DSTDEVP in Collabora Office: Population Standard Deviation

The official DSTDEVP reference was retrieved on October 22, 2026. That saved /latest/ page is labeled Collabora Office 26.04 Help; the URL is rolling, not fixed to a release. The first release, current free-edition entitlement, installed-build behavior and Online/mobile parity remain unestablished. No native Collabora execution supports the examples here.

=DSTDEVP(Database; DatabaseField; SearchCriteria)

All three arguments are required. Include headers in both the database and criteria range. The measured field may be a quoted header, a header-cell reference or a one-based database-column position. Criteria can use other fields: the filter decides which records to use, while DatabaseField decides which values to measure.

Use population standard deviation to describe the complete set of numeric observations you have defined as the population. The function does not check that all planned observations were collected. A smaller population result than a sample result is not evidence of better data quality.

Enter this original dataset in A1:D11. Enter pending as text, and physically leave empty the two cells marked leave empty. Do not type that label or enter an empty-string formula.

Put Batch in F1 and numeric 10 in F2. Use F1:F2 as the criteria range. Enter outputs in an unused area, such as H1 downward. Formulas use English function names and semicolon separators.

=DSTDEVP(A1:D11;"Minutes";F1:F2)

Expected result: approximately 3.265986. The selected numeric values are 4, 8 and 12. Their mean is 8, and squared deviations total 32. Divide by the three numeric observations, then take the square root: sqrt(32/3).

The result is in minutes. DVARP gives the corresponding variance, 32/3 square minutes. Do not take another square root of DSTDEVP’s result.

There are five selected records but only three numeric measurements. =DCOUNT(A1:D11;;F1:F2) gives 5, whereas =DCOUNT(A1:D11;"Minutes";F1:F2) gives 3. Report the missing measurements separately rather than silently treating them as zeros.

For exactly the same data and criteria:

Both use squared deviations totaling 32. DSTDEVP’s population calculation divides by 3 before taking the square root; DSTDEV’s sample calculation divides by 2. Their ratio for this fixture is sqrt(2/3), approximately 0.816497.

This difference comes from the statistic, not a different selection. Compare record counts and measured fields first if the functions disagree by an unexpected amount. Choose the population or sample interpretation based on what the observations represent, not the desired result.

Put Batch and Shift in F4:G4, then numeric 10 and 1 in F5:G5. The conditions share a row, so both must hold.

=DSTDEVP(A1:D11;"Minutes";F4:G5)

Expected result: 2 minutes. The selected values are 4 and 8, with mean 6. Their squared deviations total 8; divide by 2 and take the square root of 4.

The decrease from approximately 3.265986 to 2 reflects a narrower selection. It does not establish that the original batch became more consistent. For criteria spanning several alternatives, follow DSUM’s shared criteria examples and calculate the statistic over the intended records once.

These no-data outcomes are explicitly documented for DSTDEVP, not outputs observed in the application during this review:

• Set F2 to 99: no records match, and the reference specifies #NUM!.
• Set F2 to 30: two records match but neither Minutes value is numeric, and the reference specifies #NUM!.

For a valid zero-spread comparison, set F2 to 20. The numeric observations are 6, 6 and 6, so the independently calculated population standard deviation is 0. The count of numeric observations is three, not zero.

If you see an unexpected zero or error, check both DCOUNT forms, the exact Minutes header, the header row in A1:D11, and the intended criteria range. Do not replace every error with zero: that would make missing measurements look like perfect consistency. Do not import DSTDEV’s documented no-match zero into DSTDEVP.

These controls do not establish behavior for numeric-looking text, formula-empty strings, error-valued fields or every locale. Text-pattern criteria also depend on calculation options; the examples deliberately filter numeric Batch and Shift codes.

• Collabora Help: DSTDEVP contract, syntax and boundaries.
• Shared database arguments and criteria.
• Editorial testing standards.

Row selection, counts, exact variances, square roots and the sample/population ratio were checked independently using rational arithmetic and Python statistics. Source-defined errors are separate documentation claims. No native execution, screenshot or cross-platform certification is claimed.