<HTML>Stephen Tonkin writes:
---------------------
The example (thanks to Alice Coleman, who first brought it to my
attention, IIRC) that I used to use was that the increase in the
Canadian divorce rate in the late 1960s and early 1970s correlated
perfectly with the increase in the proportion of Canadian households
with stainless steel sinks.
-----------------------
There might well BE a connection between these statistics - all connections are not cause and effect. Reducion to cause and effect is the mistake that is often made (both intentionaly and unintentionally) by statistical propagandists.
Statistitians speak primarily of *relationships* between phenomena. The type of relationship is a secondary excercise - though establishing these is often the goal of investigative work.
The example you cite is an excellent example of a possible correlative effect stemming from a common source phenomena.
There is a large amount of anthropological evidence (much, also based on statistics) which suggests that the divorce rate rises in paralel with female affluence and economic power. Societies in which females have little purchasing power (in relation to males) *tend* to have lower divorce rates than those societies in which there is less economic disparity between the sexes.
It might be reasonable to conclude that, as female purchasing power increased, women's roles at home remained somewhat traditional (statistical analyisis does in fact tend to support this notion). It is therefore logical to expect that greater female influence on household budgets might result in a similar rise in the purchase of material goods associated with traditional female roles (such as stainless steal sinks which are ussually located in the kitchen or laundry).
I would predict that throughout this same period, we should see a rise in the purchase of similar household goods: vacum cleaners, pots and pans, toaster ovens, washers and dryers etc - a proportionally larger increase that we might see in the area of fishing poles and golf clubs.
Your error is forgetting that it is not necissary to establish a "cause and effect" relationship in order to establish a relationship.
Hense our goal: to demonstrate by statistical anylisis that the probability of the OCT arising by chance is so close to the zero point as to render the question "statistically proven."
This does *not* mean we know why the architects paraleled the belt stars on the Giza platue, nor even does it tell us if they themselves were *aware* that that was in fact what they were doing. All it tells us is that the belt star arrangment did not arrise by coincidence at Giza. Somehow, the pattern made its way from the stars into the Egyptian cultural mix to emerge in their architecture. Whether it took an immediate or indirect route from the sky to the ground cannot be established statistically. But that fact that it did make that journey, most certainly can - provided the alignment falls within a sufficiently narrow margin of error.
ISHMAEL</HTML>
