The Perception of Time
Interesting stuff from Tom Vanderbilt:
Talking about the city’s “Transit Tracker” program, which allows people to get real-time info on bus arrivals via their cell phones, Hansen mentioned a study that had been done in the U.K. of a similar program. What was noteworthy was that people using the service felt that the bus service itself had improved, that more buses were running, that they were running closer to schedule, even though none of this was empirically true.
I have a particular interest in the fluid nature of time, and the way travel, queuing, and even routing can play additive and subtractive games with this. Paco Underhill, for example, notes that people who wait in airport lines overestimate the time they waited by some 50 percent. I’ve also seen it noted that a train trip with a transfer feels longer to people than it really is, that people overestimate the time it will take to walk somewhere and underestimate the time it will take to drive somewhere. Of course, one of the masters of managing time is Disney, with its posted wait times (just posting the time makes it feel shorter for people) at queues, wait times which are then inflated — so the payoff at the end is even better: That wasn’t long at all!
Among other things, this is important because I think improving the level of objective and subjective service people get from the bus is important to our transportation future. Ultimately, I think rail is essential as the backbone of a major metropolitan area’s mass transit, but that rail backbone can have its utility massively extended if supplemented by good buses. This is also why if you’re ever taking the Subway in New York City you’ll generally be happier if you get on the local train rather than waiting for the express even if the express would be faster. Waiting around makes people very unhappy for some reason.
36 Responses to “The Perception of Time”
Register
RSSLogin
Topic Cloud
ArchivesBlog RollAbout Matt YglesiasContact Me
Donate
January 16th, 2009 at 1:17 pm
I just wasted two minutes of my life reading that.
January 16th, 2009 at 1:21 pm
to Ron: it only felt like two minutes.
January 16th, 2009 at 1:22 pm
This brings to mind a distinction we sometimes draw in computer science between performance and predictability of performance. This has actual practical consequences beyond subjective emotional reaction; it’s easier to plan around a wait if you know how long it will be. For example, imagine the train will arrive in 5 minutes and I have a 2 minute phone call I need to make. If I know the train will be here in 5 minutes, I can make the call. If I don’t, I may be afraid to start the call in case I have to interrupt it when the train arrives. So I can make good use of my time as long as I know the schedule.
Of course, it’s that much if there’s an actual reliable schedule so I don’t even have to show up at the station until a minute before.
January 16th, 2009 at 1:24 pm
Isn’t the local vs. express more about you know what you have (a local train on the platform taking x minutes) but not what you’ll get (an uncertain amount of wait time (y) for a train that’s faster (z, where z < x)), so you don’t whether you’ll win out by waiting (that y+z < x)?
January 16th, 2009 at 1:32 pm
This also points to a complaint I’ve heard but never understood, that it’s wrong for airlines to “pad” their timetables by indicating travel times that may well be longer than they actually (or usually) are. I don’t see why this is a problem, especially when the alternative is for airlines to promise times they might routinely fail to deliver on. It’s all well and good to say, “Well, they ought to just get the time right,” but I’ve worked in an airport and given the variety of variables airlines face each day, that’s darn difficult. We don’t like it when other companies promise products they can’t deliver. Why should our expectations of airlines be any different?
January 16th, 2009 at 1:37 pm
People tend to greatly overestimate the time they spend waiting for elevators. Fifteen or 20 seconds often seems interminable.
January 16th, 2009 at 1:43 pm
The bus service in Salt Lake City is abysmal. You walk through a driving snowstorm to get to the bus stop minutes ahead of time, only to see the bus barreling past ahead of schedule. Then you can wait an indeterminate time for the next one, which will probably be late.
January 16th, 2009 at 1:45 pm
It’s natural selection — waiting around exposes one to predators.
January 16th, 2009 at 1:51 pm
I was recently involved with a project at a university in the Metro-Boston area and the MBTA, trying to implement some sort of wait signal or clock to let people know when/where the next bus was. We spent a couple weeks discussing sensors, GPS systems, signal location, computer programming, which bus stops to test, etc.
Well the project was scrapped when we were told by an upper level MBTA manager -his first and only sit in- that while he appreciated our (free) work, it would never be implemented because they do not run the posted schedule. A bus that’s supposed to run 4 times an hour typically runs only 3 times or even twice an hour, and they didn’t want to “get busted”. Sad really.
January 16th, 2009 at 1:57 pm
That’s way too simplistic, of course, on the local vs. express train advice. It makes a big difference how far you’re going. I work way down town but live in Harlem. It would almost never be worth it to take the C train all the way to work- the A would have to be more than 30 minutes late, maybe more, for that to be worth-while. But it’s almost never that late. So, it’s almost always worth waiting. I know this so it doesn’t bother me. But, if you’re only going a few stops, and the local w/ stop 3 times while the express only once, then it makes great sense to take the local.
