Reliability in testing is best defined as the extent to which scores are

• A. The extent to which scores are consistent across repeated administrations
• B. The extent to which scores measure what the test intends to measure
• C. The extent to which scores are free from measurement error across occasions and scorers
• D. The extent to which scores correlate with unrelated measures

Answer: (C)

Reliability measures the consistency of test scores, reflecting how much of the observed score is due to the true trait rather than random error. The best definition here captures the idea that scores should be free from measurement error across different occasions and across scorers, so they stay stable regardless of when the test is taken or who scores it. In other words, if the same person retakes the test or if different raters score the same response, reliable scores will be similar because they primarily reflect the trait being measured, not random fluctuations or scoring differences.

Consistent scores across repeated administrations are a part of reliability, but reliability also covers consistency across scorers. Validity, not reliability, is about whether the test measures what it’s intended to measure. Correlation with unrelated measures would reflect poor validity and tells us nothing about reliability.