We now know more about the built environment and transportation

According to Reid Ewing and Robert Cervero, the most intensely researched topic in urban planning is "the potential to moderate travel demand by changing the built environment." Just within the last decade there have been dozens of published studies asking this question from many different angles, using different methodologies. So this dynamic academic duo has decided to do the rest of us a favor and consolidate these studies, pull out a common thread of measurement between them, and weigh the variables against each other. They did something like this in 2001 when there were 14 studies to look at. Now they've compiled results from more than 200 and included about 50 of these in their meta-analysis. I think it's fair to say this is the closest we've yet been to answering this complex question.

The study is published in the Summer 2010 Journal of the American Planning Association, and there are already a couple of nice discussions on the internet from Laurence Aurbach and Kaid Benfield. I have to say that one of the things I appreciate about the planning field is the real interaction between the academic world and practitioners. Working professionals and activists really do read this stuff, and most of the journals make every effort to eschew jargon and ask questions that have relevance.

Here are the land use variables that are traditionally considered in relation to travel behavior:

The D Variables	Summary Description
Density	Concentration of a variable of interest (population, dwelling units, activity centers etc.) per unit of area
Diversity	Number of different land uses in a given area.
Design	Quantifiable characteristics of the street network, such as density of intersections, connectivity, or streetscape features.
Destination Accessibility	Measures the ease of access to common trip destinations, usually in terms of distance
Distance to Transit	Shortest route along the street network to nearest train station or bus stop

The impact that each of these variables has on travel behavior, whether it's vehicle miles traveled or mode choice, is referred to as elasticity. It's "the percent change in the outcome variable [like vehicle miles traveled] when a specified independent variable [like density of dwelling units] increases by 1%."

There are two more D's mentioned that have little to do with the built environment. They are Demand Management and Demographics. Demand Management is mostly parking supply and pricing, but I imagine any economic factor could be considered under this category. Economic triggers factors can be considerable, especially if fuel prices are taken into account, but I understand how this is beyond the scope of what they are doing. Demographics are contolled for in all of the studies in the meta-analysis.

The following chart presents the findings in ranked order from the most significant factor to the least in achieving three outcomes. The variables are color-coded according to the categories defined above.

	Reduction in VMT	Increase in Walking	Increase in Transit Use
1	Distance to downtown	Intersection/street density	Distance to nearest transit
2	Job accessibility by auto	Distance to nearest store	% 4-way intersections
3	Intersection/street density	Jobs-housing balance	Intersection/street density
4	% 4-way Intersections	Land use mix	Land use mix
5	Land use mix	Job within one mile	Household/population density
6	Job accessibility by transit	Distance to nearest transit	Job density
7	Distance to nearest transit	Commercial floor to area ratio
8	Household/population density	Household/population density
9	Jobs-housing balance	Job density
10	Job density (no effect)	% 4-way Intersections (negative)

Some additional points ...

•  None of these variables are gamechangers, so don't be a physical determinist. Even the most significant factor, the effect of street density on walking, has .39 elasticity. This means a 10% increase in connectivity would lead to a 3.9% greater probability that someone will choose to walk. That being said, these figures are cumulative, so adding the effects together can make a notable difference. There are no solutions to anything, only means for incremental improvement.

• The findings seem to show that density itself is not as important as some make it out to be. As they put it, "almost any development in a central location is likely to generate less automobile travel than the best-designed, compact, mixed-use development in a remote location." Absolutely. Although I understand how, in the absence of coherent regional planning, an architect would want to do the best she can with the site she's given. Who knows, maybe in the future it will be the next "central location"? Still, even if it takes cleaning up a brownfield site or working delicately with the neighbors, infill seems to always be more effective.

• One of the limitations of elasticity is that it measures relative change only. You could be on the verge of a tipping point, and a tiny little nudge would lead to big outcomes, but these numbers would not tell you that. Can those tipping points be identified empirically and built into the same model?

• A big win for connectivity, which is great because this is something that can actually be done. Some cities and states are starting to write codes to ensure a robust street network in new developments. Even more important is retrofitting connections into existing networks. Hopefully these results will spur localities to look for those odd scraps of land and consider punching a street or multi-use trail through them. Although cycling was not considered in this analysis, I can attest from personal experience that street connectivity is the single most important factor for enhancing safety and convenience. Cyclists would much rather take an alternative back route than ride along a busy road with bike lanes.