How to set experimental design with replications in Spotfire Statistica DOE

How to set experimental design with replications in Spotfire Statistica DOE

book

Article ID: KB0138130

calendar_today

Updated On:

Products Versions
Spotfire Data Science - Statistica 14.x
Spotfire Statistica 14.x

Description

Using the example data set CellCountA.sta, we show how to set experimental design with replications in Statistica DOE

Environment

OS:  Windows

Resolution

 

  1. Go to Statistica menu bar, select "Statistics | DOE". let's take $2**(K-p) standard design as example. Click ok.

image.png

 

  1. Specify the number of factors in your experiments, for example, 3 factors, each factor has 2 levels, there will be a total of 2^3=8 runs in the experiments. Also decide the number of randomization blocks for your experiment, it could be a single 1 block, 2 blocks or more. Click Ok.

image.png

 

 

  1. In DOE dialog, under "Add to design" tab, specify the number of replications to add for each of your experiment runs. For example, if there are 3 replications for each of the $2^{**}3=8 runs, there will be a total of 8*3=24 observations in the experiment. User would need to add extra 1 replications, and specify 1 as shown in below figure.

image.png

 

  1. In DOE dialog, under "Display design" tab, click on "Change factor names, values, etc" button to specify the factor variable names & level values.  For example, “Temperature “ (52,58), “pH’ (6,9),”Duration” (35,40).

image.png

Change the defaults to match the factor names and values:

image.png

New Values:

image.png

 

  1. In DOE dialog, under "Display design" tab, click on "Summary: Display designs" to show the experimental design that was set. The design table lists the 'Replication" id, "Block" id, and Factor value combinations.

image.png

 

  1.  Once all the experimental runs are done, one can fill up the design table with observed values for each run. For example, variable "DV_1" are observed dependent variable, and there are 3 replicated observations for each combination of the 3 factors. 

image.png

 

 

 

  1. Once the observed data set is ready, go to Statistica menu bar, click on "Statistics | DOE", and e.g., select 2**(k-p) design for analysis. In the DOE dialog, click on "Analyze design" tab to specify the variables for analysis: dependent, independent factors and block variables. Below, the dependent variable, DV_1, is filled with random values.

image.png

image.png

image.png

 

 

  1. Click Ok to proceed with the analysis. In the Analysis dialog, user can tick 2-way or 3-way interactions under "Model" tab to produce the results spreadsheet.

 

 

The output would look like this:

 

image.png

 

 

 

 

 

 

 

 

 

Issue/Introduction

How to set experimental design with replications in Statistica DOE?

Powered by Wolken Software