Monday, May 24, 2010

A transit hierarchy of needs

Note: Much of the material in this post was taken from comments which originally appeared in this thread at, Jarrett Walker's excellent transit blog.  It has been significantly edited for the current article.

One famous treatise in the field of psychology is Abraham Maslow's Hierarchy of Needs, which arranges human needs in a hierarchy from basic survival needs (food, water) at the bottom, to more abstract sources of happiness (self-actualization and such) at the top. The paper was controversial when written, and is still controversial, but its influence on the field cannot be disputed.

A hierarchy for transit

A similar model can be applied to many other endeavors, including to transit. I was going to make one up :) (and did so anyway) but then discovered this paper written by researchers at the University of Florida, after surveying transit users. The "Hierarchy of Transit Needs" as documented in the paper, is a five-level pyramid, with the most basic needs at the bottom. The articulated needs (starting with the most basic) are as follows--all text is quoted from the paper.

  • Safety and security. Personal safety, safety for personal property, familiarity with route, mode, and destination.
  • Time. Trip efficiency, in-vehicle time, waiting time, transfers, walk time, trip chaining.
  • Societal acceptance. Acceptance. Personal and peer/society attitudes toward modes (for or against).
  • Cost. Best value. Fixed (vehicle, assurance) and variable (gas, care, tolls, parking).
  • Comfort and convenience. Better travel experience. Comfort. More reliability. Easy access.

The questionnaire given to survey participants consisted of a set of questions asking which mode of travel among two choices (each optimizing a given value) participants would choose.  The vehicle type was the same in each question--questions were of the sort "do you prefer the fast bus through the bad neighborhood, or the slow bus through the nice part of town?"  (That wasn't one of the questions; just an example to demonstrate the flavor).  One major issue with the paper is that the survey participants don't appear to be anything near a representative sampling of society.  Survey participants were recruited from a psychology class, a campus listserv, and a local church--and the demographic profile of the respondents was overwhelmingly white, overwhelmingly students, and overwhelmingly childless.  (And geographically limited--the survey was conducted only in two locales; Gainesville, FL, and Savannah, GA--both Southern cities of a similar size to Eugene, OR).  Beyond that, I don't have any particular complaints about the methodology; though I'm not qualified to dive too much more deeply beyond the obviously-skewed sample.

My own musings

Prior to reading the paper, I had informally come up with my own list--which is based on my opinion and my opinion only, and has no scientific justification whatsoever.  But this is a blog, not a journal, so here goes anyway. :)

  • Access. Does the mode serve my current location and my destination, at all?  Right now, I can't take TriMet to get to Mount Hood (you could in the past); so TriMet is useless to me for that trip.
  • Safety and security. Pretty much the same as that presented in the paper. Is a user likely to get mugged, killed, or have his/her stuff stolen while making the trip?
  • Reliability. Can users depend on the service? If a Greyhound bus coming to a town is frequently two hours later (or one hour earlier) than its scheduled time, the service is too unreliable to be useful to any but the most desperate traveler; conversely, a Greyhound bus that arrives within ten minutes of its scheduled time is more useful; even if it only runs once per day.
  • Time and convenience. What Jarrett Walker calls "mobility"; what the paper calls "time". How far must one walk, how long must one wait, how long must one wait, and how often must one transfer? Is fare payment easy and convenient, or do riders need exact change everywhere they go? Is there sufficient information available, presented in a useful form, for users to make informed trip-planning decisions (including routing around delays or service disruptions, if necessary?)
  • Comfort. This refers to basic issues like ride quality, seating vs standing, are the seats too hard or too cramped, climate control. Can something useful be done while riding the bus or train (reading, working on computer)? Likewise, are the stations/platforms/stops comfortable?  A train may have plush seats but if it is frequently crushloaded, it is likely to be uncomfortable.
  • Cost. Cost to the user, obviously. (Externalized costs are another issue, of course; but many people are more than happy to externalize their costs). Pretty much the same as in the study. It is useful to distinguish three types of cost--incremental costs (what must one pay for an additional trip) vs access costs (what must one pay to have this mode available to them) vs system costs (what must one pay, as a member of society, to have this mode available to society?)
  • Amenities. Niceties which don't provide a direct mobility or comfort benefit, but which may make the trip more pleasant or productive--does the vehicle/platform/station have WiFi? Restrooms? Food and beverage service (whether it be a dining car, or a vending machine)? Nice art hanging on the walls?
  • Societal acceptance. Being a nerd, my (uninformed) gut instinct is to put this on the bottom; but some users simply won't use certain modes/routes/etc. due to various cultural prejudices. Dislike for busses among upper classes in much of North America is well-documented. Many transit services (chiefly long-distance services in the US) sell first-class tickets where, for a premium price, one can have nicer amenities, a comfier seat, and segregation from the riffraff--and don't think for a minute that the last isn't important. And bus segregation was an important concern for the US civil rights movement.

