This is a sequence of animations broken into five parts on the topic of Reliability.
  1. Part 1 of 5:
  2. The on-screen text says "Any kind of measuring device, whether it’s a scale, a stopwatch, or a psychological test, should be reasonably consistent. Repeated measurements should yield reasonably similar results. Reliability refers to the measurement consistency of a test or of other kinds of measurement techniques."
  3. Part 2 of 5:
  4. The on-screen text says "Although a test’s reliability can be estimated in several ways, the most widely used approach is to check test-retest reliability, which is estimated by comparing subjects’ scores on two administrations of a test. For example, if we wanted to check the test-retest reliability of a newly developed test of assertiveness, we would ask a group of subjects to take the test on two occasions a few weeks apart."
  5. Part 3 of 5:
  6. The on-screen text says "This chart shows eight individuals' scores on the first administration of the assertiveness test. Click on the diagram to see their scores on a second administration of the same test a few weeks later. If people get fairly similar scores on both administrations of the scale, as shown here, the test measures assertiveness consistently and has high reliability. On the other hand, if people get very different scores on the two administrations of the test, as shown here, the test is an inconsistent measuring device and has low reliability."
  7. Part 4 of 5:
  8. The on-screen text says "A precise reliability estimate requires the computation of a correlation coefficient--a numerical index of the degree of relationship between two variables. In estimating test-retest reliability, the two variables that must be correlated are the two sets of scores from the two administrations of the test. If people get fairly similar scores on the two administrations of the test, this consistency yields a substantial positive correlation."
  9. Part 5 of 5:
  10. The on-screen text says "The closer the correlation comes to +1.00, the more reliable the test is. The reliability estimates for most psychological tests range from the .70s through the .90s. The higher the reliability coefficient, the more consistent the test is. As reliability goes down, concern about measurement error increases. You can click on each of the three examples shown here to see some representative reliability coefficients."
  11. End of animated sequence.