File > Options > Linear Models > Coding of Predictors
Specify how to code categorical predictors separately for regression models and Fit General
Linear Model. Regression models include the following analyses:
Fit Regression
Model
Fit Binary Logistic
Model
Fit Poisson
Model
Linear
Regression
Binary Logistic
Regression
You can also choose whether to standardize continuous variables. The changes you
make to the defaults remain until you change them again, even after you exit Minitab.
Coding for categorical predictors in Regression
Effects: Estimate the
difference between each level mean and the overall mean.
Binary: Estimate the
difference between each level mean and the reference level's
mean.
Coding for factors in General Linear Model
Effects: Estimate the
difference between each level mean and the overall mean.
Binary: Estimate the
difference between each level mean and the reference level's
mean.
Standardize continuous predictors
Do not standardize: Use your original
data for the continuous predictors.
Minimum coded as -1; maximum coded as 1: Use this option to
transform the data linearly. The minimum value in the sample is
coded as -1. The maximum value in the sample is coded as +1. The
remaining data are coded to fall between -1 and +1.
Subtract the mean, then divide by the standard deviation: Center the
predictors and put them on a comparable scale.
Subtract the mean: Center the
predictors.
Divide by the standard deviation: Use a comparable
scale for all predictors.
