The extent to which a test yields consistent results, as assessed by the consistency of scores on two halves of the test, on alternate forms of the test, or on retesting.

• A. Standardization
• B. Reliability
• C. Normal distribution
• D. Stanford-Binet

Answer: (B)

Reliability is about whether a test yields stable, consistent results across different ways of measuring the same attribute. The described methods—comparing scores on two halves of the test, using alternate forms, and retesting after a period—are classic ways to check this consistency. If scores align closely across the halves, across forms, and over time, the test is considered reliable, meaning you can trust that the score reflects the trait rather than random noise or fluke.

Standardization, while important for fair administration and scoring, is about procedures rather than the stability of the results themselves. Normal distribution describes the shape of score spread in a population, not the consistency of results from one administration to another. Stanford-Binet is a specific intelligence test, not a property describing how consistently a test measures something.