CHITEST Function in Excel: Chi-Square Test (Older Name)

CHITEST Function in Excel: Chi-Square Test (Older Name) 1
⏱ 1 min readUpdated 12 October 2026

StatisticsLevel: IntermediateAvailable in: Excel 2007+

In this article
  1. Syntax
  2. Examples
  3. Example 1
  4. CHITEST in practice
  5. Where you will use it
  6. Worked example
  7. Mistakes people make
  8. Related functions

CHITEST returns the p-value comparing observed and expected counts. It is a compatibility function: Excel keeps it so workbooks from Excel 2007 and earlier still calculate, but Microsoft recommends CHISQ.TEST for new work. Both give the same result.

Advertisement

Syntax

=CHITEST(actual_range, expected_range)
Argument What it means
actual_range Observed counts
expected_range Expected counts

Examples

Example 1

=ROUND(CHITEST(INDEX($H$1:$J$3,A2,0),INDEX($L$1:$N$3,A2,0)),4)

Result: 0.2626. See the worked example table below for more cases.

💡 For new workbooks use CHISQ.TEST: it gives the same answer and is the name Microsoft keeps documenting.

CHITEST in practice

Where you will use it

  • Opening and checking older workbooks that use CHITEST
  • Understanding formulas shared by colleagues on older Excel versions
  • Knowing when to swap in CHISQ.TEST for clarity

Worked example

CaseFormulaResult
1=ROUND(CHITEST(INDEX($H$1:$J$3,A2,0),INDEX($L$1:$N$3,A2,0)),4)0.2626
2=ROUND(CHITEST(INDEX($H$1:$J$3,A3,0),INDEX($L$1:$N$3,A3,0)),4)1
3=ROUND(CHITEST(INDEX($H$1:$J$3,A4,0),INDEX($L$1:$N$3,A4,0)),4)0

Results calculated in Excel. Same result as CHISQ.TEST.

Mistakes people make

MistakeWhat to do instead
Using CHITEST in new filesPrefer CHISQ.TEST: same result, clearer name, and the function Microsoft keeps improving
Mixing old and new names in one modelPick one style per workbook so formulas are easy to audit

CHISQ.TEST

📚 Part of the free Excel course: Beginner → Expert · Try it in the Formula Lab or ask the AI Helper.

✨ Ask AI about this article

Stuck on a step? Ask a question and the AI answers using this article.

Free · AI can be wrong