Quote:
Originally Posted by Autobacs
Well the fact that they disclosed it was based on a limited sample because they actually had a minimum cutoff should already tell you it's not going to be conclusive so you cannot treat is as gospel. Obviously you can think that every BMW model is now very reliable because it clearly says it is overall BASED on models that met their minimum cutoff. To throw away the results because on an assumption made when the survey already discloses its limitation is really just throwing bricks at a glasshouse. Anybody in here going to dispute that a Lexus overall very reliable just because they feel CR's survey is worthless for whatever reason? I don't think so  |
I'm confused... why are you taking this so personally? do you work for CR?
I never said to throw away the results or that they were worthless. I was just pointing out a statistical weakness in their analysis that could potentially skew the results we see. To be discerning of the results rather than blindly following them, I thought, was the point of a discussion.
I'm actually a subscriber to CR, so I'm well aware of what they DO disclose vs. what they don't. And they've never fully disclosed their data (which to some extent I understand - data is $$). But as someone who works closely with data in day to day profession, I have to take what they show with a grain of salt, ESPECIALLY when something jumps 10 places from one year to the next.