In Part 1 of this post, I looked at walking distance from my own life experiences. In a nutshell, the idea that people will only walk 1/4 mile doesn't seem to hold up very well in the real world.
A nice discussion started by Steve Mouzon and carried on by Kade Benfield illustrates some of the reasons why. Mouzon states that walking distance is a variable, that it is affected by "Walk Appeal," which could be summarized as "how enjoyable is a walk?" Benfield's post adds that there are probably non-aesthetic factors that matter as well.
These conversations are very informative. They also illustrate a weakness of this subject matter, which is that aside from measuring trips to transit stops there hasn't been a lot of scientific study of general purpose walking. Writing for Slate, Tom Vanderbilt describes the situation this way: "Walking in America is a bit like sex: Everybody’s doing it, but nobody knows how much."
Right now we have only a little data, and without some more it's hard to have any hard scientific conclusions about why and how far people will walk. However, summarizing my own experience and what existing writing I've seen, I'll offer a hypothesis, and then a guess about the value of studying it further.
Walking distance is variable. Collectively, people's tendency to walk to and from a given place is the result of a mental calculus consisting of the following variables, weighted in order of priority:
Safety. When an environment is physically or socially unsafe (or perceived as unsafe), voluntary walking approaches zero.
Convenience. If walking is competitive with other modes in terms of travel time and cost, then people will do it. People routinely walk a mile to avoid paying high parking costs, or to avoid dealing with heavy traffic.
Practicality. People are more likely to walk when they are chaining together many close-by destinations, especially when they are meeting other people for part of the trip, or going out drinking etc. Conversely people are less likely to walk if they need to carry a heavy load on the trip, or if they are going somewhere in formal dress and the weather is not ideal, etc.
Enjoyability. People are more likely to walk when and where walking is enjoyable, and conversely less likely to walk where it is boring or uncomfortable.
Finally, it's worth noting there is another factor that is very important but strictly individual in nature: fitness. For individuals who do not have difficulty walking, this isn't part of the calculus. But for people who have physical disabilities or injuries, this is a significant factor.
How do these factors play out in the real world? Here's another example from my life experience:
From my house, it's about 0.5 mi to two shopping and entertainment districts. These are nice walks; one of the routes I could take on foot includes crossing a lovely city park. The weather in Raleigh is pretty fantastic, aside from the very peak of summer there are few days where it's uncomfortable to be outside.
One of the nearby commercial areas has a little grocery store. On Saturdays, when I don't have a lot going on, I routinely walk to the grocery store. On a Monday evening, when I realize just before dinner that I've run out of ketchup or something, I drive there.
As much as I'm a walking advocate, it takes me only about two minutes to drive to either of these areas, and it takes about 10-12 minutes to walk. So, for a round trip, I can either spend 5 minutes total in-transit, or 20-25 minutes in transit. It's simply a calculus of time.
If parking at the grocery store cost $2, I would probably walk more often. If it cost $10, or if there was no parking, I would probably walk all the time. I also would probably just forgo the last-minute ketchup run on a Monday night. Keep this last option, forgoing the trip, in mind for a moment.
The hypothesis I've made above gives us a loose framework for understanding how people think about walking. But, as far as trying to study or model this more scientifically to come up with a hard-science formula for predicting walking distance, I would actually say it may not be worth it.
The option to forgo the trip illustrates the problem with trying to model walking distance. Walking distance is variable, depending on a large number of individual and contextual factors.
Cities are complex adaptive systems, and transportation choices are an emergent pattern within the system. Scientists refer to patterns like this as having "high sensitivity to initial conditions," meaning your model won't work if you're even a little bit off about the variables you put in.
Most importantly, there are always other options that a model or formula will not be able to account for. People adapt their behavior to changing circumstances, consciously or subconsciously. Small changes in circumstances can result in big behavior changes.
This reinforces of the importance of trial and error experimentation . When dealing with complex systems it is usually more productive to simply try things and see what happens.
So, a final takeaway for urbanists of all vocations: when we're dealing with transportation and built environment, asking questions about how people will get to and from a place, let's not worry too much about trying to fit into engineering models. Instead, let's step back and take a common sense approach. Try things in small increments, and see what happens. Learn from these little experiments, and embrace the idea that small failures are a healthy part of how we learn.
 Tim Harford's TED Talk on the subject of trial and error experimentation explains further why this is the best approach to many kinds of scientific questions.