FieldsPharmaceuticals
Experiments here usually have about 4 factors and 20 runs. Fractional factorials and full factorials are the most common.
- Full factorial
- Fractional factorial or screening
- Central composite or Box-Behnken
- Mixture
- Other designs
Show the numbers
| Kind of design | Experiments | Factors | Runs |
|---|---|---|---|
| Full factorial | 6 | 3 | 29.5 |
| Fractional factorial or screening | 7 | 5 | 16 |
| Central composite or Box-Behnken | 2 | 2.5 | 15 |
| Mixture | 4 | 4 | 14 |
| Other designs | 5 | 6 | 32 |
Factors and runs are the medians of each kind.
Full factorial
Every combination of the factor levels.
- Activity against C. albicans of an antifungal antibiotic from Thermomonospora sp MTCC 3340: carbon source, nitrogen source and temperature, a 3^3 (Problem 9.28)Montgomery (2013) · 3 factors, 27 runs
- Enzyme level in the heart muscle of beagle dogs: three brands of lidocaine, three dosage strengths and three dogs, a 3^3 run twice (Problem 9.4)Montgomery (2013) · 3 factors, 54 runs
- Moisture titrated from a freeze-dried product: chloroform by shake time, duplicatedSchmidt and Launsby (2000) · 2 factors, 4 runs
- Penicillin in surface culture: corn steep liquor, lactose, precursor, sodium nitrate and glucose, a 2^5 in two weeks (and in four blocks of eight)Davies (1956) · 5 factors, 32 runs
- Penicillin in surface culture: corn steep liquor, lactose, precursor, sodium nitrate and glucose, a 2^5 in two weeks (and in four blocks of eight), in 4 blocksDavies (1956) · 5 factors, 32 runs
Fractional factorial or screening
A planned part of the combinations, to find the factors that matter.
- Percentage yield of a reaction in drug manufacture: five factors in a 2^(5-1) resolution V fraction, the path of steepest ascent and its confidence coneBox and Draper (2007) · 5 factors, 20 runs
- Moisture titrated from a freeze-dried product: methanol, chloroform, formamide and shake time in a duplicated half fractionSchmidt and Launsby (2000) · 4 factors, 8 runs
- Yield of a medicinal product: five factors in eight runs, a 2^(5-2) with D = -AC and E = ABDavies (1956) · 5 factors, 8 runs
- Yield of penicillin: five factors in a 2^(5-1) with E = ABCD, in four fermenters confounding BC, BE and CEDavies (1956) · 5 factors, 16 runs
- Biological drug yield: steepest ascent from a 2^(4-1) fractional factorial modelDunn (2026) · 4 factors, 8 runs
Show the rest
Central composite or Box-Behnken
Extra runs that fit curvature, to find an optimum.
- Microgel-hydrogel composite: sodium fluorescein released, central composite design with axial runs at 1.4Dunn (2026) · 2 factors, 10 runs
- The chromatographic exponential function of a capillary electrophoresis separation of ranitidine in a central composite design (J. Chromatography A) (Problem 11.36)Montgomery (2013) · 3 factors, 20 runs
Mixture
The ingredients add up to the whole.
- Becker's ratio-term mixture model against the Scheffé quadratic on the solubility data (Exercise 19.1)Box and Draper (2007) · 4 factors, 14 runs
- Dissolution of 14 tablet blends of binder, disintegrant, surfactant and other ingredients: a Scheffé quadratic and a component-amount fit (Exercise 18.1; Cornell and Gorman 2003)Box and Draper (2007) · 4 factors, 14 runs
- Glass transition temperature of pseudolatex films with three nonionic surfactants: path of minimum response (Exercise 19.7)Box and Draper (2007) · 3 factors, 11 runs
- Solubility of a four-ingredient pharmaceutical mixture on an extreme-vertices design, and its ridge paths under linear restrictionsBox and Draper (2007) · 4 factors, 14 runs
Other designs
Optimal, blocked, split-plot and other designs made for the problem.
- Absorption time of an antibiotic capsule: three technicians, three dosage strengths and four wall thicknesses in four blocks, the split-split-plot design of Example 14.4 (Problem 14.22)Montgomery (2013) · 3 factors, 144 runs
- Stability of riboflavin in light: a 32-run D-optimal screening design in eight blocks (days) of four for six two-level factorsGoos and Jones (2011) · 6 factors, 32 runs
- Penicillin yield: four process variants run in random order within each of five blends of corn steep liquorBox, Hunter and Hunter (1978) · 1 factor, 20 runs
- A random 32-run design in eight blocks of four for six two-level factors, which estimates every effect but poorlyGoos and Jones (2011) · 6 factors, 32 runs
- The 64-run design in 16 blocks of four printed as D-optimal for the main effects and two-factor interactionsGoos and Jones (2011) · 6 factors, 64 runs