Design & Analysis of Mixture Experiments - Design Experiment Tab
Select the Design experiment tab of the Design & Analysis of Mixture Experiments dialog to access options to design a mixture experiment. These options generate Simplex-lattice and Simplex-centroid designs. (To generate designs for constrained mixtures, e.g., with upper and lower bound restrictions on factors, select Designs for constrained surfaces and mixtures from the Design and Analysis of Experiments (Startup Panel) - Advanced tab.) Note that Statistica generates a standard design; the Design of a Mixture Experiment dialog provides numerous options for enhancing the basic design (e.g., by adding replications, etc.). For a discussion of these designs, refer to the Introductory Overview.
- Simplex-lattice designs
- Select the Simplex-lattice designs option button to select a standard simplex lattice design. In this arrangement of design points, m+1 equally spaced proportions are tested for each factor or component in the model:
xi = 0, 1/m, 2/m, ..., 1 i=1,2,...,q
and all combinations of factor levels are tested. The resulting design is called a {q,m} simplex lattice design. For example, a {q=3, m=2} simplex lattice design will include the following mixtures:
A {q=3,m=3} simplex lattice design will include the points:
- Number of factors
- Enter the number of factors, or q, for the simplex-lattice design in the Number of factors box (or adjust the number using the microscrolls).
- Polynomial degree
- Enter the polynomial degree m for the simplex lattice design in the Polynomial degree (m) box.
- Augment with interior pts & centroid
- Select the Augment with interior pts & centroid check box to augment the standard designs (simplex lattice and simplex centroid) with interior points (e.g., see Khuri and Cornell, 1987, page 343; Mason, Gunst, Hess; 1989; page 230). For example, for 3 factors you could add the interior points:
and the overall centroid:
- Simplex-centroid designs
- Select the Simplex-centroid designs option button to select a standard simplex centroid design. Here the design points correspond to all permutations of the pure blends (e.g., 1 0 0; 0 1 0; 0 0 1), the permutations of the binary blends ( ½ ½ 0; ½ 0 ½; 0 ½ ½), the permutations of the blends involving three components, and so on. For example, for 3 factors, the simplex centroid design consists of the points:
- Number of factors
- Enter the number of factors for the design in the Number of factors box.
- Augment with interior points
- Select the Augment with interior points check box to further enhance the standard designs by additional interior points, as described in the context of simplex-lattice designs.