SKEW Function (LibreOffice Calc)

=SKEW(Argument1 [; Argument2 [; ... [; Argument255]]])

There are no default values. LibreOffice 26.8 Help shows up to 255 separate function arguments; a single range can contain substantially more than 255 cells. The function uses ; between arguments in the cited English-language Calc syntax, and : for an inclusive range. A localized Calc installation may use different decimal or list separators and translated names; follow the function tooltip on your installation rather than pasting a formula with incompatible punctuation. The published OpenFormula name is SKEW, and the specified operation is sample skewness. Official references: LibreOffice Help, Statistical Functions Part Five, SKEW section; ODF 1.4 OpenFormula §6.18.67 (both checked 2026-10-13).

The quickest calculation

Enter the following data vertically. Do not type quotation marks around the numbers.

Independently calculated result, matching saved ODS and CSV values: 1.5491313813. Formatting B2 as a number with four decimal places shows 1.5491, not 154.91%. A positive skewness indicates right-side asymmetry; it is not a probability or a percentage. This sample has only five observations, so its estimated shape is sensitive to the single high value.

If your observations occupy a row instead, =SKEW(B2:F2) works the same way. To combine nonadjacent groups, use =SKEW(B2:B20;D2:D20). Those two arrangements are standard syntax examples; the specific 23-case numerical fixture focused on explicit single-range and direct-argument data.

The following examples are recorded in the saved Calc 25.2.3.2 evidence and their numerical values were reproduced independently using the ODF sample-skewness equation.

Each row is a separate dataset, not data that should be appended to the previous row. The sample 2;3;1;6;8;5 was independently checked as an additional six-observation example; its result is not the same as the LibreOffice Help SKEWP example, because SKEWP uses a different population formula. If you enter =SKEW(2;3;1;6;8;5) instead of a range, the result reported in the saved evidence is likewise 0.3190464128.

Why SKEW and SKEWP differ

SKEWP is a separate function. A function name ending in P is not an optional switch inside SKEW. Never use SKEW(...;TRUE) to request population behavior: TRUE would be treated as another argument rather than a mode selector.

For valid numeric observations x1 through xn, define x̄ as the arithmetic mean and s as the sample standard deviation. The ODF 1.4 OpenFormula sample-skewness expression is:

SKEW = n / ((n - 1) × (n - 2)) × Σ [((xi - x̄) / s)^3]

It is defined for at least three numerical observations and nonzero sample variation. Positive and negative deviations are cubed, preserving their signs. Squared deviations, in contrast, are nonnegative and do not cancel each other. The correction factor means SKEW should not be substituted with a formula that divides the third central moment directly by the population standard deviation cubed.

Manual audit of the first example (independent mathematical reproduction; the individual intermediate formulas below were not separate native execution assertions):

1. Enter 1, 2, 3, 4, 9 in A2:A6.
2. Put =AVERAGE(A2:A6) in D2 for the mean and =STDEV(A2:A6) in D3 for the sample standard deviation. Put =COUNT(A2:A6) in D4 for numeric sample size.
3. In B2, enter =((A2-$D$2)/$D$3)^3 and fill it through B6 to calculate each cubed standardized deviation.
4. In D5, enter =$D$4/(($D$4-1)*($D$4-2))*SUM(B2:B6). The independently calculated result is 1.5491313813.

This decomposition is useful for an audit trail. Ordinary analysis should favor the native SKEW function; it is easier to read and less vulnerable to reference mistakes. To check the independent inputs, the AVERAGE and STDEV functions provide those two statistics.

Suppose you measure package weights from two production batches. Put batch A in A2:A101 and batch B in B2:B101. Use =SKEW(A2:A101) and =SKEW(B2:B101) as exploratory measures of asymmetry; separately show =COUNT(A2:A101), =AVERAGE(A2:A101) and =STDEV(A2:A101) for each group. A batch with a high positive SKEW may contain a few unusually heavy packages; investigate the actual values before making conclusions about a process.

Why the comparison needs context: skewness is unaffected by adding the same constant to every observation or multiplying all values by a positive constant. In the saved fixture, 1,2,3,4,9, 101,102,103,104,109, and 10,20,30,40,90 all produced approximately 1.5491313813. Multiplication by a negative constant flips the sign. But estimates from samples of different sizes can have different uncertainty; SKEW alone is not a statistical significance test.

