A sensitivity table shows how one model behaves across a list of trial inputs. Calc's Multiple Operations can calculate that comparison without requiring separate copies of the model for every trial. Its input field must identify the model cell being varied, which is different from the list containing the trial values.

Use a disposable example before applying the command to a complex workbook. Label B1 as the variable input and enter 4. Put =2*B1+5 in B3. Its initial result should be 13. Keep the formula and its input visible together while setting up the comparison.

Map the trial column to the model

Enter 3, 5 and 7 in D2, D3 and D4. Leave E2:E4 empty for the results, and give the columns clear labels above the selected data area. These are trial values for the same input, not three additional inputs in the formula.

Select D2:E4 and open Data, Multiple Operations. Set the Formulas field to B3 and the Column input cell field to B1. B1 is the cell that Calc substitutes while evaluating each value from the vertical trial list.

Complete the dialog and inspect the result beside every trial. The expected values are 11, 15 and 19. Check them independently using the simple expression before treating the table arrangement as established.

Investigate repeated or misplaced results

If every row gives the same unexpected number, inspect the formula reference and the model input reference. Pointing the input field at a cell that the model does not use cannot demonstrate the intended sensitivity. Also check that the selected range contains the trial column and the empty output column in the documented arrangement.

For a real model, choose trial values that are meaningful and label their units. Include at least one case whose result is already known. A table full of changing numbers is not enough if it varies the wrong assumption.

Keep the original model inputs and the trial table clearly distinguished when handing over the workbook. This calculation compares outcomes for values you supplied; it does not find an optimum, predict which trial will occur, or verify that the model's other assumptions are correct.

Reference: LibreOffice documentation.