Calculate Cronbach’s alpha (reliability)
Get the free Statistics toolkit
Reliability, measurement, and test-selection templates & worked examples from the Statistical Test Selection Workbook. We’ll email you the download link.
How Cronbach’s alpha works
Cronbach’s alpha measures internal consistency — the extent to which the items on a scale move together, as you’d expect if they tap the same construct. There are two equivalent ways to compute it:
- Standardized alpha: α = k·r̄ / (1 + (k−1)·r̄), from the number of items and the average inter-item correlation.
- Raw alpha: α = (k / (k−1)) · (1 − Σσ²ᵢ / σ²total), from item variances and the variance of the total score.
- It grows with k. Adding items raises alpha even if they’re mediocre — so a high value on a long scale isn’t automatically reassuring.
- Reliability ≠ validity. Consistent items can still measure the wrong thing.
For item-level diagnostics (item-total correlations, alpha-if-item-deleted) and modern alternatives such as McDonald’s omega, run your raw data in jamovi, JASP, R (psych), or SPSS.
Frequently asked questions
What is a good Cronbach’s alpha?
≥ 0.9 excellent, 0.8–0.9 good, 0.7–0.8 acceptable, 0.6–0.7 questionable, < 0.6 poor. Above ~0.95 may mean redundant items. Standards vary by field.
Can alpha be negative?
Yes — a negative alpha usually means some items are reverse-keyed and need recoding, or the items don’t form a coherent scale.
Does a high alpha mean my scale is valid?
No. Alpha is about reliability (consistency), not validity. And it rises with the number of items regardless of their quality.
Where do I get the variances?
Any stats package reports each item’s variance and the variance of the summed total score; sum the item variances and read off the total variance.
Does it store anything?
No. The calculation runs entirely in your browser; nothing is uploaded or saved.