RELIABILITY AND VALIDITY

1. INTERNAL CONSISTENCY ESTIMATES OF RELIABILITY: TWO GENERAL PARTS BROKEN INTO TWO; WHICH ONLY NEEDS TO BE USED WHEN A TEST MEASURES A SINGLE TRAIT

2. TEST-RETEST ESTIMATES OF RELIABILITY: OBTAINED BY ADMINISTERING THE SAME TEST TWO TIMES TO THE SAME GROUP

3. KURDER-RICHARDSON METHODS: THIS TEST DETERMINES WHICH TEST REPRESENTS A SINGLE AND FAIRLY CONSISTENT CONCEPT

3.1. SPLIT-HALF METHODS(ODD-EVEN): DIVIDED INTO TWO HALVES AND CORRELATE THE HALVES ONE ANOTHER

4. INTERNATIONAL CONSISTENCY ESTIMATES OF RELIABILITY: BROKEN INTO TWO PARTS CATEGORIES AND SHOULD ONLY BE USED WHEN THE TEST MEAUSRE A SINGLE TRAIT

5. CRITERION-RELATED VALIDITY EVIDENCE: THE CORRELATIONS OF TEST SCORES WITH EXTERNAL STANDARD OR CRITERION TO OBTAIN A NUMERICAL ESTIMATE OF VALIDITY EVIDENCE

6. CONTENT VALIDITY EVIDENCE: THE PROCESS OF COMPARING A TEST ITEM WITH AN INSTRUCTIONAL OBJECTIVE TO ENSURE IF IT CORRELATES WITH THE OTHER

6.1. CONSTRUCT VALIDITY EVIDENCE: DETERMINED BY FINDING WHETHER THE TEST RESULTS CORRESPOND WITH SCORES ON OTHER VARIABLES AS PREDICATED