Originally Posted by jnbacon

I would bet that if you added density (# of people/sq. mi) to the regression analysis, the actual effect of green space on bicycling/walking behavior would be fairly small. It would be interesting to know what other factors were considered to have an effect, and what their contributions to the overall effect are.

And as to counting, in the Phoenix metro area, any employer with more than some number of employees (500 maybe?) has to administer a transportation survey and return certain percentage (pretty high - 90% or more, I think). These numbers would seem to give pretty accurate numbers, at least for a certain type of employmer.