CORREL in Collabora Office: Linear Association and Preserved Observation Pairs

• Introduced: original Collabora release and launch date are not established by the reviewed sources.
• Availability: currently documented in Collabora Office 26.04 desktop help.
• Platform scope: online and mobile deployments are not inferred from this desktop-help release.
• Checked: October 14, 2026. Function-specific documentation and independent calculations; no native Collabora execution.

=CORREL(data1;data2)

Both datasets are required. The Collabora reference identifies the result as a linear correlation coefficient but does not give a complete type-conversion, missing-pair or native error table. These examples use aligned, equally sized numeric columns.

The independent numeric model divides the centered cross-product total by the square root of the two centered square totals multiplied together. This makes the arithmetic inspectable without claiming native execution or equivalence with every other spreadsheet’s invalid-input behavior.

Enter A1:B6:

=CORREL(A2:A6;B2:B6)

Independent expected result: approximately 0.956182887. Mean X is 3 and mean Y is 5. X deviations are -2,-1,0,1,2; Y deviations are -3,-1,-1,1,4.

The paired products are 6,1,0,1,8, totaling 16. Squared X deviations total 10, and squared Y deviations total 28. The coefficient is therefore 16/sqrt(280).

A helper audit can calculate the two means, each row’s centered product and each column’s squared deviation. Those totals reveal how the final normalization is constructed.

The primary dataset has a large positive coefficient, but its exact least-squares line is Y=0.2+1.6×X. Predicted values are 1.8,3.4,5,6.6,8.2, which do not reproduce every observation.

For a separate exact line with the same X and Y values 5,8,11,14,17, the independently expected coefficient is 1. Its slope is 3, demonstrating that correlation and slope are not the same number or unit.

Use SLOPE when the required output is a fitted change in Y per unit change in X.

Multiply every original Y by ten and add 7. The new values are 27,47,47,67,97. The independent correlation remains approximately 0.956182887: the added constant disappears during centering and the positive multiplier cancels during normalization.

Negating every original Y instead gives approximately -0.956182887. The association’s direction reverses while its magnitude is unchanged.

Reversing both columns together also preserves the original coefficient. These are mathematical controls for complete paired numeric data, not a blanket missing-data rule.

Use X -2,-1,0,1,2 with Y 4,1,0,1,4. Every Y is exactly X squared, yet the centered cross-products sum to zero and the expected linear correlation is 0.

The relationship is curved, not absent. A scatter plot can reveal structure that a single linear coefficient misses. Likewise, a high coefficient does not prove that changing one variable causes a change in the other.

Never sort or filter one data column independently and then assume the remaining rows still describe the original pairs. Keep an observation identifier with both measurements. A successful formula result cannot detect every accidental re-pairing.

If one series is constant, the mathematical normalization denominator is zero. The dedicated Collabora section does not specify that native error display, so no particular code is certified here. Empty inputs, unequal lengths, Boolean/text conversion, source errors and extreme numerical precision remain unverified.

PEARSON names the Pearson coefficient explicitly, and RSQ returns its square. Similar numeric purposes do not establish full type/error alias equivalence. None of these calls supplies a significance test or uncertainty interval by itself.

• Collabora Office 26.04: CORREL.
• Editorial testing standards.

Complete function-specific documentation reviewed. Worked outputs are independent statistical models or explicitly documented rules, not application screenshots or native test results. Unverified type, error and numerical boundaries retain their stated scope.