PSY 380 Lecture Notes - Lecture 7: Inter-Rater Reliability, Takers
8 Mar 2016
School
Department
Course
Professor
Document Summary
Not ideal but good enough correlation coefficient. Consider ability level and educational background of test takers included in test-retest study. Higher baseline levels of ability benefit more from practice alone. Uneven effect of practice depending on composition of sample members. Time varies between time one and time two. Rule of thumb: the more time passes, more likely to be different. Time frame: 2 weeks to 8 weeks. Good enough period of time to make sure practice or exposure effects have time to decline. More problematic with tests developed for children. Relative standing may stay but there are other factors to consider. Test retest reliability is best for estimating reliability of a tool for a stable trait. Mindful and critical of reliability values that are . 9 or higher. Best to demonstrate as many types of reliability and validity as possible. Can be used if too expensive or inappropriate for test retest. Administer test to a group, one time only.