January 16th, 2009 at 1:58 pm
Oh I’m going to miss life in Zurich when I have to move back to Texas this year. Not only are buses a lifeline for every segment of the population, they are strictly coordinated with the tram and train schedules. They have their own lanes. Then run on electric. They’re never late. I drive maybe 3k miles a year, usually for vacation or personal reasons, because I don’t have to! And parking in the city costs about $7 an hour…
January 16th, 2009 at 1:58 pm
Not too surprising, my girlfriend often mentions how much she likes the electronically displayed wait times on NYC’s L subway line. For a lot of people, a five minute wait is much less stressful when you know how long you’ll be waiting.
In the case of a bus, where you may be waiting outside in the rain/cold, uncertain arrival times become even more stressful. Plus on buses you only pay once you board the bus, which means while you’re waiting, you haven’t actually committed to riding the bus, so you’re prone to second guessing your decision more.
That said, internet/cell-phone accessible wait time info really DOES improve the quality of bus service. It allows people to time when they go outside and walk to the bus station, effectively reducing average wait time across the board.
January 16th, 2009 at 2:06 pm
ConorClockwise, if you’re interested in a system that has been around for decades in providing real-time information to passengers of bus, tram, and train, see this PDF:
http://www.andynash.com/zrh-pages/zrh-docs/Zurich_Transport_QA_2007.pdf
Excerpt:
What sort of monitoring is there to ensure compliance with standards for buses and trains, such as frequency etc are not compromised?
In Zurich there is an automatic system that monitors the position of buses and trams. The control center knows exactly where vehicles are at all times. The bus tram operators even have a display on their control panel that tells them if they are ahead of or behind schedule. This helps them drive the vehicles so that they maintain the schedule. The automatic location system was implemented in the 1980s before today’s GPS and wireless communications and is based on relatively simple technology (wires in the ground).
Transport planners regularly analyze the tram bus schedule data to identify systematic delays or unreliability on routes. If problems are identified the routes are studied in detail to determine the problem cause and to identify solutions.
The real-time schedule adherence data is also used to make announcements to passengers about delays and problems as well as to control bus tram operations (in other words controllers redeploy the vehicles to avoid problems like traffic accidents that block tracks etc.). Controllers have pre-defined detours for all routes and scripts to read in announcing detours delays.
It is also important to emphasize that schedule reliability is naturally high due to Zurich’s comprehensive public transport priority program, which optimizes the network to reduce public transport delays. The control strategies outlined above would not work as well in networks with many delays.
Switzerland’s national railroad also uses automatic monitoring to ensure trains operate reliability and without delays. The IVT has helped develop some of the information technology tools used to identify problems and evaluate solutions. The Swiss Federal Railways has an extensive on-going research program to develop new strategies for increasing reliability.
January 16th, 2009 at 2:08 pm
This has been known for a long time. When I was a transportation planning intern at the RTD over ten years ago the models we used to predict ridership on light rail lines used a higher factor for wait time and transfer time than travel time. All of this was based on people’s observed behavior in choosing transportation modes, not stated preference. ‘Smart Travel’ has been around for a while too.
Incidentally, research at the time showed people underestimated train travel time compared to buses by something like ten to twenty percent.
January 16th, 2009 at 2:16 pm
This is true for the 1-2-3 and the 4-5-6 in Midtown/Downtown (especially with the East Side’s level-shifting requirement).
Often not so at 59th Street Northbound for the A/D v. B/C.
A ca. 1-1.5 mile trip (20, 24, or 30 blocks) with two or three additional stops (e.g., 18/23/28, 79/86, or 50/59/66) doesn’t have that much extra delay; switch if it’s there, otherwise go local, even in “frequent service” times.
Otoh, 72/81/86/96/103/110/116 is a long stretch of stop-and-go driving that can make a three-minutes wait worthwhile, especially if you’re going to 168th or above anyway.
One might do calculations and realise that people who live in poorer areas get less optimal service. But that would surprise exactly no one.
January 16th, 2009 at 2:22 pm
Matt may be too young to remember this, but back when the NY subways were pulled from the abyss by finally doing the necessary deferred maintanance, replacing antiquated trains, and so on, one of the biggest improvements was requiring that conductors actually tell people why their train had come to a halt in a tunnel, rather than just keeping everybody in the dark (sometimes literally back then).
In the old days, I had been in trains where people came near rioting because of a delay. Somehow just knowing the nature of trouble really helps. Now, it’s a bit unsatisfactory because the information the train crew knows, or is allowed to give, is often sketchy or meaningless (”police action,” “red signal against us”), but the occasional detailed explanation (sick passenger three stations ahead, emergency crew on scene, should be cleared in X time; smoke condition at X station, etc.) is extraordinarily helpful.
January 16th, 2009 at 2:48 pm
The bus service in Salt Lake City is abysmal. You walk through a driving snowstorm to get to the bus stop minutes ahead of time, only to see the bus barreling past ahead of schedule. Then you can wait an indeterminate time for the next one, which will probably be late.
Ding ding ding!
