Cronbach's alpha is best described as

• A. A measure of test-retest reliability over time.
• B. A measure of internal consistency reliability for a scale; values range 0-1 with .70+ typical threshold.
• C. A statistic for correlation between two variables.
• D. A method for adjusting p-values.

Answer: (B)

Cronbach's alpha measures internal consistency reliability of a scale, meaning it assesses how well the items on a test or questionnaire work together to measure the same underlying construct. It looks at how closely related the items are to one another; if they tend to move in sync, the alpha value will be higher. The statistic is influenced by both the average correlation among items and the number of items in the scale, so more items with moderately strong inter-item correlations typically produce a higher alpha. Values range from 0 to 1, with a common practical threshold around 0.70 indicating acceptable internal consistency for many research purposes. Caveats: very high values (e.g., above 0.95) can suggest redundancy among items, and alpha assumes the scale is essentially unidimensional—if the items tap more than one dimension, interpreting a single alpha can be misleading. Cronbach's alpha does not assess stability over time (that’s test-retest reliability), it is not a measure of the relationship between two different variables (that would be a correlation coefficient), and it is not a method for adjusting p-values.