Re: USHCN Temperature Record of the Week: Fort Sumner, New Mexico

From: Steve Schulin (steve.schulin_at_nuclear.com)
Date: 09/26/04 

Date: Sun, 26 Sep 2004 15:59:15 -0400

In article <Xns956FB8B143682ThomasPalmchellose@212.83.64.229>,
 Thomas Palm <Thomas.Palm@chello.removethis.se> wrote:

> Steve Schulin <steve.schulin@nuclear.com> wrote in news:steve.schulin-
> 1A6567.20272224092004@comcast.dca.giganews.com:
>
> > Gee whiz, everybody here knows that as many as 19% of the surface grid
> > boxes in the CRU surface record showed statistically significant warming
> > over the 1979-2001 period. One-fifth of the globe distinguishable from
> > no trend! Can you believe there's still anybody who doubts the meaning
> > of the data! [Ref: Jones and Moberg, J. Climate 16:206, 2003]
>
> Gee whiz, you've discovered a new statistical toy: the smaller your sample,
> the larger the statistical uncertainty and the harder to get a
> statistically significant result. It's like taking a loaded die with 25% of
> showing a six, and when someone gets the result that it is loaded you chop
> up his test into sequences all so short that the result of each of them is
> indistinguishable from a fair die.

I like the loaded dice analogy, given the cheating ways of the Mann-made
warming crowd. But I think you're misapplying it. It is good to know
that more than 80% of the measurement grids do not show statistically
significant trend over the period 1979-2001.

Very truly,

Steve Schulin
http://www.nuclear.com

Relevant Pages

  • Re: USHCN Temperature Record of the Week: Fort Sumner, New Mexico
    ... > boxes in the CRU surface record showed statistically significant warming ... > no trend! ... statistically significant result. ... It's like taking a loaded die with 25% of ...
    (sci.geo.meteorology)
  • Re: Cheese
    ... ::: the Lyon Diet Heart Study, and why wouldn't we, cheese has never ... :: In the Oxford vegetarian study consumption of eggs and cheese were ... significant, although the trend was. ... P for trend < 0.01 is statistically significant. ...
    (sci.med.cardiology)
  • Re: USHCN Temperature Record of the Week: Fort Sumner, New Mexico
    ... >>boxes in the CRU surface record showed statistically significant warming ... >>no trend! ... statistically significant trends over the 1979-2001 period. ... papers and FAR-SAR-TARs and whatnot I've had the pleasure to read over ...
    (sci.geo.meteorology)
  • Re: USHCN Temperature Record of the Week: Fort Sumner, New Mexico
    ... The CRU dataset, as analyzed over the years by Jones and colleagues, is ... different ways in which daily mean temperature has been calculated by ... > because the trend is not large relative to the variation in the data set. ... If the trend is not statistically significant, ...
    (sci.geo.meteorology)
  • Re: USHCN Temperature Record of the Week: Fort Sumner, New Mexico
    ... Nuke, you sure know how to twist things. ... Surely you realize that anyone can take a time series and calculate a trend ... The important phrase in your note is "statistically significant"*. ... Lots of places have warming trends, but they aren't statistically significant ...
    (sci.geo.meteorology)