[Husker] Dynasties

[Paul Dalen]

All but 1 was significant at the .05 level.  For something like this...which isolates a small statistic inside a complex dynamic system, .10 is very precise.  It would have been acceptable, in my opinion, to broaden the parameters to .20.  

One risk of doing an analysis like this is that it implies a level of precision or certainty that just isn't there.  I tried to communicate that in the opening paragraph. 

Parce que je suis plus beau que vous?

On Jan 26, 2013, at 11:32 AM, Theodore Heise <theo at heise.nu> wrote:

> 
> 
> 
> 
> So I'm just now getting time to read this, thanks for sharing your work.
> 
> Overall, most impressive!  That said, I did have a few comments and/or questions.
> 
> How did you happen to choose p=0.10 as the level of signifcance? Were any of the significant differences still siginificant at a level of p=0.05?
> 
> In the discussion of the red zone in section 1.4, you might consider saying something about the 21-30 ranked teams.  Some might argue that ignoring this is a bit biased.
> 
> Again, very nice work!
> 
> -- 
> Theodore (Ted) Heise     <theo at heise.nu>     Bloomington, IN, USA
> 
> 
> 
> 
> 
> On Thu, 24 Jan 2013, Scott Stewart wrote:
> 
>> I concur.
>> 
>> NU lines should be Red and Bammer's lines White. That way they will appear as insignificant on paper as they are in reality. LOL. WDE!
>> 
>> 
>> On Thu, Jan 24, 2013 at 10:11 AM, Nick Chevance <nickchevance at gmail.com>wrote:
>> 
>>> And in classic piling on mode, I too applaud the analysis.  Very nice.
>>> 
>>> My only complaint - Why are the NU lines on the graphs in blue, and Bama in red.  That just seems .... wrong, somehow.
>>> 
>>> 
>>> On Thu, Jan 24, 2013 at 8:13 AM, David Elfering <aroundomaha at gmail.com
>>>> wrote:
>>> 
>>>> I'll ditto what some others have said and applaud. Excellent analysis!
> 
> _______________________________________________
> husker site list
> husker at tssi.com
> http://romaine.tssi.com/mailman/listinfo/husker