Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/17/84; site rduxb.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!rduxb!smh
From: smh@rduxb.UUCP (henning)
Newsgroups: net.auto
Subject: Re: Volvo's saftey -- Hard data
Message-ID: <651@rduxb.UUCP>
Date: Mon, 26-Aug-85 09:56:42 EDT
Article-I.D.: rduxb.651
Posted: Mon Aug 26 09:56:42 1985
Date-Received: Wed, 28-Aug-85 01:53:48 EDT
References: <62@zeus.UUCP>
Organization: AT&T Bell Laboratories, Reading, PA
Lines: 41

****                                                                 ****
From the keys of Steve Henning, AT&T Bell Labs, Reading, PA rduxb!smh

> >According to 8-19-85 USA Today, the auto insurance industry released data
> >on the personal injury rates for 1982-1984 cars sold in the US based on
> >the number of insurance claims.  A rating of 100 is average.
> 
> Since I did not read the article, I have a couple of questions with
> regard to computation of the rating:
> 
> 1. Is it based on the total number of claims of a model or is it
> based on the number of claims per (e.g. 100) automobile(s) of a model?

It is based upon the number of claims per insured automobile per insurance year.

> 2. Is there any relationship between the number of claims of a model
> and the average risk rating of its drivers, i.e., are we looking at the
> safety rating of the different models or are we looking at the safety rating
> of their drivers?

This is not a black and white issue.  You are looking at the model plus
1) Driving habits of people who choose that model
2) Safety habits of same such as belt use and child restraints
3) Accident avoidance capabilities of the car and of the drivers
4) Accident risk factor in areas where a particular model is usually sold
5) Miles per year driven by typical drivers of that model
6) Highway versus local miles driven by typical drivers of that model
You can see this clearly with American cars.  Many times the same
model with a different division's label will have quite different
personal injury and fatality data yet they are the same car, almost.
The main difference is in the advertising and the type of person they
are targeted for.  For example Pontiac has a high injury/fatality rate
compared to Olds for equivalent models.

Laboratory safety tests have similar problems.  They factor out all
of the above which are a part of the real world and have accident tests
which are idealized and not real world.  However lab tests are the only
fair comparisons you can make even though they are unrealistic. 

If you are a typical driver of one of the cars on the survey,
then the results probably are a good indicator of your risk to injury.