# Specify the options for Create 2-Level Factorial Design (Specify Generators)

Stat > DOE > Factorial > Create Factorial Design > Options

## Fold Design

If you selected a fractional factorial design in the Designs sub-dialog box, then you can fold the design. You can also fold the design if you chose a full factorial and then specified generators to add factors to the base design. Usually, you fold a design to decrease aliasing.

Usually, folding is a part of sequential experimentation. First, you run a small fractional design. After you analyze the design, you can fold the design to add runs that decrease aliasing. To fold an existing design in Minitab, use Modify Design.

When you create a design, you can usually maximize the resolution of the design by using a larger fraction instead of folding the design. However, folding on all factors can create a different alias structure. After you create different designs, you can compare the alias structures to determine whether one best meets your needs.

When a design is folded, a new run is added for each run in the base design, with the sign reversed for the factors you are folding on. All other factors are kept at the same level as in the base design. For more information on folding, go to What is folding?.

• Do not fold: Fold on all factors to decrease the aliasing for all of the factors. When you fold on all factors, any word in the defining relation that has an odd number of letters is omitted. For example, the defining relation for a 1/4 fraction design with 5 factors can be I + ABD + ACE + BCDE. Omitting the words with odd numbers of letters gives the new defining relation I + BCDE.

When you fold on all factors, Minitab copies the original set of runs and adds them to the design, reversing the signs of all values in the additional runs.

If all the words in the defining relation have an even number of letters, then folding on all factors adds replicates without reducing aliasing. Minitab displays a warning and does not create the design.

• Fold just on factor: Fold on a single factor to estimate all of the terms that involve that factor independently. When you fold on one factor, any word containing that factor is omitted from the defining relation. For example, the defining relation for a 1/4 fractional factorial design with 5 factors can be I + ABD + ACE + BCDE. If you fold on the factor A, the words containing A are omitted and the new defining relation is I + ACE.

When you fold one factor, Minitab copies the original set of runs and adds them the design, reversing the signs of the factor you folded the design on.

### Fraction

If the base design is a fractional factorial or if the base design is a full factorial and you specified generators to add factors, then you can select which fraction to use. Minitab uses the principal fraction by default. The principal fraction is the fraction where all of the signs for the design generators are positive. Usually, you use a different fraction because one or more combinations of factor levels are impractical to run. For example, the principal fraction always includes the run where all of the factors are at their high setting. The other fractions do not.

• Use principal fraction: The principal fraction is the fraction where all of the signs for the design generators are positive.
• Use fraction number: The fraction number distinguishes the runs in the design from another set of runs that form the same size fraction. The possible fraction numbers depend on the fraction of the design. For example, if the design is a ¼ fraction, then the possible fraction numbers are 1, 2, 3, and 4. In Minitab, the principal fraction is the highest integer of the possible fraction numbers. For example, if the design is a 1/8 fraction, then the principal fraction is 8. In the case of the 1/8 fraction, enter an integer from 1 − 7 to get a different fraction.

## Randomize runs

Select whether Minitab randomizes the run order within each block or stores the design in standard order. Randomization reduces the chances of confounding between the effects of factors in your study with the effects of factors that are not in the study, particularly effects that are time-dependent.

Sometimes, you do not want to randomize the design because factor levels are difficult or expensive to change. Instead, you may want use a split-plot design to minimize the factor level changes. For more information on split-plot designs, go to Split-plot designs in design of experiments.

Base for random data generator
Specify a base for the random data generator so that you can obtain the same random order when you create another design. When you enter the same base, the run order is the same.

After the design is in the worksheet, you can change how you view the design and re-randomize the design. To view a design in standard order or random order, use Display Design. To randomize or re-randomize a design, use Modify Design. For more information, go to Overview for Display Design and Modify Design.