Excel like a finance pro.
Statistical function
CHISQ.TEST
Returns the p-value of a chi-squared test of independence.
When to use it
The p-value of a chi-squared test of independence, comparing an observed table to the expected counts. Low p means the rows and columns are related.
The shape of it
- Syntax
=CHISQ.TEST(actual_range, expected_range)
Worked examples
2 by 2 table
=CHISQ.TEST({30,20;20,30},{25,25;25,25}) → 0.045500264
A 2×2 table where each cell differs from expected by 5: chi-squared 4 on 1 degree of freedom.
From ranges
=CHISQ.TEST(B2:C3,E2:F3) → 0.0455
The same test with the observed and expected tables on the sheet.
Larger table
=CHISQ.TEST(B2:D4,E2:G4) → 0.31
A 3×3 table with 4 degrees of freedom: no evidence of a relationship.
Worth knowing
- Expected counts are row total × column total / grand total; build them with formulas.
- Cells with expected counts below 5 make the test unreliable.
- Degrees of freedom are (rows - 1) × (columns - 1).
Where it goes wrong
- #N/A when the two tables differ in shape.
- Works on counts, not percentages.
Related
Learn the moves here — or let Wauvel run them on your numbers.
Meet your AI CFO →One CFO-grade Excel tip a week
A short, practical email for finance operators — functions, shortcuts, and the moves that save an afternoon. Free, unsubscribe anytime.