STDEV in WPS Spreadsheet: Sample Standard Deviation and Reproducible Checks

• Introduced: documented by 2019-05-24 in an official WPS lesson; the first release is not established.
• Availability: the lessons identify WPS Office for Windows Personal, without a numeric build range.
• Scope: exact current supported builds and other WPS clients remain unverified; historical documentation does not establish retirement.
• Checked: October 14, 2026 against function-specific sources. Worked results were independently calculated; no native WPS test was performed.

=STDEV(number1, [number2], ...)

The May 24, 2019 Windows Personal lesson describes 1–30 arguments, which may include a single array or reference. A range is one argument even when it contains many observations. This is the historical lesson’s limit, not a verified maximum for a current WPS build.

The source explicitly names the n-1 method and says text and logical values are ignored. Its wording does not establish every direct-argument conversion, empty-cell, formula-empty or error case. Primary examples therefore use numeric ranges with at least two observations.

The source calls the method “unbiased,” but this guide makes the precise claim that the squared-deviation total is divided by n-1. That does not generally make the resulting square-root estimator unbiased for population standard deviation.

Enter 3, 5 and 7 in A2:A4. In D2 enter =AVERAGE(A2:A4), giving 5. In B2 enter =(A2-$D$2)^2 and fill through B4.

=STDEV(A2:A4)
=SQRT(SUM(B2:B4)/(3-1))

Both independently expected results are 2. The squared distances total 8; dividing by two gives sample variance 4; taking the square root returns the original measurement units.

Keeping the mean reference anchored matters. If each helper row used a different mean cell, the reconstruction would no longer describe one shared dataset.

For the same three observations:

=STDEVP(A2:A4)

The population model divides 8 by 3, giving expected standard deviation approximately 1.632993162. Whether the sample or population formula is appropriate depends on the intended population, not on which result is smaller.

The STDEV lesson incorrectly names STDEVA as its whole-population alternative. The separate official STDEVP lesson explicitly defines the n-denominator population function and recommends STDEV for a sample. This guide uses that product-specific distinction instead of repeating the inconsistent recommendation.

Copy the three observations twice into A2:A7 on another sheet:

3, 5, 7, 3, 5 and 7

The mean remains 5 and the squared-distance total becomes 16. There are six observations, so the expected sample variance is 16/5 = 3.2 and the expected STDEV is approximately 1.788854382.

The result is not 2 even though the same values were repeated. Both the squared-distance total and n changed, but n-1 did not simply double. Accidental duplicate rows can therefore alter a sample statistic without adding any genuinely new information.

For group 1, 3, the expected sample standard deviation is SQRT(2), approximately 1.414213562. Group 7, 9 has the same result.

Pool the four observations 1, 3, 7 and 9. Their mean is 5 and their squared-distance total is 40. The expected pooled STDEV is SQRT(40/3), approximately 3.651483717.

Averaging the two subgroup standard deviations would miss the difference between their means. Retain the underlying observations, or the necessary group counts, means and squared-deviation information, when a pooled calculation is required.

Adding 100 to all values in the primary fixture leaves the expected STDEV at 2. Multiplying all three values by -4 gives an expected STDEV of 8. The sign reversal changes the mean and ordering but not the nonnegative meaning of spread.

For 5, 5 and 5, all squared distances are zero, so the expected STDEV is 0. This multiple-observation control does not establish a singleton or empty-input result; their exact native error behavior remains unverified.

The historical source’s text/logical rule predicts a result of SQRT(8), approximately 2.828427125, for a reference containing numeric 3, numeric 7, text pending, and =TRUE(). Direct quoted arguments and imported errors are separate unverified cases.

STDEV.S has a separately documented later spelling and lesson. Similar sample-purpose descriptions do not prove identical current limits, removal of STDEV, or every import/export behavior.

• WPS Academy: STDEV, 2019-05-24 (Chinese; Windows Personal).
• WPS Academy: population STDEVP definition.
• Editorial testing standards.

Expected outputs are independent calculations, not recorded native results. Current-build limits beyond the cited documentation remain unverified.