[IamIan]

Originally Posted by ksa8907

I disagree with your numbers, because it is not stated, any assumptions and extrapolations made cannot be taken as fact.

1 + 1 = 2
I don't need someone else to 'state' or tell me it's 2 .. I can do the math myself .. that is all that I did .. and I didn't even do any complex math either.

Originally Posted by ksa8907

SUV death rates fall

From this article, "But drivers of today's SUVs are among the least likely to die in a crash, the Institute's latest calculations of driver death rates show. The change is due largely to the widespread availability of electronic stability control (ESC), which helps prevent rollovers. With the propensity to roll over reduced, SUVs are on balance safer than cars because their bigger size and weight provide greater protection in a crash."

Bold added.
You're comparing apples to oranges.

I did an analysis of the actual numbers of what actually happened in the real world .. my numbers did not exclude all the things they tell you they excluded in that paper you site .. I did not go out of my way to make a bias as that paper you sited did ... as shown bellow.

You're pointing to the odds of death (in the event of a crash) ... which can only relate to the number I posted ... IF

1. Odds of getting in a crash are the same... they are not.
   Your own link bellow also directly refuets this .. in their own words:
   • "because they are more likely to get into situations that ESC is designed to prevent"
2. Sample size of how many chances are the same... they are not.
3. If all the exclusions they made .. they didn't make... but they did.
4. etc ...

Originally Posted by ksa8907

My wife and I are very good drivers. We allow plenty of following distance, obey the speed limit, and brake early.

Good

Originally Posted by ksa8907

My primary concern is for the safety of my family. I want a vehicle that will protect us from all the other idiots out there. Simple physics, and the IIHS, show that larger vehicles are safer.

If that is your primary concern .. than take a bus .. they are safer per mile than personal vehicles (as previously sited) ... they are safer per trip than personal vehicles (as previously sited).

And even if you want to insist on the belief that larger and heavier is safer ... than the bus is both larger and heavier than any normal personal vehicle you could drive.

If that truly is your 'primary' concern ... than your course of action is clear.

The math I showed is even simpler than the physics ... and there are numerous simple physics arguments that can be made about the better overall safety of the lighter vehicle.

Originally Posted by ksa8907

Here's the .pdf with more info and numbers http://www.iihs.org/externaldata/srdata/docs/sr4605.pdf

Read closer.
Not only is that still focusing on the 'in the event of a crash' ... with all the errors sited before about such a PoV.

But..

They also come right out and tell you some of the other choices they made ... that it doesn't take a genius to tell would change the results if they didn't do these things... these are all from your own link... in their own words.

1. "The Institue computes driver-only death rates, because the presence of passengers varies."
   • Which is them coming out and tell you ... they ignored / didn't count any of the deaths of passengers ... easy to see the flaw in that approach.
2. "A registered year is 1 vehicle registered for 1 year or 2 vehicles for 6 months each"
   • wow .. they come right out and tell you they intentionally count 2 vehicles as if they were only one ... hmm .. again not hard to see that could change the results.
3. "Although the vehicles span 2005-08 models, only those equivalent to 2008 models are in-cluded"
   • They come out and tell you they are intentionally excluding vehicle data .. based on whatever standard they have chosen as "equivalent" ... hmmm ... again... not hard to see how that might alter the results.
4. "In other words, if a vehicle was completely redesigned for the 2007 model year, the 2005-06 versions weren’t counted. The exception is the Malibu."
   • Wow .. the hits keep coming ... not only to they repeat their cherry picking of the data to crunch ... but they aren't even consistent with the method .. they just randomly make an exception for the Malibu ... again ... not hard to see this can effect the results.
5. "Among 4-door midsize cars, for example, the lowest death rate was 19 for the Honda Accord, and the highest was 99 for the 2007 Chevrolet Malibu"
   • Sooo ... the exception they made before to include in 'cars' ... was the very worst of the cars ??? ... hmmm ... again ... not hard to see this kind of cherry picking can easily stack the deck for the results.
6. "Death rates are adjusted with information on driver age and gender and vehicle density"
   and
   "Because the latest round of driver death rates has been finetuned in a way that previous ones weren’t"
   • Something wrong with just using the actual death rates number ... instead of adjusting them ? ... again ... not hard to see this can effect the results... and they even tell you this new adjustment is very different from their previous adjustments.
7. "Although vehicle performance in the tests varied, the results didn’t correlate with insurance loss patterns"
   • They come out and tell you the real world loss patterns of insurance companies don't agree with their vehicle test results ... all the more reason to just use the real world death numbers I used previously.