Covariates for Analyze Definitive Screening Design

Stat > DOE > Screening > Analyze Screening Design > Covariates

Select the columns that contain the covariates to include in the model. You can include up to 50 covariates in your model.

In a designed experiment, a covariate is typically used to account for the effect of a continuous variable that is measurable, but difficult to control. Include a covariate in the model to reduce the error variance. For example, an engineer plans an experiment to study the drying time of paint. The factors in the experiment are different types of binders, different types of diluents, and different types of additives. The engineer wants to control for the effect of the covariate ambient temperature on the drying time of paint.