Obviously, my list is a bit more fine-grained than the one in the paper. The top of my list (the bottom of my pyramid) is an item (access) that appears to be considered a "given" in the paper. The big difference, as noted above, is the inversion of social issues and "comfort"; but as a self-proclaimed "nerd", I'm perhaps less sensitive to certain cultural attitudes and cues than are most people--or so I like to think.

A few things to note:

  • Safety is very important, particularly in the developed world.  System safety is generally a given in developed nations, where most safety concerns revolve around personal safety (i.e. not becoming victims of crime while using or waiting for the service).  The study paper noted this--many who claim to value safety expressed a preference for private automobiles, where they wouldn't encounter unsavory characters, despite the greater chances of being involved in an accident in a personal auto.  (Hence the infamous Chevy add to the right).  In other parts of the world, safety is often sacrificed for greater access and coverage--not a month goes by, it seems, without a report of a train wreck with multiple casualties (often in the dozens) in a developing nation.
  • Time is also very important, which to me is unsurprising.
  • I was a bit surprised that in the UF study, comfort rated low--given that many of the pro-rail arguments in the bus/rail debate center around comfort.  (Indeed, the original thread at humantransit centered around how bus driver skill can increase or decrease ride quality).  
  • One surprising thing was the low placement of cost--especially given that the sample was dominated by students, who are generally more sensitive to price than other segments of the population.  Income levels were not included in the demographic information published in the study report--though respondents were asked if they were presently employed on either a full-time or part-time basis.
  • Social acceptance placed high in the UF study, behind only safety and time--and ahead of cost.   
So, what does it all mean, anyway?
Ignoring any particular methodological issues with the UF study, it is worthwhile to ask:  Is this sort of model useful?  I would say yes, with the following caveats:
  • This sort of model is only useful as a first-order approximation--sufficient for abstract discussions of customer behavior; insufficient for planning purposes.  One issue with the "Maslow hierarchy" comparison--an issue acknowledged in the study report--is that it implies a stricter ordering of priorities than is probably manifest.  Customer behavior is doubtless too complicated to accurately describe with a bullet-ed list of values.
  • The UF study probably has limited use outside of the geographical area(s) in question.  It is doubtful whether it applies to larger nearby cities such as Jacksonville, let alone ethnically diverse areas like Miami or Atlanta--or any place outside of the South.  (And for this reason, any attempt to perform a study like this on a national level, to devise the views of an entire country in aggregate, are probably of limited use).  
  • The UF study also distills questions to their most abstract forms--values (safety, time) independent of any particular implementation.  Even with valid results for an area, one must consider what it all means in concrete terms.  In some parts of the world, for example, it is widely believed that busses are burdened with a stigma of poverty but rail is not--in other parts of the world, that is decidedly untrue.  
  • A transit agency or planning agency looking to improve or expand its service, would do well to perform more detailed market research than this--in particular, segmenting the marketplace even finer than a citywide or regional aggregate.  (Many transit agencies do this already, or plan to). 
With those caveats in mind, though, it is useful to keep in mind that the factors which drive passenger behavior (and which planners need to be aware of) are varied and complex.  A proposal which focuses on only one aspect of the hierarchy, even if at the upper levels, is probably going to not be successful.  While it is useful, for instance, to focus only on time-and-distance issues for a particular discussion; actual decision-making will have to be concerned with much more.  How much?  That depends on regional values, and is very much a political decision--not something that can be computed by rote.

A common mistake is to assume that values, or their implementations, are global--this is a common phenomenon in bus/rail debates, where claims like "choice riders don't like busses" are frequently heard. 
And in some cases, it is probably desirable to disregard the stated values of some customers--particularly when those values include exclusion or segregation of other customers.  (A common manifestations of the "social acceptance" value is avoidance of persons or demographics considered inferior or suspect--which in all too many cases, means minorities or the poor). 


