alpha = betai ( 0.5*df, 0.5, df/(df+tval**2) )
>
>Thanks for your advice. I want to show that a time series
>of annual temperature data has significant warming.
>I don't believe that I can use regline or regcoef to
>test this, since it's one time series.
>
Yes you can. :-)
The null hypothesis is that the regression coef is zero..
You can use NCL's incomplete beta function, to calculate
the probability.
http://www.ncl.ucar.edu/Document/Functions/Built-in/betainc.shtml
see example 2 [Remember the N returned by the regline/coef func
yields N-2 degrees of freedom]
Reference:
brownlee
statistical theory and methodology
j wiley 1965 pgs: 334-346.
D
_______________________________________________
ncl-talk mailing list
ncl-talk@ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
This archive was generated by hypermail 2b29 : Wed May 25 2005 - 10:03:12 MDT