Contact
12 programs
Explore the site

FISHER Function in Apple Numbers: Transform Correlations

Statistical Advanced Apple Numbers

Learn the Apple Numbers FISHER transformation with domain limits, table examples, Fisher z confidence intervals, and independent calculation checks.

Verification status: Official documentation referenced; example outcomes independently reasoned or arithmetically checked where stated. Not executed in Apple Numbers. How formula examples are checked

Quick answer: =FISHER(0.75) gives approximately 0.972955074527657. The function transforms a Pearson correlation coefficient r, which must be strictly between āˆ’1 and +1, into its Fisher z form. This is useful in correlation interval calculations, but it does not calculate a correlation from a dataset. Apple’s FISHER reference confirms the syntax and the domain. [A1, C1]

Exact syntax and arguments

=FISHER(x-value)
Position Argument Type Required Default Meaning
1 x-value Number value; a literal, numeric expression or numeric-cell reference Yes None Pearson correlation coefficient r, with āˆ’1 < r < 1.

There is one required argument and no optional argument. Apple’s wording is number value, not a collection/range argument. Enter a single numeric result or reference such as B2, not a whole range like B2:B7 as if the function automatically iterated. A multi-row worksheet uses one formula per row. Apple’s argument-types reference distinguishes scalar number values from collection values. [A1, A2]

The mathematical relationship (provided for explanation; Numbers’ internal algorithm is not documented here) is:

z = Ā½ Ɨ LN((1 + r)/(1 āˆ’ r)) = atanh(r).

At r = 0, the transform is 0; as r approaches +1 or āˆ’1, z grows in magnitude and is not restricted to the interval āˆ’1 to +1. A negative input produces a negative transformed value. The FISHER transform differs from a correlation coefficient and should not be interpreted as a correlation. [A1, C1]

Reproducible worksheet: six correlation results

Create a Numbers table named Correlations. Put the text heading Correlation r into B1 and Fisher z in C1. Enter the numeric values shown below into B2:B7, then type the formula from C2 and fill down through C7. The table title is optional for same-table references; if you calculate from another table, Apple documents Correlations::B2 with :: between the table name and cell. [A3]

Row B: Correlation r C: Formula Expected result in C
2 0 =FISHER(B2) 0
3 0.25 =FISHER(B3) 0.255412811883
4 0.5 =FISHER(B4) 0.549306144334
5 0.75 =FISHER(B5) 0.972955074528
6 -0.6 =FISHER(B6) -0.69314718056
7 0.9 =FISHER(B7) 1.47221948958

The values are independently calculated mathematical expectations, not screenshots or native output. Display formatting may round 0.972955074527657 to 0.973; increasing decimal places in the cell makes the distinction visible. The row for āˆ’0.6 shows that FISHER retains the sign while changing the scale. [C1]

A simple cross-table example is:

=FISHER(Correlations::B5)

This expects approximately 0.972955074528 with the example data. Select the source cell while editing the formula to let Numbers insert the table reference rather than guessing the punctuation or sheet name. [A3, C1]

Reproduce Apple’s published checks

Apple documents these expected identities: =FISHER(0.75) ā†’ approximately 0.972955074527657; =FISHER(0) ā†’ 0; =FISHER(FISHERINV(0.5)) ā†’ 0.5. The last one is a round-trip identity: FISHERINV accepts the transformed value and reverses it. Numerical round-off may affect the least significant digits. These are documentation-checked examples, independently supported by math.atanh and math.tanh; they were not run in Numbers. [A1, C1]

Advanced use: approximate confidence interval for a correlation

For a suitably sampled set of independent paired observations from an approximately bivariate-normal population, the usual asymptotic Fisher-z interval treats 1/SQRT(nāˆ’3) as an approximate standard error for the transformed correlation. This is an independent statistical method, not a special built-in Numbers confidence-interval mode and not a guarantee that assumptions are satisfied. [C1]

Create a table Study with B2 = 0.6 (observed correlation) and C2 = 25 (number of independent pairs). Calculate lower and upper approximate 95% limits:

