-40min.