In a more popular version of a forthcoming article in the Journal of Transportation and Statistics, Donald Shoup chronicles the problems with the ITE Parking Generation and Trip generation manuals that have become standards for planners and transport professionals in the U.S. and even abroad. He demonstrates that their ostensible accuracy is unmerited and misleading; that average rates are most often presented despite lack of statistical evidence for any relation between floor area and trips generated; and, more profoundly, that the tables are based on assumptions of car dependent land use that the manuals help become self-fulfilling, as follows. Rates are derived from studies of peak parking demand and number of trips in suburban sites with ample free parking--lacking transit, pedestrian facilities, or TDM; these are then used to mandate the design of new developments, thus ensuring that parking in these will be free, and densities low (so traffic generated at the predicted rates does not overwhelm roads.) When such sites are surveyed in the future, they will, indeed, validate these rates. Voila!
Thanks to Yaakov J. Garb, ITDP