Actually, if a reasonable argument can be made that driving a car for transportation statistically leads to increased travel distances, and/or that riding a bicycle for transportation does the opposite, the per-mile consideration needs to be set aside. Potentially with notation referring to the exclusion, but not necessarily. If it is included, a sound argument needs to be made in support of inclusion, comprehensive work needs to be done to nail down the effect; as accurately as possible because this additional variable will compound any current discrepancy, and the effects of inclusion vs exclusion of this additional material would need to be outlined individually.
Unfortunately, the study is barely cited on the Yahoo page. Off to Reuters.
Nothing more there, but there is that token "Health Affairs" mention. Must be what qualifies as an in text citation for professionals; two words short of plagiarism. Google... And here's the abstract. Need a password for the full text, bah! Off to the college website, hee hee, free scholarly reference material for me, the evening student... Double bah! They don't have the May/June publication online yet! Slackers!
Oh well, it'll be up soon enough if the orderly row of Health Affairs publication dates is any indication. It'll be interesting to see who they cited and how the odds were stacked.