Advanced application: catching insufficient data

This illustrative guard formula avoids displaying a result for fewer than three numeric cells:

=IF(COUNT(A2:A100)<3;"Need at least 3 numbers";SKEW(A2:A100))

It does not guard against three or more identical numbers, which have zero standard deviation. For a more defensive calculation, use:

=IF(COUNT(A2:A100)<3;"Too few numbers";IF(STDEV(A2:A100)=0;"No variation";SKEW(A2:A100)))

These two IF/COUNT/STDEV combinations are editorial illustrations derived from ordinary Calc function semantics, not members of the 23 reported SKEW formula assertions. In a changing spreadsheet, prefer explicit checks like these to silently hiding every error with IFERROR.

Advanced application: spotting tail direction without overinterpreting it

The saved fixture also included [-5,0,1,2,6] (−0.3568646218) and [1,3,5,9,12,30] (1.7236192374). A negative skew does not imply every observation is negative, and a positive skew does not prove a distribution follows a particular distribution family. A value near zero does not automatically mean data are normal: different asymmetric features can cancel in a single skewness statistic. If a decision matters, inspect a histogram or at least the raw sorted observations in addition to reporting the statistic.

LibreOffice Help says text and empty cells within referenced ranges are ignored. The final saved CSV/JSON reports Calc 25.2.3.2 diagnostics illustrating why users should check imported data carefully:

The Boolean finding is a reported 25.2.3.2 result in the final CSV/JSON, not a guarantee about every version, import path or argument form. Do not extrapolate it from Excel’s behavior. In the same saved fixture, directly entered =SKEW("1";2;3;4) produced Err:504, so text in a range and a direct text argument are not necessarily interchangeable.

Practical quality check: after an import, use Calc’s value highlighting or inspect the column format. A cell that looks like 7 can still be text and be silently excluded. COUNT(A2:A100) tells you how many numeric values a referenced range supplies; COUNTA(A2:A100) counts nonempty text too. If these counts differ unexpectedly, validate the source data before relying on SKEW.

Error strings above are reported observations attributed to one Calc 25.2.3.2 build. Different releases, input types, function contexts, or locales may use different error codes or display labels. OpenFormula defines constraints and mathematical semantics, not every UI error string. Do not convert undefined cases to zero without documenting that policy.

LibreOffice Calc: SKEW is documented in online LibreOffice 26.8 Help and ODF 1.4 OpenFormula. The numerical fixtures were reported for Calc 25.2.3.2 on Linux. Existing ODF (.ods) files can contain the standards-based function name, but this run did not test every LibreOffice version, operating system, locale, or historical migration path.

Microsoft Excel: Microsoft documents an SKEW function with sample-skewness semantics. Formula punctuation in Excel’s English-language UI is normally comma-separated (for example, =SKEW(A2:A6) is unchanged for one range). That is a documentation comparison, not an Excel runtime assertion; see Microsoft’s official SKEW help.

Google Sheets: Google’s official help lists SKEW for data skewness. The behavior of text, empty cells, errors, and international separators should be tested on the actual target sheet before migrating critical workflows. This review did not execute Google Sheets; see Google Sheets SKEW documentation.

Function naming: The ODF standard and the LibreOffice Help section use SKEW. The population variant SKEWP is a separate function, not a punctuation or localization alias for SKEW.

• STDEV — sample standard deviation — the denominator in the standardized sample-skewness expression.
• AVERAGE — arithmetic mean — the center used in the calculation.
• KURT — sample excess kurtosis — describes a different aspect of distribution shape, not a substitute for skewness.

Claims were reviewed on 2026-10-13 (UTC). Vendor Help currently labels itself LibreOffice 26.8 Help; ODF 1.4 is the cited standard. The executed program was Calc 25.2.3.2, Linux.

Evidence limitation: Three type diagnostics differ between the saved workbook and final reported output: logical-in-range (0.8885835616 versus 1.2433904445), numeric-text-in-range (0.0880413931 versus 0.8885835616), and direct numeric text (0 versus Err:504). The tables give the final report’s values as reported observations requiring confirmation in your version. All 23 ordinary numerical examples agree across those artifacts and with the independent calculations.