How to use
- Copy only the items of one scale or subscale, with one row per respondent and one column per item. A header row is optional.
- Tick any reverse-scored items and check the lowest and highest answer options (for example 1 and 5).
- Read alpha and its rating, then look at items marked ⚠ in the item-total table.
- Copy the APA line.
Options explained
- Alpha
- α = k/(k−1) × (1 − sum of item variances ÷ variance of the total score), where k is the number of items. The more consistently the items move together, the closer α is to 1.
- Corrected item-total correlation
- The correlation between an item and the total of the other items. Items below about 0.3 may be measuring something else.
- Alpha if item deleted
- Alpha for the remaining items without this one. If it's higher than the current alpha, the item is lowering reliability.
- Reverse scoring
- Negatively worded items such as "I rarely feel calm" must be flipped (1↔5, 2↔4) before calculating, or alpha comes out low or even negative.
Worked example
The sample has 10 respondents answering five 5-point items:
| Statistic | Value |
|---|---|
| Cronbach’s α | 0.868 |
| α based on standardized items | 0.872 |
| 95% CI | 0.672 – 0.963 |
| Mean inter-item correlation | 0.576 |
| Item | Mean | SD | Item-total r | α if deleted |
|---|---|---|---|---|
| Q1 | 3.0 | 1.155 | 0.764 | 0.822 |
| Q2 | 3.0 | 1.247 | 0.835 | 0.802 |
| Q3 | 3.2 | 1.229 | 0.596 | 0.865 |
| Q4 | 3.2 | 0.919 | 0.675 | 0.849 |
| Q5 | 3.0 | 1.247 | 0.627 | 0.858 |
α = 0.868 is good. Every item correlates above 0.3 with the rest and dropping any of them would lower alpha, so all five stay. With only 10 respondents the confidence interval is wide (0.67 to 0.96); a real study needs a much larger sample.
APA style: Cronbach’s α = .868 (5 items, n = 10).
Fixing a low alpha
- Check whether a reverse-worded item was left unticked.
- Read items with an item-total correlation below 0.3 for ambiguous wording.
- Alpha depends on the number of items: with only two or three items it tends to be low, so also report the mean inter-item correlation.
- When exporting from Google Forms or other survey tools, export the numeric codes (1, 2, 3…), not the answer text.
Frequently asked questions
What is a good Cronbach's alpha?
Common rules of thumb are 0.7 acceptable, 0.8 good and 0.9 excellent; exploratory work sometimes accepts 0.6. Above 0.95 often means items repeat each other.
My questionnaire has several subscales. How do I calculate alpha?
Calculate alpha separately for each subscale using only its items. You can also report alpha for all items together.
Does it match SPSS?
Yes. Alpha, standardized alpha, corrected item-total correlations and alpha if item deleted match SPSS Analyze → Scale → Reliability Analysis with "Scale if item deleted" ticked.
Why is my alpha negative?
Usually a reverse-worded item wasn't reverse-scored, or some items correlate negatively with the rest. Look for a negative item-total correlation.
Is my survey data uploaded?
No. The calculation runs in your browser.
Privacy
This tool runs entirely in your browser. Your files and text are processed on your device and are never uploaded to our servers.