Not writing an academic paper about it doesn't make it right either. This is the problem, when it's just hearsay, the rest of us have nothing to judge its accuracy with.

When looking at something as complex as people's behaviour, then it's worth allowing for seasonal variations in public transport use (e.g. when it's raining people may decide to get the bus rather than walk). Could it be related to the length of their journeys and/or their net worth? (There's a strong correlation between health and wealth). How does it compare to a similar section of the population that drive or cycle instead?