Archive for July 8th, 2008

Imagine there were two roads that could be fixed to improve drivers’ travel times. Imagine also they’re the same length, but have different average speeds.

Fixing the first road will see drivers’ speeds increase from 40 kph to 50 kph.

Fixing the second road, meanwhile, will see drivers’ speeds go from 80 kph to 130 kph.

Given that the construction costs would be the same to fix each road, which road would you choose to improve? Which would result in more time saved (and, hence, more social benefit)?

If you’re a regular, math-challenged sort like me, you probably picked the second option. But apart from some minor rounding differences, the amount of time saved for each is actually the same.

This example comes from a fascinating new paper, “Driving Speed Changes and Subjective Estimate of Time Savings, Accident Risks and Braking,” by Ola Svenson, head of the Risk Analysis, Social and Decision Research Unit at Stockholm University.

I’m admittedly someone who has trouble computing things like how much time an increase in speed will save — once you get me off the nice 60 mph mark (a mile a minute), things get a little fuzzy. But now I’ve found I’m apparently not alone. When Svenson asked this to a group of respondents, a majority thought the 80 to 120 kph option was better. The reason, he speculates, has to do with a sort of “proportion heuristic,” an effect that’s been found in many other contexts (the work of Daniel Bartells is instructive here); in what he calls a “ratio rule,” we’re biased by the ratios in speed changes, rather than employing the actual, more complex formula.

What this means on the road is that we may underestimate time savings of increases in lower speeds and, perhaps more importantly, overestimate the time savings we gain when we begin to accelerate from an already high speed. He also looked at estimations of crash risk and braking distance as it applies to speed — both of which are non-linear — and found similar mis-estimations. He notes: “Intuitive arguments for higher speeds may be biased.”

Astute readers will note the curious echo here of another study, the so-called MPG illusion, by Duke University researchers Richard P. Larrick and Jack B. Soll, which found that increases in MPG from 16 to 20 can save as much fuel, relatively, as going from 33 to 50 mpg.

Previous research has found that we tend to overestimate trip lengths, and we think the waits at things like traffic lights were longer than they really were. The lesson in Svenson’s new paper, I suppose, is that that a car’s instrument panel is more complicated than it appears.