Which is not a reason for using ANCOVA:

Which is not a reason for using ANCOVA:

to control for some pre-existing characteristic of subjects which may affect scores on the outcome variable

in place of a well-controlled laboratory experiment that uses random selection and assignment