Cronbach Alpha Calculator
Cronbach's alpha from a table of item scores, with the item and total variances printed.
Cronbach Alpha Calculator
Cronbach's alpha from a table of item scores, with the item and total variances printed.
What the page prints, and what it refuses
Every figure is printed with the rule or formula that produced it, and any numerical tolerance is stated on the result rather than implied. The page refuses an input the formula does not define and names the value it read; it prints no answer it cannot support.
The calculation runs entirely in this browser tab; nothing is uploaded.
The page computes Cronbach's alpha from a table of item scores. Four respondents on three items, with item variances 1.6667, 0.6667 and 0.6667 and a total variance of 8.3333, give alpha = 0.96.
Alpha measures how consistently the items move together. The page prints the item variances and the total variance because the coefficient is a ratio of those two quantities, and a reader should be able to check it.
What alpha is, and what it is not
Alpha = k/(k − 1) × (1 − Σs²ᵢ ÷ s²total), with k items, each item's sample variance and the variance of each respondent's total. It rises when items correlate and falls when they measure unrelated things.
It is not a measure of unidimensionality, and the page assigns no small/medium/large band: those cut-offs are conventions that differ by field. The number is arithmetic; the judgement is the reader's.
How to use Cronbach Alpha Calculator
Enter one respondent per line, with that respondent's item scores separated by commas. Every row must have the same number of items.
Press Run. The result prints alpha, the item and respondent counts, the variance sum and the total variance.
What to Expect
Best for
- Checking the internal consistency of a scale
- Comparing alpha before and after dropping an item
- Teaching why item variances matter
Not ideal for
- A single item, where alpha is undefined
- A formative index, where internal consistency is not the goal
- A sample so small that the variances are unstable
What this tool keeps
- Every respondent's scores and their order
What may need cleanup
- Nothing: alpha is a final number
Common errors
- Ragged rows are refused with the row number
- A zero total variance is refused with the reason
Example use cases
- Compute alpha for a four-respondent scale
- Check a published reliability claim
- See alpha fall when an item is unrelated
Sample input
three items, four respondents as shown
Sample output
alpha = 0.96, with item variances 1.6667, 0.6667, 0.6667
Who this is for
- Researchers checking a questionnaire
- Students learning reliability
Frequently asked questions
Is 0.96 a good alpha?
That is a field-specific judgement, so the page does not make it. It prints the number and the variances; whether a scale needs 0.7, 0.8 or 0.9 depends on what it is used for.
Why does alpha use sample variances?
The coefficient is defined on the observed sample, and the n − 1 divisor is the usual unbiased variance. The page states the divisor so the number can be reproduced.
What if an item is reverse-scored?
Reverse-score it before pasting. The page does no recoding, because a recoding rule would be an assumption about the instrument.
Build custom tools for your store
Join the SDK program to create useful ecommerce utilities for your store.