Definitive Screening Designs with too few degrees of freedom (Q962)

The information in this article applies to:

MODDE^® 12.1

MODDE^® 12

Summary: 

DSD are parsimonious designs and the minimum number of runs is 2k+1 (excluding replicates) for k factors. In MODDE 12 the minimum design alternative was presented resulting in 0 degrees of freedom (Df) for the design part excluding the replicated center. 

Workaround:

Add extra factors but exclude them from the model. 2 extra Df are then added for each extra factor. 
This procedure increases the possibility to detect the interesting factors.

More information:  

In MODDE^® 13 the Definitive Screening Designs are available with optional 2, 4, 6 extra runs. Those alternatives are recommended as they significantly increase the possibility to detect the interesting factors. This can be seen when investigating the estimated POWER for the alternative design selections. 

The original work of Definitive Screening Designs recommends that the hypothesis shall be a detection maximum of k/2 significant factors. 

Ref; Bradely.J , Nachtsheim.C.J , A class of three level designs for definitive screening in the presence of second order effects, Journal of Quality Technology, Vol. 43, No. 1, January 2011.