Posted by Ueli (212.28.158.132) on February 05, 2000 at 21:33:24:
In Reply to: Interested CH'ers posted by Miguel on February 05, 2000 at 18:51:49:
Miguel,
Nothing against you personally, but I'm sorry to say the method you propose to find a correlation between attack frequency and weather phenomenas does not follow rigorous statistical requirements. You ask, that the correlations you are looking for, are already made when collecting the data points. As an example: in this way most instances of "huge weather front but no attack" would be missed. The result of such an effort would be rather skewed, with at best anecdotal value.
However,
there are many of us who are keeping a diary of their attacks and it would be great to run these data against all sorts of weather records, supplied by a local weather station. Correlations found in this way would yield meaningful information.
Just my opinion, I could be wrong,
Ueli