
 
Dataplot: Tabulated Designs 

Introduction 
The Dataplot distribution comes with a number of data files
that contain many common designs.
The links for these file are to the NIST ftp site. If you have downloaded and installed Dataplot, local copies of the files are available in the "DEX" subdirectory of the Dataplot auxillary directory. For Windows, the default auxillary directory is "C:\DATAPLOT". For Unix/Linux, the default auxillary directory is "/usr/local/lib/dataplot". 

LIST Command 
You may view any of these files within Dataplot by
entering the command
The header lines in the specified will, in most cases, contain the instructions for reading the file into Dataplot. 

Specifying the Location of the Dataplot Auxillary Directory 
If Dataplot cannot find the requested file when you enter
a READ or LIST command, this indicates that the Dataplot
auxillary directory is not installed in the expected
location on your local platform. Contact your local
system installer to determine the location of the Dataplot
auxillary directory on your local platform.
If the Dataplot auxillary directory is not in the default location, you can define the environment variable DATAPLO$ (on Windows and Unix/Linux platforms) to tell Dataplot where the Dataplot auxillary directory is actually located. The Dataplot installation notes contain instructions for defining this variable for Windows and Unix/Linux platforms. For other platforms, contact your local system installer for guidance. 

Categories for the Dataplot Design Files 
The designs are organized by problem/design category.




Comparative Designs  
Completely Randomized Designs 
Because of their intrinsic simplicity (only one factor
and no blocking factors) and their ease of construction,
there is no formal index of Completely Randomized Designs.
All completely randomized designs are defined by three specifications:
l_{1} = number of levels r = number of replications
We present two simple examples of completely randomized designs:


[ Dataplot / Dataplot Designs ]  


Randomized Block Designs 
The following table shows the layout for randomized block
designs.
l_{3} = number of levels (settings) of factor 3 l_{4} = number of levels (settings) of factor 4 l_{k} = number of levels (settings) of factor k
l_{1} = four levels of factor X1 l_{2} = three levels of factor X2 r = one replication per cell n = l_{1} * l_{2} = 4 * 3 = 12 runs


[ Dataplot / Dataplot Designs ]  


Latin Square (and Related) Designs 
The following Latin Square, and related, designs are available:


[ Dataplot / Dataplot Designs ]  


Screening Designs  
Full Factorial Designs 
The following lists the full factorial designs that are available
as builtin data files in Dataplot.


[ Dataplot / Dataplot Designs ]  


Fractional Factorial Designs 
The following table lists the fractional factorial designs that are
available as builtin data files in Dataplot. The diagonal
elements of the table are full factorial
designs that are given above.
Additional discussion of fractional factorial designs is contained in the NIST/SEMATECH eHandbook of Statistics. 

[ Dataplot / Dataplot Designs ]  


Taguchi Designs 
The following table lists the Taguchi designs that are available
as builtin data files in Dataplot.
Note: The L36B design is temporarily unavailable.
Additional discussion of Taguchi designs is contained in the NIST/SEMATECH eHandbook of Statistics. 

[ Dataplot / Dataplot Designs ]  


Plackett Burman Designs 
Dataplot contains builtin data files for the following PlackettBurman designs: Additional discussion of PlackettBurman designs is contained in the NIST/SEMATECH eHandbook of Statistics.  
[ Dataplot / Dataplot Designs ]  


Regression Designs  
Principles in Choosing Design Points 
Regression designs do not have explicit designs.
Designs in the regression context refers to how
one selects the values for the points of the
independent variables.
Although there are not specific designs, there are several principles to use in selecting these points.


[ Dataplot / Dataplot Designs ]  


Optimization Designs  
Some Common Optimization Designs 
Dataplot contains data files for the following optimization
(also referred to as response surface designs):


[ Dataplot / Dataplot Designs ]  


Mixture Designs  
Simplex Lattice Designs 
Simplexlattice designs are defined by the following
parameters:


[ Dataplot / Dataplot Designs ]  


Simplex Centroid Designs 
Dataplot contains data files for the following
simplexcentroid mixture designs:


[ Dataplot / Dataplot Designs ]  


Simplex Centroid Augmented Designs 
Dataplot does not currently provide any builtin simplexcentroidaugmented designs.  
[ Dataplot / Dataplot Designs ]  


Snee Marquardt Designs 
Dataplot does not currently provide any builtin SneeMarquardt designs.  
[ Dataplot / Dataplot Designs ]  


Privacy
Policy/Security Notice
NIST is an agency of the U.S. Commerce Department.
Date created: 06/05/2001 