First thing that came into my mind

http://xkcd.com/552/

Unless they have a nice biological explanation, a random correlation will always be the most probable explanation. 

There is a book/article around the association of the size of the city with a bunch of other factors. I do think that a lot of factors in a city will be established, even without thinking about it, by the size/distribution of the city. It is kind of how development restricts the possible evolutive pathways that an organism may take.

Second...The association is logaritmic. It makes me always nervous, because it increases the probability that the authors massaged the data until they got one approach that gave them what they wanted.

Third...they better have a very objective way to measure waking speed. Humans know that they are being studied, and this clearly could affect the person's gait speed. A person living in a large city is much more likely to have heard about research projects. They may want to "help" the researcher by walking a little bit faster. Maybe they used video from surveillance cameras or a video left somewhere in the city. There also could be biases due to the place or time, etc

Interesting study, but I think it is way harder to do it than it looks.

From my G+ post by blog contributer Kate:

What's your angle? It seems like it'd be a simple dynamic...bigger cities have more busy people in them, and once you have a few fast-moving (busy) people on the streets, it becomes increasingly inconvenient to stroll around slowly, so everyone's pace picks up...