Example of creating a 2-level split-plot design
main topic     interpreting results     session command     see also  

A plastic manufacturing company wants to increase the strength of its plastic. The research team identifies additive percentage, agitation rate, and processing time as possible contributors to strength. The temperature at which the batches are baked also impacts plastic strength. To run a completely randomized 24 design requires each combination of the within-batch factor levels to be baked individually at one of the two temperature settings. The team realizes that this process is too time consuming. They decide to use a split-plot design; all 8 combinations of additive percent, agitation rate and processing time are baked at the same time for each temperature setting. This is replicated so that each temperature level is used twice. This results in 32 observations, run in 4 whole plots of 8 subplots each.

1    Choose Stat > DOE > Factorial > Create Factorial Design.

2    Under Type of Design, choose 2-level split-plot (hard-to-change factors).

3    From Total number of factors, choose 4.

4    Click Designs.

5    From Number of hard-to-change factors, choose 1.

6    In the box, highlight the line for Full Factorial with 2 whole plots and 8 subplots.

7    From Number of whole-plot replicates, choose 2. Click OK.

8    Click Results. Under Printed Results, choose Summary table, alias table, design table, defining relation.

9    Click OK in each dialog box.

Session window output

Full Factorial Split-Plot Design

 

 

Factors:          4   Whole plots:            4

Hard-to-change:   1   Runs per whole plot:    8

Runs:            32   Whole-plot replicates:  2

Blocks:           1   Subplot replicates:     1

 

 

Hard-to-change factors: A

 

 

Whole Plot Generators: A

 

 

All terms are free from aliasing.

 

 

Design Table (randomized)

 

Run  Blk  WP  A  B  C  D

  1    1   2  +  +  +  -

  2    1   2  +  +  -  +

  3    1   2  +  -  +  +

  4    1   2  +  +  +  +

  5    1   2  +  -  -  +

  6    1   2  +  -  -  -

  7    1   2  +  +  -  -

  8    1   2  +  -  +  -

  9    1   1  -  -  -  -

 10    1   1  -  -  -  +

 11    1   1  -  +  -  -

 12    1   1  -  +  -  +

 13    1   1  -  -  +  -

 14    1   1  -  +  +  +

 15    1   1  -  -  +  +

 16    1   1  -  +  +  -

 17    1   4  +  +  +  +

 18    1   4  +  -  -  -

 19    1   4  +  -  -  +

 20    1   4  +  +  -  +

 21    1   4  +  -  +  -

 22    1   4  +  +  +  -

 23    1   4  +  -  +  +

 24    1   4  +  +  -  -

 25    1   3  -  -  +  +

 26    1   3  -  -  -  +

 27    1   3  -  +  -  +

 28    1   3  -  +  +  -

 29    1   3  -  -  -  -

 30    1   3  -  +  -  -

 31    1   3  -  -  +  -

 32    1   3  -  +  +  +

Interpreting results

The first table gives a summary of the design: the total number of factors, whole plots, hard-to-change factors, runs per whole plots, runs, whole-plot and subplot replicates, and blocks.

No alias table is shown because no terms are confounded.

Because you chose to display the summary and design tables, Minitab shows the experimental conditions or settings for each of the factors for the design points. When you perform the experiment, use the order that is shown to determine the conditions for each run. For the first 8 runs of this split-plot experiment, Factor A is set high. The first subplot run in the first whole plot has Factor B set low, Factor C set high, and Factor D set high.

Minitab randomizes the design by default, so if you try to complete this example your run order may not match the order shown.