The big problem with buses isn’t that they are often late — it’s that they are sometimes early. The bus to work that picks me up on my property is scheduled to leave the station a mile down the road at 8:08 AM. That means that at the earliest, it could be at my stop at 8:10 if it hits all of the lights just right (more like 8:13 if it has to stop at the busy intersection) — except, of course, sometimes it’s five minutes early (8:05). But usually it’s five minutes late. So when I roll down at 8:07, I’m sitting there waiting until usually 8:18 or so. But during all of this time, I don’t know if perhaps this was the one day a month that the bus came a full five minutes early, in which case I’ll be waiting until the 8:38 picks my up (usually) about 8:48. So why the hell don’t I go back into my house, check email, and have a cup of coffee? Well, of course, it’s because it probably will swing by around 8:18 or so.
And it’s the same story in the evening, when I have to take a quick cross-downtown bus to catch the 5:05 that could come anywhere between 5:00 and 5:20. Except the transfer is right across the street from a bar and I usually can tell if the bus has already been by (no other people at the stop), so I will just grap a drink or two while I wait for the 5:35 (which will arrive between 5:30 and 5:45 but is usually late).
January 16th, 2009 at 3:04 pm
Thank you for saying that. I hear that complaint a lot and when I say, “well, that’s good — the airlines should build in some extra time if they might be late, rather than quoting a best-case ETA” — I looks of confusion and diagreement but never a cogent argument against “padding” the schedule. It seems to be conventional wisdom that it’s bad, and even smart people just swallow and regugitate this stuff w/o thinking about it.
January 16th, 2009 at 3:04 pm
Thank you for saying that. I hear that complaint a lot and when I say, “well, that’s good — the airlines should build in some extra time if they might be late, rather than quoting a best-case ETA” — I looks of confusion and diagreement but never a cogent argument against “padding” the schedule. It seems to be conventional wisdom that it’s bad, and even smart people just swallow and regurgitate this stuff w/o thinking about it.
January 16th, 2009 at 3:41 pm
There is an urgent need to quantify these insights and begin building them into the computer models used to estimate ridership. Such models routinely put heavy penalties on waiting and transferring, based on real experience with situations where these things are really unpleasant. A generation of modelers internalized these factors and can recite them with confidence as though they were facts of mathematics.
This is a crucial issue in transit development, because these models drive Federal funding decisions, and often also influence decisions at state and local levels.
January 16th, 2009 at 4:57 pm
In one of my psychology textbooks it mentioned a building where lots of people were complaining about the slow elevator service. So they put full-lentgh mirrors at the elevators. People got distracted checking themselves out and their perception of the speed of elevator response improved.
January 16th, 2009 at 7:07 pm
I was thinking of this yesterday while waiting for a bus in Portland. Once I called TriMet’s tracker number and knew I only had 12 mins to wait it wasn’t that bad. When I lived in Boston I never had any idea when a bus might appear (or even where it would go..).
January 16th, 2009 at 7:38 pm
Man, I would love something like this in Japan. It’s just too much of a pain to go to the bus stop, look up times, then wander back. Being able to check online if you’ve already missed the bus that won’t come for another 20min would be usefule, especially when it’s raining/cold/sucky weather in general.
I have the train schedules available via internet on my cell, but it’s not real-time, no info on delays and whatnot.
January 17th, 2009 at 8:09 am
Jeremy– next time you’re at the relevant bus stop, just take a photo of the timetable with your cell phone. No need to go out in the rain and check it next time…
January 17th, 2009 at 11:57 am
This is why short feeder bus routes feeding into a stopping train station, light rail line, or mass transit line is so much more effective than a sprawling bus route system … short bus routes with layover times at the station structured around the train schedule are far more predictable than long haul bus routes.
And certainly, for routes that can support less than the 10,000 passengers per day for a light rail line, focusing bus priority improvements on a particular set of trunk routes, and converting the system into short feeders into prioritized express buses is better than nothing … indeed, if the trunk buses are given overhead electric lines and made into trolley bus routes, that system can start tapping into the property development benefits of a light rail line at passenger levels well below the levels that justify a light rail line.
It is true that Bus Rapid Transit is normally proposed in the US in settings where a Light Rail line is preferable, but that’s more a function of the specific bias against funding dedicated transport infrastructure and the general bias over the last forty decades against necessary public investment in infrastructure … if we were rapidly putting together all the regional rail, light rail, and mass transit lines that made sense, then the accompanying increase in bus usage would also be driving substantial investment in priority improvements for the main bus routes.
March 2nd, 2009 at 4:26 am
levitraGreat site. Good info
March 5th, 2009 at 3:10 pm
Here is a look the best site in 2008 for adults.
I do not know the right sort of a link and it is not clear to you at the forum how to do it.
If you are under 18 do not go to links
March 22nd, 2009 at 5:50 am
tramadol
I want to say – thank you for this!
March 23rd, 2009 at 4:14 am
viagra
Incredible site!
April 2nd, 2009 at 4:19 am
Excellent site. It was pleasant to me.
buy cheap viagra
April 3rd, 2009 at 3:58 am
I bookmarked this site. Thank you for good job!
cheap brand pfizer viagra