Home / DOE / Additional Information / Stacked vs. Table
Stacked vs. Table¶
Design format is a choice on the outputs page, Stacked or Table, and it decides how a replicated design is laid out on the worksheet. It changes the shape of the sheet, not the analysis.
The two layouts¶
| Stacked | Table | |
|---|---|---|
| A replicated run | one row per replicate | one row per design point |
| Where the replicate responses go | down the response column, one per row | across, one response column per replicate |
| Replicate column | yes, it says which replicate each row is | not needed |
| Row count for 8 runs replicated 3 times | 24 | 8 |
Why the choice matters¶
Weights and events and sample size columns exist only in the stacked format. The outputs page says so, and the reason is structural: a weight or a sample size belongs to one observation, and in the table format a row holds several observations. Choose Table and the checkbox on each output is greyed out and ignored. A box that was already ticked stays visibly ticked, but no column is written for it.
So a design that needs a weights column, or a binary output entered as events and trials rather than one row per unit, must be stacked.
Converting an existing design¶
Two ribbon entries convert a design sheet in place:
- QXL DOE New > Modify Design > Replicate Format > Convert From Table To Stacked
- QXL DOE New > Modify Design > Replicate Format > Convert From Stacked To Table
Both act immediately, with no dialog. Converting to stacked spreads each design point's responses down the rows and adds the replicate column. Converting to table collapses them back.
Converting a sheet that is already in the chosen format does nothing. The command reports Design is already in table format. or Design is already in stacked format. and leaves the sheet alone.
Converting to table drops the weights and sample size columns, because the table format cannot hold them. Convert back to stacked and the columns are available again, but the values are not restored.