JD Power does two ratings. The first, an "initial quality" rating, is based on surveys of new cars. This rating measures quality off the showroom floor, I believe defined as defects reported in the first 90 days. It also measures consumer satisfaction with the functionality of the car as measured by surveys of customers. JD Power does not discuss how many vehicles (or customers) they evaluate to determine their ratings, but I'm quite sure it's more than a half dozen.
Their other rating is a "dependability" rating, which measures reliability after 3 years of ownership. Again, they don't reveal how many vehicles they survey to obtain a statisically significant sample, but I'm quite sure it's based on more vehicles than any one of us would own in a lifetime.
Obviously, for current year cars, or in fact any car that has been on the market for less than 3 years, they have to use some sort of projection to forecast the predicted reliability. They make these projections based on their collected history on similar models, or in some cases where they have an entirely new model they base their prediction on their overall history with a particular brand.
You might suggest that these predicted ratings are then based on a very small sample, but in reality they are based on a very large sample of cars. For example, where there is an earlier version of a particular model the prediction would be based on a large sample of that earlier model. Given the amount of commonality across models that often exists "under the skin" this turns out to be a reasonable approach. Think about how engines, transmissions, and even smaller components such as alternators, water pumps, and so on are used across many models and you can see why this method will predict reliability reasonably well. To be sure, it's not as good as actual data on the car in question, but for a new model where the historical data simply can not be had it's a pretty good approach.
The accuracy of these predicted reliability ratings can suffer when you have a major new model introduction, or a major new component such as a new engine. But then, that's exactly why many advise against buying such new models until their reliability is proven.
Back to the question at hand. In the case of the 2008 JKU they show a overall dependabiltiy rating of 2 out of 5, and a powertrain dependability rating of 2 out of 5. In the case of the 2008 FJC they show an overall dependability rating of 4.5 out of 5, and a powertrain dependability rating of 5 out of 5.
I chose those years because they should have the full three years of data behind the ratings. All JD Power ratings are freely available at
www.jdpower.com
It is true that the domestic manufacturers have made great strides in reliability. Unfortunately, Jeep's overall rating still runs to the middle of the pack at 2 out of 5. I don't mean to suggest people shouldn't buy a Jeep. I appreciate the capabilities they offer, and as I said in an earlier post, if your needs run towards the rock crawling end of the spectrum the Jeep very well may be the best choice for that application. For most of us though, reliability should also be a consideration, and to the extent it is important it should be evaluated based on the best data around, not anecdotal evidence.
BTW, where the sun rises is a meaningless example to this discussion as it is not a function of statistics in any meaningful sense.
