[PVRCNC] More bad SS statistics

Thu, 7 Nov 2002 11:55:56 -0500

Actually, all the U stations that I listed WERE high power. For whatever reason, the 3830 robot asks for power level. There were TWO low power U stations in the listing. I did not include them. 

I stand by my earlier postings, including the proper use of median in this case.

73, Guy.

> 
> From: "Jim, K4QPL" <[email protected]>
> Date: 2002/11/07 Thu AM 11:22:04 EST
> To: "alsopb" <[email protected]>,  "PVRCNC" <[email protected]>
> Subject: Re: [PVRCNC] More bad SS statistics
> 
> My prior comment to the contary, maybe empirical evidence and logic is more
> persuasive:
> 
> Howie and Will are usually pretty evenly matched. Howie, unassisted got more
> Q's. Will, Assisted, got a sweep. Draw your own conclusions!
> 
> (The above logic is similar to that familiar to all college students in the
> 60's:
> 1. Love is blind
> 2. God is love
> 3. Ray Charles is blind
> Therefore: Ray Charles is God.
> 
> ----- Original Message -----
> From: "alsopb" <[email protected]>
> To: "PVRCNC" <[email protected]>
> Sent: Thursday, November 07, 2002 8:08 AM
> Subject: [PVRCNC] More bad SS statistics
> 
> 
> > Guys,
> >
> > To provide an alternate analysis of whether assisted helps or hurts, I
> > submit the following analysis.
> >
> > 1.  Data set - the most recently published 2002 CW SS results #3
> > 2.  Data included:
> > For assisted-- all stations
> > For SO -- all class B stations
> > 3.  Comparison method:  compute mean and standard deviation
> > 4.  Hypothesis to be tested-- On the average assisted operation
> > improves scores.
> >
> > Suffers from the following inadequacies.  No control group
> > established. Sampling is questionable.  Not all U stations may have
> > been high power-- thus the affect of power is possibly a confounding
> > factor.
> >
> > Results:  SO   mean = 97103 standard deviation 48752  20 data points
> >                mean =101639 standard deviation 45719  (ignoring the
> > one 10K outlier)
> >
> > Results:   SOA  mean = 129218  standard deviation 39816   11 data
> > points.
> >
> > Just comparing the means shows a benefit of over 30% for packet.  If
> > one also considers the standard deviations, one can't conclude
> > anything.  The distributions are just too scattered.
> >
> > Note: If one were to include A stations in the SO results, the mean
> > would have been even lower.
> >
> > 73 de Brian/K3KO
> > _______________________________________________
> > PVRCNC mailing list
> > [email protected]
> > http://mailman.qth.net/mailman/listinfo/pvrcnc
> 
> _______________________________________________
> PVRCNC mailing list
> [email protected]
> http://mailman.qth.net/mailman/listinfo/pvrcnc
> 