At 01:18 PM 3/31/98 -0600, Patrick Larsen wrote:

The large standard deviation leads me to believe
>there are variables other than just a design flaw. These variables may or
>may not include coolant type, water type, flush intervals, driving style,
>climate, stored in winter or not, ever overheated...I'm sure there are many
>more but you get my point.
>
>Note: I am neither a statistician nor a waterboxer expert, these are the
>opinions of a layperson :)
>
I AM a statistician, and my guess is that we can find out why these head
fail with a high degree of certainty. There is a large enough N (number of
cases) to make a statistically valid test possible.  If anyone is serious
about this, here are the obvious variables.  If you can think of any
others, let me know (reply to the list so everyone can see.)  If you think
I've listed some that are not important, say so.  Once we've agreed on the
important variables, I'll structure a survey form and owners can answer it
and someone (not me) can collect and collate the data.  I'll do a
regression, and presto! we'll (maybe) have a better idea as to what goes
wrong.
VARIABLES
Coolant type, pH of water, time between coolant changes, average length of
trip, average speed during average trip (including stops), range of ambient
temperature during driving year.  More?  Fewer?
Ian Cameron
U of Victoria
Victoria, B.C.