Note on the relationship between internal consistency and test-retest estimates of the reliability of a test