Correcting for hierarchy will now always select hierarchical terms even if they are aliased.
Analysis nodes now have the response ID prepended to them (e.g. ‘R1:’).
Import Data Set now has an option to mark a column as a discrete numeric factor.
The Box-Cox plot legend now defaults to showing the confidence interval.
Multiple fixes for scrolling and selecting data in the design layout.
Fixed the colors on the perturbation legend when categorical factors are present.
Fixed an issue where curvature was incorrectly reported as 0.
Changed the analysis name change interface to a dialog.
Fixed an instance where the legend high and low colors were inverted.
Removed the “Duplicate Graph” menu item when not viewing a graph.
Fixed a failure to converge in REML diagnostics that was causing the Externally Studentized Residuals to be blank.
Fixed a case where the normal probability guide line was not appearing.
Fixed several issues with graphing a categoric factor when one or more levels are not present in the data.
Improved import of data when the response columns are not at the end.
Fixed issue preventing entry of a constraint with a factor id of H or higher.
Placing a reference line at a specific Y value no longer gives a value very close but not exactly at the requested value.
Fixed a crash when using the Modify Design Space tool with a constrained mixture design.
Scrollbars no longer disappear when toggling value/colors in the correlation matrix.
The value or color selection in the correlation matrix will be remembered when changing tabs.
Fixed a bug where inserting a response could cause the optimization to fail.
The correlation table in the Graph Columns toolbox is now exportable.
Separated the preference for one or two-sided intervals in Point Prediction and Confirmation from the preference for the graph intervals.
Fixed a bug related to the one or two-sided interval preferences that was preventing them from working in Point Prediction.
Fixed a crash when importing data that does not vary then trying to analyze it.
Improved selection of multiple cells in the design layout grid.
Added equations used in the mean bias correction to the help documentation.
Fixed a display issue with the graph legend that was causing it to get cut off when placed at the bottom of the graph area.
Fixed some focus issues when using the response data grid in the Confirmation node.
Runs and center point counts are now reported as integers in the Summary node.
Delta/sigma values entered into the response grid for power are now saved with the file.
Right-clicking and choosing “Clear” on a single cell now clears the cell.
Added a warning when using the “Round Design” feature while viewing Coded values. The rounding always applies to Actual values.
Fixed an issue where an optimization goal was flipping from maximize to minimize in rare cases.
Logit transformation limits are no longer allowed to be equal to the high and low values of the response, as it is undefined at those values.
Loading a design with Equation Only response data while using a decimal separator other than a period now works correctly.
Fixed crash in Diagnostics when a Verification row has an individual response cell ignored.
Fixed crash in Simple Sample designs on the Confirmation node.
Fixed a bug where editing an axis name in the Scatterplot affected the Box Plot axis name.
Fixed a crash in the Graph Columns node.
The View menu in the Diagnostics tab now correctly reflects the residual option that is displayed.
Fixed a bug where a categorical factor was blank when using Import Design.
Fixed crash selecting X axis in Box Plot.
Fixed crash analyzing count data with Poisson when there are many zeroes.
Fixed a bug where a categorical factor was changed to numeric when using Import Design.
Removed the “Hide Legacy Designs” preference. These designs are now available in a new “Legacy Design” node.
Removed Constraint node when there are no continuous factors.
Fixed a bug preventing input of the upper limit in optimization when the response has a logit transformation.
Categoric contrast rows in the “Edit Info…” dialog are now updated when rows are changed from Ignored to Normal.
Fixed a bug preventing optimization criteria from being entered when there is an analysis that has not been completed.
The correct combined Fit Summary will now be shown in combined designs with an ignored factor.
Fixed incorrect labels on the interaction plot when there are entire levels of a factor marked as ignored.
Augmenting a factorial design to a Central Composite design will now switch the default analysis model to polynomial.
Requesting a network license in macOS will no longer give an error about a temp directory.
Exporting to a PNG now defaults to an opaque background.
Copy and pasting values from the design layout into a spreadsheet will copy the full precision values instead of what is displayed.
The Coefficients Table now shows p-values for analyses done with logistic regression.
Fixed a crash when clicking on the Confirmation Location tab in certain circumstances.
Fixed a bug with the Import Data feature where assigning categorical factors in reverse order could cause them to be reordered in the final design.
Removed an incorrect simulation warning about missing factors when using the sine function.
Fixed multithreading bug that was causing combined optimal builds to run extremely slow.
Fixed a bug where discrete factor data were changed after cancelling an augment.
Upgraded to RLM 14.1.
When changing a font preference the previous font type will be pre-selected.
Turning on/off the legend now works correctly on the Pareto plot.
Added One Factor RSM tutorial that covers Propagation of Error and Optimization. Updated the remaining tutorials.
The dex_activate.exe utility will now copy over custom ports from the previous license to the new one.
Fixed display error in the factor tool where the input box appeared in the corner.
Changing the name of a response will now change the corresponding analysis name, if the analysis name has not already been modified.
Fixed a bug causing missing intervals in some combined analyses.
High/Low labels on the interaction plot are now draggable.
Tick mark labels on axes with two mixture components no longer overlap.
Using the Import Data tool with categorical data now consistently codes the levels in the order they appear.
The default model for split-plot analyses is now the model that the design was built for.
When adding a new analysis the dropdown will default to the response that is currently selected.
In some graphs the flags could not be expanded. This has been fixed.
In the optimization, when changing from an analysis with a transform to one without, the criteria settings are preserved.
Fixed simulation equation errors when using a non-US locale.
Added analysis name to several reports to distinguish between multiple analyses.
Fixed some multiple-selection and pasting issues in Windows.
Fixed a crash when selecting an empty response from the Evaluation dropdown.
Fixed a bug that caused some valid rows to be ignored when leaving the Design Layout.
Fixed an indexing error that would cause p-values to be out of order in the Coefficients Table.
New runs added via the Modify Design Space augment will now go into a new block.
Added missing labels back to interaction plot when the legend is off.
Flags can now be expanded when showing all responses in the Numerical Optimization graphs.
Labels on the Pareto chart now have backgrounds, and can be clicked to bring to the forefront if they are overlapping.
Support for Big Sur.
Improved visibility in data tables when using a dark background (e.g. macOS Dark Mode).
Deleting observations in the Confirmation node now deletes the correct row.
Fixed a crash when using ctrl+z to undo a duplication of more than 50% of rows.
Using the arrow keys after typing in a cell will now move to an adjacent cell, similar to the behavior of other popular spreadsheet software.
Removed KCV order for double mixture designs.
Fixed several crashes related to deleting analyses.
When columns are rearranged after using Import Data the names are correctly reassigned.
If factor levels are missing when using Import Data, the runs are now ignored.
When the number of blocks changes due to an augment of a design, the block column is automatically shown.
Treatment means no longer show INF Std. Error for a mean model, instead they are blank.
Fixed a crash when clicking on Graph Columns for certain designs.
Fixed an issue with the Y axis changing when the legend is undocked on a histogram.
Importing Data now does a better job of displaying floating point numbers.
Fixed several crashes when using Import Data and the columns are in non-standard order.
Responses with count data can be analyzed with Poisson Regression.
A design space augment wizard is available to expand an existing design.
Editable constraints in the constraint node.
Runs that are outside the constraints are flagged, but can still be used in the analysis.
Runs that are outside the constraints can be moved back in to valid space via the right-click menu.
Importing data is much easier using the Import Data Set option. You just have to paste in your data.
A ribbon has been added to the Design Layout screen for easy access to several features.
You can now round factor values by right-clicking the column header.
Adding verification runs after an analysis has been performed will no longer invalidate the analysis.
Continuous and discrete numeric factors are now indicated in the Design Summary.
You can now create multiple analyses for a single response.
When copying equations to Excel, the response name is now included.
Pearson, Deviance, and Hosmer-Lemeshow goodness-of-fit tests are available for logistic regression.
There is a new preference for the default layout of the Diagnostics tabs.
Control multiple graphs at the same time with the factors tool.
There is a Box and Whiskers plot available in the Graph Columns node.
Perturbation and Trace plots are now colored by factor.
A new option in All Factor graphs will only show factors that are used in the model.
If there are a large number of ticks on a graph only a subset will be shown.
For large designs the Leverage graph is scaled to the maximum value, rather than 1.0.
When the crosshair on an FDS graph is above 80% it changes to black, rather than red.