Cronbach’s alpha: what it means and how to read it

In the report: Alpha — Cronbach’s alpha: how consistently the items measure the same thing, on the rows where every item is answered.

How to read it

At most 1, and it can be negative. 0.7 is the usual minimum for a research scale, 0.8 good, 0.9 excellent (above 0.95 the items may repeat each other). A low or negative alpha means the items do not hang together: check the coding of reversed items, and whether the items measure one thing at all.

In a paper: "The scale was reliable, α = .84 (7 items)."

In plain words

Say you want to know how much kids like the school canteen. One question is not enough, so you ask five: "Is the food tasty?", "Do you look forward to lunch?", and so on. Before adding the answers into one canteen score, you count Cronbach's alpha and get 0.84. It means: when kids' scores differ, about 84 parts of 100 of that difference come from the one thing all five questions catch, how much a kid likes the canteen, and only 16 from single questions on their own. So the five work like one big question. A low alpha means some question is about something else, or was asked the other way round, "The food is bad", and was not flipped.

How it is computed

α = k / (k − 1) · (1 − Σ s²ᵢ / s²ₜ), s²ᵢ — the variance of item i, s²ₜ — the variance of the row sums

Example

The cell with Alpha is selected in each report; point at a label to see its note.

— the sample data (50 rows) the examples use
— all columns starting with q; includes 'if item deleted' =STATS_ALPHA(Data!A1:T51, "/^q/", TRUE)
— old syntax, 3 reports =STATS_ANALYZE(Data!A1:T51, "pearson(x1,x2)alpha(/^q/)anova(x1,x2)")

Where you see it

STATS_ALPHA — Computes Cronbach’s Alpha — a measure of internal consistency (reliability) for a set of items, such as Likert-type questions.
STATS_ANALYZE — Runs several analyses on one range and returns one report.

Standardized alpha — Alpha from the mean correlation of the items, as SPSS "Based on Standardized Items": as if every item had the same variance.
Alpha if the item is deleted — The alpha (or the standardized alpha) of the remaining items without this one.
Corrected item-total correlation — The correlation of the item with the sum of the other items (the item itself left out of the sum), as SPSS RELIABILITY "Corrected Item-Total Correlation". Empty for an item without spread.

Get it in your sheet

One formula in a cell gives the whole report, Cronbach’s alpha included. ALS Stats is a free add-on for Google Sheets™.

Install ALS Stats from Google Workspace Marketplace (opens in a new tab)