Notice
You are now leaving support.minitab.com.
Click Continue to proceed to:
The coded and uncoded units define the factor levels in an experimental design. For example, you want to determine which combination of pressure settings and primer type optimize paint adhesion. The low settings in your experiment (Pressure = 310 and Primer Type = One) are identified by -1 in coded units and the high settings (Pressure = 380 and Primer Type = Two) are identified by 1 in coded units.
| C1 | C2 | C3 |
|---|---|---|
| Pressure | Primer Type | Adhesion |
| 310 | One | 4.52 |
| 380 | One | 4.55 |
| 310 | Two | 5.05 |
| 380 | Two | 4.69 |
| C1 | C2 | C3 |
|---|---|---|
| Pressure | Primer Type | Adhesion |
| -1 | -1 | 4.52 |
| 1 | -1 | 4.55 |
| -1 | 1 | 5.05 |
| 1 | 1 | 4.69 |
By default, Minitab uses coded units to do the analysis. Coded units let you compare the size of the coefficients (on a common scale) to determine which factor has the largest impact on the response. If a design is analyzed in uncoded (or natural) units, it might no longer be orthogonal. Orthogonality lets you estimate model terms independently, making analysis easier because you can remove terms that are not significant without changing the estimates for the terms that remain in the model.
You are now leaving support.minitab.com.
Click Continue to proceed to: