RSQ in Collabora Office: Squared Correlation and Residual-Based Fit Checks

• 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.

=RSQ(data_y;data_x)

The Collabora reference specifies Y data first and X data second. It calls the result a determination coefficient and relates it to regression analysis. These examples use equally sized, complete numeric columns.

The independent model squares the centered Pearson coefficient. The residual identity shown below applies to ordinary simple least-squares regression with an intercept; it is not asserted for every model forced through the origin or another constrained fitting procedure.

Enter X 2,4,6,8 in A2:A5 and Y 3,5,4,8 in B2:B5:

=RSQ(B2:B5;A2:A5)

Expected result: 0.7. Both means are 5. The centered cross-product total is 14, X squared deviations total 20, and Y squared deviations total 14.

The Pearson coefficient is 14/sqrt(280), approximately 0.836660027. Its square is exactly 196/280=0.7 in this numeric model.

This is not a slope of 0.7 by definition, even though this particular dataset happens to have a fitted slope with the same numerical value. Their interpretations and units are different.

The fitted line is Y=1.5+0.7×X. Predicted values are 2.9,4.3,5.7,7.1, and residuals, observed minus fitted, are 0.1,0.7,-1.7,0.9.

Their squared total is 4.2. The total centered Y square sum is 14:

=1-4.2/14

Expected result: 0.7. In this intercept-containing model, the line reduces squared residuals relative to predicting every Y by its mean.

A descriptive in-sample reduction is not evidence that the same relationship will hold for future observations or that X causes Y.

Negate all original Y values. The coefficient becomes approximately -0.836660027, while RSQ remains 0.7.

For exact positive and negative lines at the same X values, use Y 7,11,15,19 and Y 19,15,11,7, respectively. Both have expected RSQ 1, even though their slopes are +2 and -2.

Report SLOPE or the signed PEARSON coefficient alongside RSQ when direction matters.

Use X -2,-1,0,1,2 and Y 4,1,0,1,4. Centered cross-products sum to zero; expected Pearson correlation and RSQ are both 0.

Every Y still equals X squared. A plot exposes the deterministic curved relationship that the straight-line measure misses. Zero RSQ therefore does not prove that the variables are unrelated in every sense.

Multiply original Y by ten and add 100. The transformed list is 130,150,140,180, and expected RSQ stays 0.7. Positive scaling cancels in the normalized coefficient and a shared offset disappears during centering.

Repeating every complete original pair also preserves 0.7 in the model, because both squared totals and the cross-product total scale together. That invariance does not make duplicate records acceptable; counts, uncertainty and provenance may still be wrong.

Reversing both columns together preserves the same pairs. Sorting just one variable can instead manufacture a different relationship.

RSQ has no argument for a future target value and does not supply a prediction interval, a significance test or a check of the model’s assumptions. STEYX describes the residual scale in Y units, a different aspect of the fit.

Constant inputs make the correlation denominator zero. The dedicated Collabora section leaves its exact native error display, empty or unequal inputs, mixed types, source errors and numerical extremes unverified. Diagnose the data before replacing an undefined result with zero.

• Collabora Office 26.04: RSQ.
• 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.