Stepwise/Best-Subset Descriptive Statistics Different than Least Squares
Nov 1, 2018 3:38 AM(433 views)
I'm encountering a problem where I find a good model using the Stepwise Regression Control utility, but then when I "Run Model", the same model shows completely different descriptive statistics. Please help!
I think it might be a result of where the missing values are located in your data table. In the Stepwise report, it says that 8 rows were not used due to missing values. But, if those missing values are located in cells in either the "Fill Jobs" or "IQA In-Process" columns, the corresponding rows could still be used in the Fit Model that appears on the left of your screenshot.
Michael Crotty Sr Statistical Writer JMP Development