PlanMaker COVAR: Population Covariance and the COVARIANCE.P Name
Reconstruct PlanMaker population covariance, compare COVAR with the identical COVARIANCE.P name and inspect unit scaling and zero covariance.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
COVAR returns population covariance: the average product of the two variables’ deviations from their respective means. It is a signed quantity, with units formed by multiplying the two measurements’ units.
SoftMaker describes COVARIANCE.P as an identical function that supplements COVAR. COVAR remains documented; the reference does not identify it as retired. COVARIANCE.S answers the different sample-covariance question.
Availability and version history
- Introduced: Documented by the PlanMaker 2010 manual; the first introduction release is not established.
- Availability: Documented in PlanMaker 2026; this is not a final working-version endpoint.
- Scope: The reference belongs to the 2026/NX manual family. Individual FreeOffice and mobile-build availability remains unverified.
- Checked: October 14, 2026, against the official function reference. Examples use vendor results and independent calculations; no native PlanMaker test was performed.
Syntax and documented input limits
=COVAR(Range1, Range2)
Both arguments are required references. The current PlanMaker entry says the ranges must contain integers, that empty cells, text and logical values are ignored, and that different numbers of values produce #N/A.
The older 2010 description and the 2018 manual do not state that integer restriction. This guide keeps inputs inside the current wording rather than claiming fractional measurements are accepted in every build. The source also does not fully specify how differently positioned ignored entries affect alignment. Use complete paired numeric rows for the calculations below.
Formulas use English names and comma separators. Adapt the argument separator to your configuration.
Rebuild the official example
Create CovariancePairs with x and y in A1:B1. Enter the reference’s data in A2:B4:
| Row | A: x | B: y |
|---|---|---|
| 2 | 2 | 2 |
| 3 | 5 | 7 |
| 4 | 3 | 4 |
Enter =COVAR(A2:A4,B2:B4) in H2. The independently calculated result is 23/9, approximately 2.555555556, which rounds to the manual’s 2.55556.
Put =AVERAGE(A2:A4) in H4 and =AVERAGE(B2:B4) in H5. These give 10/3 and 13/3. Put dx, dy and Product in C1:E1. In C2:E2 enter =A2-$H$4, =B2-$H$5 and =C2*D2; fill through row 4.
| Row | dx | dy | Product |
|---|---|---|---|
| 2 | -4/3 | -7/3 | 28/9 |
| 3 | 5/3 | 8/3 | 40/9 |
| 4 | -1/3 | -1/3 | 1/9 |
These fractions describe expected results, not additional literal inputs. Keep the formulas’ full precision. In H7 enter =SUM(E2:E4); expect 23/3, approximately 7.666666667. In H8 enter =H7/3; expect 23/9, matching H2. Population covariance divides by the three paired observations.
Compare the old and explicit population names
On CovariancePairs, enter:
| Cell | Formula | Expected result |
|---|---|---|
| H10 | =COVARIANCE.P(A2:A4,B2:B4) |
2.555555556 |
| H11 | =COVARIANCE.S(A2:A4,B2:B4) |
3.833333333 |
| H12 | =COVAR(B2:B4,A2:A4) |
2.555555556 |
The first result matches COVAR as SoftMaker explicitly documents. The sample result divides the same cross-product total by 3−1, rather than 3. Swapping the complete paired series leaves covariance unchanged.
SoftMaker’s note about Microsoft Excel 2010 or later concerns the newer dotted function names in Excel. It does not establish when PlanMaker first introduced them. Use the population or sample model because it fits the analysis, not merely because one name is newer.
Covariance depends on units, not only association
Keep the original table intact. Put x times 100 and Negative y in J1:K1. Enter =100*A2 in J2 and =-B2 in K2; fill both through row 4.
- M2:
=COVAR(J2:J4,B2:B4)gives approximately 255.555555556, 100 times the original covariance. - M3:
=COVAR(A2:A4,K2:K4)gives approximately −2.555555556. - M4:
=CORREL(A2:A4,B2:B4)gives approximately 0.997176465. - M5:
=CORREL(J2:J4,B2:B4)gives the same 0.997176465.
Multiplying one measurement by 100 changes covariance’s units and magnitude. The unitless correlation remains unchanged under this positive scaling. A larger covariance after a unit conversion does not mean a stronger relationship.
Zero covariance is not proof of no relationship
Create CurvedPairs, with −2, −1, 0, 1, 2 in A2:A6 and 4, 1, 0, 1, 4 in B2:B6. Put x and x squared in the header row. Enter =COVAR(A2:A6,B2:B6) in D2; expect 0.
Every y value is exactly x squared, yet the positive and negative cross-products cancel. Covariance measures a particular linear co-variation summary. It does not establish independence or causation.
Troubleshooting and related functions
#N/A is explicitly documented for unequal value counts. The short page does not enumerate every native error, empty-only case, fractional-input behavior or extreme precision boundary. Keep observation identities together and inspect ignored or imported cells before interpreting a result.
- COVARIANCE.P for the explicitly named population function.
- COVARIANCE.S for sample covariance.
- CORREL for unitless linear correlation.
- AVERAGE for the centering calculation.