PSYC37H3 Lecture Notes - Lecture 4: Classical Test Theory, Standard Deviation, Face Validity
Document Summary
Sources of error variation might be the test itself, test-taker, environment, or how the test was scored: errors refer to the gap measuring the person"s actual knowledge and the way they"re marked. Reliability: degree to which test scores are free of measurement errors: the accuracy, dependability, consistency, or repeatability of test results. Observed score = true score + error. No, but the idea of this concept does exist. Assumption: errors of measurement are random (as opposed to systematic: rubber yardstick analogy: randomly stretches or shrinks each time you take a measurement, sampling theory suggests that the distribution of random errors is bell-shaped, or normally distributed. Degree of spread around the middle (mean) reflects the amount of sampling error. Parallels the concept of standard deviation around a mean. The standard deviation of errors is the basic measure of error is classical test theory = standard error of measurement. Three distributions of observed scores, with varying amounts of error.