On Mar 12, 2010, at 2:35 PM, Simon Royal wrote:
According to MacInTouch the numbers are surprisingly high.
That was a self-reported survey, which any statistician will tell you is an invalid method of measuring anything in an entire population..it is NOT a random sample, which is the crucial step in statistical sampling.
If you run a self-reported survey for defects in anything, you'll soon think that there were an overwhelming number of defects, but people who participate in these do so in greater numbers when they have a defect to report. People lie on these things, regularly. People misreport on these things, regularly.
Non-randomness in surveys like this basically mean that statistically speaking, any numbers you derive are meaningless.
(No one runs to this list and posts "Hey, NOTHING went wrong with my Mac today! What can I do about this?")
Second the numbers are very small. No G3 iBook represented more than 1% of all the respondents, which were only 10,600 or so systems....Apple sold tens of millions of laptops between 1999 and 2005.
Look at the large chart. Of all the responses only 1316 systems (only 12% of all of them) needed logic board replacements for any cause (which would be the fix for the video chip issue in the iBooks).
The chart reports that a staggering 55% of the 700Mhz G3 14" ibooks needed logic board replacements.
55% of 13 is 7 systems. Only 7 respondents to this survey said they needed new logic boards (for any reason, this could be the video issue, this could be 'Tommy spilled Coke into the iBook issued by the school'.)
What is NOT shown, and is not knowable from this survey, is what proportion of the total population of 700 MHz G3 14" iBooks that the 13 systems in question represent, and what the chance is that this deviation from the actual population %'age is due to chance rather than a real estimate of the repair rate.
(and given the inherent bias of the study design towards systems displaying defects this could be an enormously wrong number. 6 of the 7 could be due to chance, if not all 7. )
In a population of millions, 13 systems is even remotely a statistically valid sample, even if it was properly randomly drawn (which is is not).
On a broader scale, the first thing you need to ask yourself, when looking at ANY study like this, on any subject, is "Do the conclusions reached actually make sense?"
There is no way that 43% (the charts say 43%, the text says 41%) of all Apple laptops made during this six year span needed repairs under warrantee.
Apple would have been out of business *decades* ago if their products had a failure rate that high.
"The answer is not 12.445634 because my calculator said so." is the very first lesson you learn in statistics class.
AT BEST, this survey can tell us that the G3 iBooks SEEM to have a higher rate of repairs than other models, *which we already know*, since Apple instituted a special repair program for these models.
As I said in an earlier post, only Apple knows the actual failure rate of any given model because they have the actual warrantee repair numbers, and without a statistically valid sample of laptops to study there's no way we could estimate it either.
-- Bruce Johnson "Wherever you go, there you are" B. Banzai, PhD -- You received this message because you are a member of G-Books, a group for those using G3 iBooks and PowerBooks (we run a separate list for G4 'Books). The list FAQ is at http://lowendmac.com/lists/g-books.html and our netiquette guide is at http://www.lowendmac.com/lists/netiquette.shtml To post to this group, send email to [email protected] To leave this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/g-books Support for older Macs: http://lowendmac.com/services/