=FISHERINV(FISHER(B2)-1.96/SQRT(C2-3))
=FISHERINV(FISHER(B2)+1.96/SQRT(C2-3))

For these inputs the expected lower limit is 0.268525236321 and the upper limit is 0.804422855475. 1.96 is the usual normal-theory critical value for an approximate two-sided 95% interval. Here n=25 is greater than three, r=0.6 lies strictly inside the accepted domain, and the two endpoints are converted back to the correlation scale with FISHERINV. [C1]

Before using this for scientific reporting: inspect the sampling method and data quality. Overlapping observations, clustered samples, outliers, poor normality or very small n can compromise coverage. This tutorial does not assert that the interval is an exact test or that any particular medical or financial dataset meets its assumptions.

A different advanced scenario is pooling multiple correlation estimates. One may transform each r to Fisher z, average transformed estimates (with justified weights where appropriate), then apply FISHERINV. Do not simply average transformed correlations without considering sample sizes and study design. [C1]

Errors, limits and troubleshooting

Problem or input What to verify Reliable action
=FISHER(1) or =FISHER(-1) Apple requires a strict open interval; endpoints are outside it. Investigate why a coefficient is exactly Ā±1. Do not silently clamp a legitimate result for publication.
A value such as 1.2 Domain violation, even if a cell is formatted as a percentage. Confirm the input is a Pearson coefficient, not a percent, percentage point or Fisher z.
Whole column passed as one argument Apple labels the argument a number value, not a collection. Put one =FISHER(B2) formula in each result row.
Text like "0.5", a blank or an imported value Apple’s page does not describe every coercion rule or resulting error code. Use a confirmed numeric cell; test any dataset-specific missing-value handling in the installed version.
Rounded displayed result Cell formatting controls visible decimals independently of mathematical value. Increase displayed decimals before comparing to the worked example.
Formula name not recognized on an older device This guide is based on the 15.4 reference; earlier-release support is not established by that page. Check local Numbers version and use the in-app function browser.

These are preventative diagnosis steps. No exact error string has been asserted for a value that was not executed in Apple Numbers. For general error navigation see Apple’s error guidance. [A1, A4]

Use FISHERINV to reverse the transform and CORREL or PEARSON to first calculate the actual correlation from paired data. Apple’s function references explicitly connect FISHER, FISHERINV and CORREL, and those official links were opened/checked. [A1]

The formula is mathematically comparable to standard Fisher-z transforms used in statistics, but this guide contains no verified Microsoft Excel, Google Sheets, LibreOffice or imported-workbook execution claim. Field names and future release support must be checked in those products individually.

Platform, version, locale and validation boundaries

This guide uses the English Apple Formulas and Functions Help 15.4; the reference explains formula arguments for Numbers and the wider iWork functions browser, not a live test in any particular installation. Apple’s Numbers 15.4 Mac and iPhone user guides describe the current release and OS requirements (Mac: macOS 26 or later; iPhone: iOS 26 or later). Apple’s release notes say new statistical functions were added in 15.4, but they do not individually certify this formula on older releases. Check Numbers > About Numbers on Mac or the iPhone app version and test on your own device if backwards compatibility matters. [A5, A6, A7, A8]

The formulas here use the English-language identifiers and decimal points; English examples separate multiple arguments with commas. Formula names, numeric punctuation and argument separators are locale-dependent. For example, Apple’s Brazilian Portuguese help translates EXPONDIST as DISTEXPON and shows semicolons as separators with decimal commas; do not create a second English-language function page or silently replace a translated name with an English name in that environment. This regional demonstration establishes a specific documented locale example, not a blanket rule for every locale. [A9]

All displayed results are mathematically derived expected values, with checks in an independent Python calculation harness. The documentation and links were inspected on 2026-10-13 (UTC). No Apple Numbers workbook was executed and no screenshot, production link, Excel import/export behaviour, coercion for every cell type, or installed-version compatibility was tested. Differences in rounding or presentation can occur if the cell’s displayed number format has fewer decimals. [C1]

Apple references and verification methods