Reliability in Testing Is the test or assessment tool consistent and dependable?• Student-related reliability• Rater reliability• Test administration reliability• Test reliability (Brown, 2004, 20-22)
Student-related reliability • Give students consistent materials/time for test preparation • Plenty of time for studying • Consistent test time/conditions (always on Wednesdays)
Rater reliability • For subjective scoring in high stakes tests: Have more than one rater Use rubric and have traning/norming sessions Have outside periodic oversight Keep tests anonymous • For low stakes tests (quizzes): Use rubric Read through all tests before rating
Test administration reliability • Consistent rules for all classes/teachers: Dictionary use Notes/books Strict time limits
How to measure test reliability?One way is through test-retest method:• The same group of students takes the same test twice. (drawback- motivation and washback)What else could we do to test whether thesame students (or very similar students)score the same on the same test twice?
Split half method Have the test split Then measure the This is a good way to into two halves, scores for each half look at reliability, but equal in tasks and it requires that the difficulty. split tests are really equal
Split half method – Your task • What can you tell about the reliability for the following split sections on my test? Joanna 88 95 Jenny 90 87 Asuka 97 95 David 92 88 Annick 74 78 Mercedes 84 89
Checklist for Reliability (1) • Have many independent items • Delete items that do not discriminate between weaker/stronger students • Limit choices – restrict student responses • Keep items clear and unambiguous • Provide clear instructions and examples • Keep type large enough and cleanly copied
Checklist for Reliability (2)• Provide practice with testing format• Keep administration uniform• Have clear answers for all items (as possible)• Provide detailed answer key• Train raters as necessary• Determine acceptable responses before scoring starts• Keep test-takers anonymous John Bunting (2004) presentation in the Course: Testing, Assessment and Teaching- A program for EFL Teachers at UABC. Facultad de Idiomas, UABC
A particular slide catching your eye?
Clipping is a handy way to collect important slides you want to go back to later.