Hacker News Reader

Comment by dangoldin | Hacker News Reader

I think it's unfair to compare cars from 50 years ago against cars today without comparing the behavior of the drivers. Although the cars themselves may be safer, more people may be on the roads and may be driving more recklessly which would lead to a higher number of accidents as well as having the damage be equal.

I remember reading something about a study being done that showed people that cars drove much more dangerously around bikers with helmets than bikers without; if you feel that other cars are safer you may drive more dangerously.

Edit: Here's the study: http://www.bath.ac.uk/news/2006/9/11/overtaking110906.html

View on HN

I believe it's a fair question. Assume for discussion that there is a higher rate of auto injuries per year today than 50 years ago. Ok, is that due to the cars being less safe, the drivers being less safe, or something else? This video isolates one possible cause (the cars), tests it, and determines that they are safer today than 50 years ago. Clearly, if today's cars are safer, the cause of the injuries must lie elsewhere (drivers, roads, some external factor, etc).

I love this comment. It gets to the heart of a scientific inquiry. Many experiments are just like this. You try to control for as many variables as you can, then you test variations on the remaining variables. Observing the effects you can start to see how the two are correlated, which would be impossible to isolate from real-world data.

Of course, the post is not necessarily applicable to the real world because the original assumption may be incorrect.

Ideally you would be able to carry out such an experiment but as troystribling mentions it's possible that an increase in the safety of cars may lead people to drive more dangerously. It's hard to isolate this correlation.

An experiment may be to show that holding everything else constant, a car from 2009 would lead to 50% fewer serious accidents. If we look at the data and notice that there were only 25% fewer accidents, we would then conclude that it may have come from the factors mentioned.

This is the 'Peltzman Effect' http://en.wikipedia.org/wiki/Peltzman_effect

"The Peltzman effect arises when people adjust their behavior to a regulation in ways that counteract the intended effect of the regulation. So, for example, when the government passes a seatbelt law, some drivers may respond by driving less safely"

How about we look at actual data?

http://www.publicpurpose.com/hwy-fatal57+.htm http://www-fars.nhtsa.dot.gov/Main/index.aspx

Total highway fatalities (in the US) have gone down since their peak in the 70s. This despite the fact that the population of the US has grown tremendously and the total vehicle-miles driven has grown even more so. The number of fatalities per vehicle mile has dropped by nearly a factor of 5 since 50 years ago (and has fallen 20% in the last decade alone).

If people are driving more dangerously, the data surely doesn't show it.

This is a very valid point and raised very well. It does not deserve these downmods.

Don't forget that in the 1950's it was socially acceptable to tank up on a few highballs, take one for the road, and get on your merry way.

There has been so much more safety education in the last 30 years. It's bound to have rubbed off a bit.

why? this feels like a red herring fallacy.

we're talking about the technology and quality of the cars across time as it relates to safety.

you pull in people as drivers and it fundamentally changes what we're talking about. more examples:

"I think it's unfair to compare cars from 50 years ago against cars today without comparing the state of modern traffic laws and speed limits"

"I think it's unfair to compare cars from 50 years ago against cars today without comparing the quality and ubiquity of roads, highways and interstates"

etc.

My general point was that although cars are safer than before, we still need to look at the system they are a part of which affects safety. I don't want people to make the conclusion that it's safer driving now than before - that may or may not be the case.

Well... it is:

http://www.roadandtravel.com/safetyandsecurity/highwayfatali...

Note that nationwide statistics are not available that far back, but consider this report from Washington State:

http://www.wsdot.wa.gov/planning/wtp/datalibrary/Safety/MVFa...

But that hasn't necessarily come from car safety although I'm sure that's a large contributor.

It can also be attributed to better education, more cracking down on drunk drivers, etc.

I think it's unfair to compare cars from 50 years ago against cars today without comparing the behavior of the drivers, improvement of safety education, improved laws, better police enforcement, etc..

see what i did there? ;)

Reply on news.ycombinator.com