Re: Global Warming-Do It With Steam Power!
Author: BOB2
Date: 06-17-2008 - 07:04
Is there a theory of causality from the correlation of the variable? NFL-AFL Supebowl victories and the stock market? Correlation does not imply causality. Auto correlaton and random variability of non-causal variables isn't really a problem, if you are testing a real hypothesis and no how to do the appropriate tests. As is ordering in iterative variables. I alway like to see what the tests for these factors show. Ordering is a big problem, in impact analysis, especially for complex time series models (which chickems came from which eggs, when?). But random alternative correlations still ain't science, no underlying hypothesis is actually being tested, (ad hominen?).
If you think some of these climate input models are difficult to understand and interpret, try long range economic or transportation models, that assume "rationality" in decisions making? And, there is virtually no validation for any of these models for radical variations which result in extrema conditions. I was once involved in modeling impacts of 5-10 dollar gas, it showed no impacts? This is also because many of the more simplistic models aren't validated or lack non linear iterative feedback loops, which can really be quite conforting, simplistic, and wrong.
Doesn't everybody read Playboy for the stories? That's what I tell my wife?
Not the same as reading the stories in Trains? I mostly just look at the pictures in Trains?