View Single Post
  #101 (permalink)  
Old 24-March-2008, 02:15 PM
Nereid Nereid is offline
Senior Member
 
Join Date: Mar 2004
Posts: 9,644
Default

Quote:
Originally Posted by JimP View Post
[snip]

I did not know what the second step was when I started into this. I have never pursued anything like this before. The high r^2 values made it worth while to look into further. So far the exercise has been worth while. I have and am learning a lot. And yes, I think at this point I need to start defining my analysis.

The Alcyone software and JPL data are highly accurate. I do not know how to quantify the accuracy of the sunspot data. I made the statement before that the sunspot data “is what it is” acknowledging the fact that the older data had a higher level of error in it. The older historical sunspot data is not good. Solanki has stated that there data has a 68 % uncertainty. I have not used this data up to this point and if I do it will be just for a trend line and nothing more. Is this wrong?

I thank you all for your constructive feedback. I came here knowing that I was going Against the Mainstream.
I may have missed it, but have you stated, quantitatively, what the 'null hypothesis' is?

You may also consider creating some fictional input data, using some kind of simple simulation, to explore how the high correlations you found might arise ...