Test-Retest Reliability

  • Test-retest reliability is perhaps the most commonly used and easiest to understand. There are only two steps:
    1. Administer the same measurement device to the same group of subjects on at least two occasions
    2. Compute a correlation coefficient of the initial test and the retest scores.
  • The higher the correlation (in a positive direction) the higher the test-retest reliability
  • The correlation coefficient obtained through this method is called the reliability coefficient
  • The biggest problem with this type of reliability is what’s called the memory effect. IOW: subjects may recall their answers on the original test and therefore artificially inflate the reliability coefficient.