- Low and High Values for Factors
- For each factor in the list, verify that the type and settings are correct.
- Type: Factor levels can be text or numeric. The initial type matches the data type in the worksheet. If a categorical factor uses numeric levels, use Change Data Type so that the levels are text before you define your design. If a categorical factor has more than 2 levels, you can use Fit Regression Model to analyze the data without defining a design.
- Low: If the factor is numeric, the initial value matches the minimum numeric value in the column. The minimum and the low level can be different if the design contains a botched run where the factor level is less than the level that you want to use as the low level. If your design contains a botched run with a lower level, type the level you want to treat as the low level for the factor. If the factor is text, type the level that should be the low level.
- High: If the factor is numeric, the initial value matches the maximum numeric value in the column. The maximum and the high level can be different if the design contains a botched run where a factor level is greater than the level that you want to use as the high level. If your design contains a botched run with a higher level, type the level you want to treat as the high level for the factor. If the factor is text, type the level that should be the high level.

###### Note

If you mistype the high or low level for a text factor in the worksheet, Minitab Statistical Software warns that text factors must have exactly two levels when you click OK to define the design.

- Worksheet Data Are
- Select how the data are recorded in the worksheet.
- Coded: High and low levels of numeric factors are coded as 1 and −1. Text factors are coded with numbers instead of text. For numeric factors, Minitab Statistical Software interprets values other than 1 and −1 as botched runs.
- Uncoded: High and low levels of the numeric factors are in the units of the data. Levels for text factors are text.