s-news
[Top] [All Lists]

Dilemma deciding on 'best' process

To: s-news@lists.biostat.wustl.edu
Subject: Dilemma deciding on 'best' process
From: "Nicholas Kormanik" <nkormanik@qwest.net>
Date: Wed, 28 May 2003 13:14:25 -0600
Importance: Normal
Organization: Synomics
Reply-to: <nkormanik@qwest.net>
 
I'm analyzing a number of various processes, trying to find the one that consistently results in the highest outcome.  Below is a subset of results.
Case Mean StDev SEMean  Q1  Median   Q3   SSQ SkewnessKurtosis CV
1    5.07  3.63  0.39  2.61  5.06   6.92  3411   0.43   0.33  0.72
2    5.02  3.66  0.52  3.17  4.63   6.60  1876   1.02   3.29  0.73
3    4.14  3.14  0.50  1.98  4.69   5.28  1042   0.37   1.39  0.76
4    3.50  2.66  0.48  1.86  3.02   5.33   591   0.67   0.68  0.76
5    3.70  2.89  0.32  1.18  3.60   5.96  1778   0.04  -0.69  0.78
6    3.49  2.78  0.43  1.16  3.40   5.66   829  -0.62   0.65  0.80
7    3.76  3.00  0.55  1.20  3.85   5.85   683   0.24  -0.64  0.80
8    4.91  3.95  0.37  2.57  4.43   6.59  4591   1.22   2.28  0.80
9    7.67  6.17  0.94  2.45  8.23  11.84  4126  -0.02  -0.68  0.80
10   4.11  3.39  0.24  1.89  4.31   5.92  5457   0.13   2.54  0.82
11   4.11  3.40  0.44  2.25  4.21   6.26  1695   0.08   0.40  0.83
12   6.28  5.32  0.89  1.86  6.20  10.55  2410   0.38  -0.02  0.85
13   3.76  3.19  0.36  1.41  4.22   5.79  1910  -0.03  -0.40  0.85
14   3.35  2.85  0.27  1.42  3.42   5.30  2215  -0.01   0.27  0.85
15   7.50  6.42  0.90  3.42  7.68  12.57  4925  -0.44   0.37  0.86
16   3.85  3.34  0.38  1.10  4.36   5.53  1965   0.12   0.12  0.87
17   6.14  5.38  0.98  2.47  4.76   7.25  1972   1.24   1.19  0.88
18   6.90  6.05  0.67  1.37  6.08  11.28  6871   0.64  -0.46  0.88
19   7.11  6.25  0.81  2.24  7.58  11.23  5340   0.06  -0.49  0.88
20   5.75  5.06  0.96  1.12  5.61   8.63  1615   0.99   1.90  0.88
My tentative hunch is to use the last column --- Coefficient of Variation --- as the over-riding criterion.  Thus case number 1 is the best.
 
The first-impression-approach of simply using the highest Mean, case number 9, would appear to be a grave mistake, as the dispersion of outcome is so great.
 
Please comment on the above thinking.  Do you agree?  Or am I disregarding something important?
 
Thank you for your time and attention.
 
Nicholas
 
Salt Lake City, Utah
 
 
 
<Prev in Thread] Current Thread [Next in Thread>