Mac look alike?

[Odd]

Colin Law wrote:
> 2009/11/30 Odd <iodine at runbox.no>:
>> Also, for those who think the survey makes Apple look bad, Apple 
>> did actually come out pretty well. And the difference from the 3 
>> ahead of Apple is not that big. Apple is only 1.8% behind the best,
>> Asus.
>> 
>> So I don't see what the fuss is about. Why does Apple have to be
>> the absolutely best?
>> 
>> Take solace in the fact that Apple is among the 4 best, rather than
>>  the 5 worst.
> 
> In addition there are no statistical details on the figures.  Without
> confidence limits we have no way of knowing whether the small 
> differences seen are real or just down to random chance. If you back
> to the original source

Exactly.

> http://www.squaretrade.com/pages/laptop-reliability-1109 and download
>  the report it says that they have enough data to provide
> statistically significant data on the two year failure rates but says
> that the three year figures are projections (though how they have
> forecast it is not clear to me).

You take the original 2-year numbers, and extrapolate, perhaps?

> This says to me that nothing can be read into the three year figures

Not nothing, but it is more uncertain, for sure.

> On the two year ones the difference between the first manufacturers
> is very small.  The report states that the results are statistically
> significant but even that is meaningless without more information.
> What is it about the results that is statistically significant?  What
> is the confidence level that Apple is 1% worse than Asus?

1.8%. I'd say it's a 1 in 4 chance that Apple is best.

> Without more information I think all one might say is that the 
> manufacturers towards the left of the chart are probably more
> reliable than those towards the right, but I would hesitate to say
> more than that.

That's probably right.

-- 
Odd