BESC1433 Lecture Notes - Lecture 11: Classical Test Theory, Observational Error, Sampling Error
Document Summary
Test results need to be reproducible, consistent (reliable), and meaningful (valid). Reliability refers to the consistency of scores obtained by the same persons when re- examined with: the same test on different occasions: different sets of items measuring the same thing, different conditions of testing. Individual"s possess stable traits (e. g. , intelligence) that can be measured: random errors contribute to measurement scores, the actual score you get on a test will reflect your actual ability plus random error. Observed score = true score + error: random error is expected and accounted for. It"s the test-maker"s job to minimise the sources of random error in scores on their test. Sd (i. e. , sd of errors") = std error of measurement (sem) how much a score deviates from the real" score. 99% level, z = 2. 58: annie scored 100 on the wechsler adult intelligence scale. I might think your tat story is unusual but another rater might disagree.
