•You would like to restrict patient
variables to those similar to your patient and relevant to
the outcome of interest.
•Uncontrolled variables may affect
the outcome variable in ways that make it difficult
to distinguish their effects from those of the intervention.