  1. Really interesting.

    One possible system for sorting this out would be: Whenever you hear a claim about how important x is to ridership, ask: "Where in Maslow's scheme does x fall?" Is x something that only matters once more basic needs are met? Or is it one of those basic needs?

    Another way to ask the same question is: "Is the statement 'x will increase ridership' a geometric fact, or a biological fact, or a cultural fact? What you're asking there is "Suppose it's true that 'x will increase ridership.'" But for how long can we expect that to be true?

    Another difficulty in doing an absolute hierarchy is that certain things matter according to different kinds of curves. They probably found little senstivity to cost because they focused on a narrow range of costs to which students aren't that sensitive; someone living on social security might have had a different response. Comfort is an issue tied to trip length. I'll ride a crush-loaded subway for 5 minutes but not for an hour, just as I'm happy in Economy class for a few hours but not overnight.

  2. Thanks, Jarrett! Folks should check out Jarret's excellent response over on humantransit; where he does an excellent job bringing a rather abstract debate back to the big picture.

    And certainly, this exploration of the Maslow hierarchy does help explain some of the strong backlash many agencies, including TriMet, have seen. Where in Maslow's hierarchy does Portland Streetcar fit, for instance?

  3. I found this post thanks to Human Transit (and just subscribed). You do note that Maslow's hierarchy has its challengers, and I personally wouldn't do too much heavy lifting to try to turn that particular pyramid into a transit decision-making filter.

    Your list is pretty comprehensive. I don't think a rank-ordering approach works because different people will value the various components from their own perspectives and priorities. I see it as more like a matrix with the variables weighted by the individual user in his/her decision making.

    I have a factor in my own transit use decision-making that, while possibly "elitist" in nature (because I can afford to choose), might be something to address in thinking about the social marketing of transit: it's the right thing to do and I'm living my personal values by choosing this instead of my car.

    My motivation stems partly from my bike commuting habit. Once I broke the driving habit it became much easier to shift to the bus when it's too snowy or icy for the bike. It also extends my effort to diminish the size of my personal carbon footprint, so it's an expression of my values.

    Your "social acceptance" gets at this concept but I'd add another element along the lines of "values expression". Even though SOV drivers probably don't stop to think that they're expressing their personal values by driving--any more than most people think about how far their food has traveled and whether they really meant to consume that much oil to eat vegetables that probably grow locally--it's a factor at work for at least some people in a demographic segment I think could be effectively targeted to market transit.

    I recognize the perception that transit riders are low-income and undesirable keeps some professional types off my local bus system. I and several other business-y types were featured in a recent series of promotions (billboards, TV spots, bus signs) specifically to target that demographic. (You can see the TV spots on their Facebook page,

    Some of those same professionals probably have environmental concerns that could be addressed by changing transportation mode if they thought about it in those terms. The other variables have to line up, obviously, but the motivation could at least get them to read a schedule and give it some thought.

    As far as the validity of any study that draws primarily on college students, forget it. Those are convenience samples for the profs (who probably assigned survey participation for extra credit). As you note, there's no way to generalize unless you're trying to study the attitudes of college students in that region and even then it's iffy.

    I recently was named chair of a policy advisory committee for our regional MPO and am participating in development of a "visioning process" for a 20-year transportation vision (which obviously also involves land use--big goals!). Looking forward to learning from your blog.


  4. Thanks, Barb!

    I think the important thing, of course, is to understand what motivates people to use a particular choice at a particular time. The listing of a particular set of categories is meant to stimulate discussion, not to propose any absolutes--the topic has simply done that.

    (As an aside for any other readers: Jarrett has a followup, which reference a pair of posts by Cap'n Transit on the issue).

    One point which is worth making: Many people, I think, believe themselves to be acting on one level when they are really acting on another. Motorists who avoid the bus out "safety" concerns in cities where actual acts of violence on the bus system are exceedingly rare is the most common example. One needs to understand these "misplaced" motivations in order to deal with them.

    But of course, the bottom in line, as Jarrett points out, in many places the system is simply not useful for many trips, for all but the most desperate.

    Now I've got to go figure out why blogger seems to have turned comment notification off.... :)


Keep it clean, please