Quote:
Originally Posted by bradnixon
|
Studies, like the one that the referenced article is based on, really need to be understood for what they are. This one appears to be an exercise in statistical analysis. It does not seem to be a meaningful study of where to, generally, put bus stops – although in its Conclusion, there are wild statements about fantastic savings for transit companies.
The study, “The Far Side Story” (
Download the PDF), involves calculating the length of time buses spent stopped at Near-side, Far-side, and Mid-block bus stops in Montreal and Portland. Various factors were taken into consideration, including the number of doors on the bus, whether there was a right-turn lane before the intersection, the bus load, etc. However, the fundamental point of the study was to compare the length of time that a bus stopped at a near-side verses a far-side bus stop
at a stop level. That is, records were used that told how long a bus was stopped at each type of stop, and the statistics provided the results.
According to the study, a near-side stop will delay a bus by an average of 4.2 seconds in Portland, and 5 seconds in Montreal, more than a far-side stop in those locations. Part of the extra time in Montreal could be attributed to the fact that right turns are not allowed on red lights there. This means that buses in Montreal can get stuck behind vehicle waiting longer to turn right. In these cases, Montreal bus operators will often open the doors before reaching the near-side stop and customers walk down to where the bus is to board it; that walk takes some extra time, lengthening the boarding time, which is included in the statistic.
It all sounds great for a stunning conclusion: If the bus company can save 5 seconds for each of 23 stops (the average number of near-side stops on the bus routes examined), then over the entire day, a particular route could have up to 9 extra runs! Sensational.
Except, nowhere in the study did they take into account the most fundamental reason for a stop lasting longer before the bus crosses a signal-controlled intersection – that a red light was involved. Remember, the measurement was from the time the bus stopped at a stop until it left that stop. It makes perfect sense that at a far-side stop the bus arrives, stops, and can leave immediately, with no extra delay. At a near-side stop, a bus arrives, stops and then has to wait for a green-light to leave. According to the study, apparently at a Stop-sign-controlled intersection, there is much less of a delay than at one controlled by lights. Well, imagine that. It sounds as if the only thing delaying a bus from leaving a near-side stop, in that case, is any vehicle that needs to clear the intersection.
And this leads us to the flaw in the conclusion stated by this study; that there can be time saved by moving the stops to after the intersections. The flaw arises from the comparison of the times
at a stop level, based solely on the records of when the bus arrived at the stop and when it left that stop. The time that a bus waits at the intersection BEFORE it stops at a far-side stop is not included in the time for that stop. That delay can never over-lap the time for the bus stop; as it can at a near-side stop.
To be a fair, apples-to-apples, comparison to determine which is better, a study should look at the following time periods:
• For a near-side stop – the time from when the bus arrives at the stop (or near-enough for people to quickly walk to it and board/alight while the bus sits in traffic) until the bus crosses the far-side of the intersection;
• For a far-side stop – the time from when the bus stops for the last time before crossing the intersection (in case it takes multiple light cycles to cross), or from when the bus crosses the near-side edge of the intersection (if it does not stop), until the bus leaves the far-side bus stop.
This study, which is often reference as proving that far-side stops save time over near-side stops, does not prove anything of the sort. All it does is show that, if you include the random timing of a signal light into the time of stops at one location, then it will, generally, take longer than if no such additional time is included for stops at another position.