2.42.2.2.2.1 The Factorial Design Dialog Box


Contents

Supporting Information

To open the Definitive Screening Design dialog

Topics for Further Reading:

Design Selection

Factorial Type

Select a factorial type from the list below. Refer to the comparison table for which type should be used for your design

2-Level Factorial A design where every factor has exactly two levels (low and high). Examines all combinations or a fraction thereof to estimate main effects and interactions.
2-Level Split Plot A 2-level factorial design where some factors are hard to change, so runs are grouped by whole plot and only easy-to-change factors vary within each group.
Plackett-Burman A screening design with two levels per factor and run counts in multiples of 4 (e.g., 12, 20, 28). Each run changes roughly half the factors.
General Full Factorial A design where factors can have any number of levels (not restricted to two). Examines all possible combinations of all factor levels.

Generators

Only available when 2-Level Factorial is selected in the Factorial Type

Default Create a standard 2-level factorial designed
Custom Use this when you need a specific alias structure, want to avoid confounding certain effects

Display Available Designs

Click the button for information to select an appropriate design. The table shows the number of factors and runs

Factors

Number of Factors Specify the number of continuous factors you want to include in the design
Number of Hard-to-Change Factors Available only for 2-level split-plot designs

Specify the number of hard-to-change factors you want to include in the design

Factor Table Enter the design information of factors. That includes
  • Name
  • Type
  • Low
  • High

Setttings

2-Level Fatorial Design

Design Selection Table Select a design to view its number of runs and resolution.
Generators Available only when fractional factorial design is selected for custom 2-level factorial design. Defines how additional factors are created from base factor interactions; determines the alias structure and resolution of the design.
Number of Center Points per Block Specifies the number of center points (runs with all factors at their mid-levels) to include in the design. Center points detect curvature in the response without affecting the estimation of main effects or interactions; set to 0 if curvature is not a concern.
Number of Blocks Divides runs into groups to account for nuisance factors like different days or operators.
Define Block by Generators Lets you specify which interactions are confounded with blocks when using multiple blocks.
Number of Replicates Sets how many times each factorial combination is repeated. Adding replicates can increase the statistical power to detect effects but also increase total runs and resource cost.
Fold Design Reverses signs of selected factors. This technique de-aliases specific effects in a fractional factorial design. For example, folding on factor A makes all two-factor interactions involving A estimable (aliased only with higher-order interactions), effectively increasing resolution for those terms.
Randomize Runs Randomizes run order to eliminate bias from time trends or lurking variables.
Random Seed Enter a non-negative integer to control the random number generator

2-Level Split Plot Design

Design Selection Table Provides fractional and full factorial design options. Select a design to view the number of whole plots, subplots, runs, and its resolution.
Number of Whole-Plot Replicates Sets how many times each whole plot (hard-to-change factor combination) is repeated. Replicates add degrees of freedom to the whole plot error term, improving the power to detect hard-to-change factor effects.
Block on Replicates Available only when whole-plot replicates are used. Puts each replicate into a separate block to account for nuisance factors like different days or operators.
Number of Subplot Replicates Sets how many times the subplot (easy-to-change factor) combinations are repeated within each whole plot. Adding replicates increases the statistical power to detect effects but also increases total runs and resource cost.
Randomize Runs Randomizes the subplot run order within each whole plot to eliminate bias from time trends or lurking variables. The order of the whole plots is always restricted by the hard-to-change factors.

Plackett-Burman

Number of Runs Specifies the total number of experimental runs to include in the design. Plackett-Burman designs are available for run counts that are multiples of 4 (e.g., 12, 20, 24, 28, 36, 40, 44, 48 runs). Select the run count that accommodates your number of factors while minimizing experimental effort.
Number of Replicates Sets how many times each factorial combination is repeated. Adding replicates can increase the statistical power to detect effects but also increase total runs and resource cost.
Number of Center Points per Replicate Enter a non-negative integer to specify the number of center points added to each replicate of the design.
Randomize Runs Randomizes run order to eliminate bias from time trends or lurking variables.
Random Seed Enter a non-negative integer to control the random number generator

General Full Factorial

Number of Replicates Sets how many times each factorial combination is repeated. Adding replicates can increase the statistical power to detect effects but also increase total runs and resource cost.
Randomize Runs Randomizes run order to eliminate bias from time trends or lurking variables.
Random Seed Enter a non-negative integer to control the random number generator

Output

Design Table Displays the complete experimental design matrix showing the factor level settings for each run
Coded Design Shows the design matrix using coded units (-1 for low level, +1 for high level) rather than actual factor values. Useful for understanding the underlying structure and for standardized analysis.
Alias Table Available only for 2-level factorial designs

Displays the confounding (aliasing) structure of fractional factorial designs, showing which effects are confounded with each other. Critical for interpreting results from fractional designs where not all effects can be estimated independently.

Interaction Order for Alias Available only for 2-level factorial designs

Specifies the maximum order of interactions to display in the alias table.

Dialog Theme

Save and load the settings of the dialog. Use this to store frequently used factor configurations, design options, or output preferences, then recall them for future designs without re-entering all values.