IFL: Prius Tops All 14 Hybrids in Safety!

[kens]

Originally Posted by Glider

Therefore, my only choices are to either not post any results for those models with limited published ratings (per IIHS and NHTSA) or to consider that missing element to be "typical". To do otherwise would be arbitrary.

Sincerely,
Michael D. Dulberger
Founder and President
Informed For Life, Inc.

Although it's nice of Mr. Dulberger to reply to the email, his reply really doesn't make much sense. He says it's arbitrary to assume that a car is identical to the most similar vehicle tested.

He is right that, as shown by his Accord 2DR vs 4DR example (and many other 2DR vs 4DR examples one can find in his site's data), it's not a perfect measure.

But, as noted by RodJo, to use an AVERAGE car instead of the MOST SIMILAR car is even more arbitrary.

More importantly, I think it's crucial to be transparent about this aspect of the methodology. Maybe it's explained somewhere on the IFL site, but if you look at the page for, e.g., a TCH there is no explanation of how the missing data could be affecting the score.

That said, I should note that I think the composite index that IFL publishes really is a great idea. My quibble is just with the implementation.

[Glider]

I edited my first post in this thread to try to give a fairer ranking based on modified IFL scores. Please see the new table shown there.

- g