Wednesday, March 30, 2011


My mother-in-law gave me a link to an online survey offered by the NTTA. This is a great example of why survey results need to be taken with a grain of salt.

I was asked when was the last weekday day, not date, that I had been in a car. That would have been for my sons one-year birthday dinner, three weeks ago. That time should make a big difference in the travel patterns this survey is trying to find.

It then asked many questions related to that trip, however, that was only part of the survey, the other, I presume, is more of a policy guide. Which route would you take: One with no tolls and a lot of time, one with a small amount of tolls and a medium amount of time or one with a lot of tolls and a shorter amount of time. It also used a time frame in there, between 5:30 and 6:30 or outside, showing a bias towards work commute patterns, when that is roughly only a quarter of all metropolitan travel. I also have a fourth option not included, a transit route.

After they asked about the preferences of the respondent, the questions migrated towards the general do you prefer tolls for a faster commute?, how should our roads be paid for? and other toll-related items. This is what I think they really are doing the survey. This seems to have more use as an opinion survey presented to decision makers and legislators than as a trip survey.

A brief decision of how modeling actually works proves that. A metropolitan area is divided into hundreds of zones (In graduate school, the example we worked with was small, at 500 zones). Within these zones, attractions and housing are determined. As an example, downtown with its many jobs, is considered a high attraction (though in reality it is a few zones) while the suburban zones generally have lower housing amounts, there are many of them. Routes are than drawn, based on the regions roads between the zones. Many mathmatical formulas exist to determine the amount of trips on the roads between zones. Many use different things, like number of jobs, houses, retail, recreational activities, density depending on the model used. Then real world measurements are taken of actual traffic counts at these locations. Once those are obtained, the model is calibrated using adjustments and run again until all the numbers are close to the observed totals. Finaly, the calibrated model is run for X number of years in advance using projections into the future of where growth, stagnation and decline are expected to occur.

Obviously, this is much more complicated than asking where your last trip was. That makes me believe this is a device being used to push an agenda. I would hate for this to be put out there that this is being used for a road building agenda. I hope this is purely being used as a road financing tool. Instead, if its being used for the generic term of transportation, then its is skewed and not worth the paper it is printed on, and since it is an internet survey, that is saying something.

No comments: