Reply to thread

Re: erobbins




You mean, like this one?


To get data meaningful enough to draw any conclusions, you would need reports from at least 20-30 percent of the ~2400 cars in the range - and that means from ALL the cars, chosen at random, rather than self-selecting because only those who have had failures decide to report.  And if you're looking to determine the repeating pattern of failure - x out of every y transmissions fail - then you would need closer to 90 percent of the cars in the range.


Back
Top