Paste your raw data or use the formula (k items, mean r̄) and get α instantly, with interpretation, quality cutoffs and what to do if it comes out low. Free.
Cronbach's alpha estimates the internal consistency of a scale: how far its items measure the same thing. It is not an index of validity or of unidimensionality, and it does not tell you whether the scale measures what you claim it measures.
| Alpha range | Interpretation | Recommendation |
|---|---|---|
| > .90 | Excellent | Excellent reliability. May indicate redundancy if very high (> .95). |
| .80 – .90 | Good | Good reliability. Suitable for research and clinical practice. |
| .70 – .80 | Acceptable | Acceptable for exploratory research. Review weak items. |
| .60 – .70 | Questionable | Low reliability. Its use must be justified and the scale reviewed. |
| .50 – .60 | Poor | Insufficient for most purposes. A thorough revision is needed. |
| < .50 | Unacceptable | Not recommended. Rebuild or replace the scale. |
Mode A, from summary parameters (generalised Spearman-Brown): alpha = (k · r̄) / (1 + (k − 1) · r̄), where k is the number of items and r̄ the mean inter-item correlation.
Mode B, from raw data (classical formula): alpha = (k / (k − 1)) · (1 − ΣSi² / ST²), where ΣSi² is the sum of the item variances and ST² the variance of the total scores.
Item analysis: if «alpha if item deleted» is higher than the overall alpha, dropping that item would improve reliability. Items with a corrected item-total correlation below .30 contribute little and should be reviewed.
George, D., & Mallery, P. (2003). SPSS for Windows Step by Step: A Simple Guide and Reference. Allyn & Bacon. Nunnally, J. C. (1978). Psychometric Theory (2nd ed.). McGraw-Hill.