Release Notes - SmartPLS

Version 3.3.9, released 2022-03-28

Fixed: Issues with trial.

Version 3.3.8, released 2022-03-27

Fixed: Minor stability issues.
Improved: Upgraded internal libraries.

Version 3.3.7, released 2022-01-23

Improved: Updated translation files.
Improved: Upgraded internal libraries.

Version 3.3.6, released 2022-01-19

Improved: Improved performance of several algorithms.
Improved: Upgraded internal libraries.

Version 3.3.5, released 2021-12-20

Fixed: Deadlocks occurred on some computers during save operations.

Version 3.3.4, released 2021-12-16

Fixed: Log4J Vulnerability fix. Log4J updated to a patched version (2.16.0)
Improved: Upgraded internal libraries.

Version 3.3.3, released 2021-01-11

Improved: Upgraded internal libraries.
Improved: GUI adjustments, for a better display under Windows 10 and MacOSX Big Sur.
Improved: Indicators that contain non-numeric values are marked with an exclamation mark + tooltip in the datafile editor.
Fixed: Display errors and refresh problems under MacOSX Big Sure.
Fixed: Performance issue with PLSC algorithm.
Fixed: Installation problem in the Persian language area.

Version 3.3.2, released 2020-04-21

Improved: Upgraded internal libraries.

Version 3.3.1, released 2020-04-21

Improved: Upgraded internal libraries.

Version 3.3.0, released 2020-04-11

Improved: Updated embedded Java version
Improved: Mac OS installer includes Java

Version 3.2.9, released 2020-01-08

Feature: Descriptive statistics for latent variable scores, residuals, and prediction errors
Feature: Preferences option to hide leading zeros in the results output
Feature: Implementation of consistent PLS-MGA and permutation
Localization: Translation to Urdu
Localization: Update of the existing language versions
Improved: Added specific indirect effects to the PLS-MGA and permutation results
Improved: PLS-MGA two-tailed p-values
Improved: Added intercepts for unstandardized coefficients IPMA results
Improved: Software speed and performance
Improved: Blindfolding missing data treatment (i.e., it does not support pairwise deletion anymore)
Improved: Added construct level results (e.g., R²) in the modeling window of bootstrapping
Improved: Descriptive statistics for indicator data (e.g., group-specific outcomes)
Improved: Excel export of results
Fixed: PLSpredict LV predictions
Fixed: Changing the sort order in the result tables (e.g., by variable name)
Fixed: Pairwise deletion to treat missing values
Fixed: Highlighting of p-values for PLS-MGA results

Version 3.2.8, released 2018-11-22

Feature: Implementation of predictive model selection criteria for PLS and PLSc
Feature: New unique case identifier (i.e., a fixed number for each observation in the dataset, which is useful, for example, when using the casewise deletion option, multigroup or segmentation analyses)
Improved: Inclusion of all specific indirect effects for the mediator analysis
Improved: Color highlighting of significant p-values in bootstrapping, permutation, and multigroup analysis (MGA)
Improved: Results presentation for bootstrapped fit indices
Improved: Software activation procedure
Localization: Translation to Korean (100%)
Localization: Translation to Malay (100%)
Localization: Translation to Polish (>20%)
Localization: Translation to Rumanian (>50%)
Localization: Correction of incompletions and little issues in the different language translations
Fixed: Residual correlation instead of covariance for PLSc
Fixed: Bootstrapping sometimes stopped when using “complete bootstrapping” with repeated indicators due to problems in the fit calculation for these models
Removed: Lohmöller’s initial weighting scheme (complexity reduction)
Removed: Bootstrap sign change options (i.e., individual sign changes, construct level sign changes; complexity reduction)
Removed: Double bootstrapping (increased bootstrapping performance; complexity reduction)
Removed: d_g1 version of the exact fit measures (complexity reduction)

Version 3.2.7, released 2017-09-18

Feature: Linear model (LM) prediction results as naïve benchmark for PLS Predict outcomes
Feature: Mean value as naïve benchmark for Q² value-based assessment of PLS Predict results
Feature: Specific indirect effects added to the results (e.g., useful for mediation analyses)
Feature: Group-specific FIMIX-PLS analysis results
Feature: Reliability pre-specification of composites
Localization: Translation to Arabic (<50%)
Localization: Translation to Indonesian (100%)
Improved: Additional multigroup analysis (MGA) results (e.g., Rho_A, f², etc.)
Improved: Additional permutation test results (e.g., AVE, Cronbach-Alpha, f², etc.)
Improved: Specific indirect effects for the importance-performance map analysis (IPMA)
Improved: Simple slope plots added to consistent PLS (PLSc) results
Improved: Cross-loadings added to consistent PLS (PLSc) results
Improved: Rho_A results added to the graphical presentation of the PLS path model
Improved: Changed order of VIF results in the report (outer model comes first, now)
Improved: Latent variables’ full correlation matrix in reports (instead of lower triangular)
Improved: Basic bootstrapping as default option (instead of complete bootstrapping)
Fixed: Studentized confidence interval’s bias correction
Fixed: p-value computation for the permutation test
Fixed: Correlation matrix in the data preview with missing values
Fixed: Missing complete bootstrapping results of higher-order constructs (repeated indicators)
Fixed: Missing complete bootstrapping results of multigroup analysis (MGA) computations
Fixed: Missing results when using the complete option and double bootstrapping routines
Fixed: Symbol encoding problem in the PLS algorithm with French translation

Version 3.2.6, released 2016-11-13

Improved: : Inclusion of results for Cronbach's alpha

Version 3.2.5, released 2016-11-06

Feature: New PLS Predict Algorithm
Localization: Translation to Italian (100%)
Localization: Translation to Portuguese (100%)
Localization: Translation to French (50%)
Localization: Translation to Chinese (50%)
Localization: Translation to Persian (75%)
Improved: : Higher speed of the PLS algorithm
Improved: : Higher speed of Bootstrapping and other PLS modules
Fixed: The "complete bootstrapping" option
Fixed: Permutation test p-value computation
Fixed: Estimated empirical correlation matrix computation

Version 3.2.4, released 2016-05-02

Localization: Translation to German
Localization: Translation to Spanish
Localization: Translation to Japanese
Feature: Composite reliability roh_A results
Feature: Composite reliability roh_A bootstrap results
Feature: PLS-POS: Option to use the FIMIX-PLS segmentation result as starting partition
Feature: PLS-POS: Weighted average R² as goal criterion and as additional results
Feature: Modeling Window: Option to display f² and HTMT results
Feature: Modeling Window: Option to show blindfolding Q² in the latent variables
Feature: Project Explorer: Option to automatically import sample projects
Feature: New model fit measures: RMSE Theta
Feature: New model fit measures: NFI
Feature: New model fit measures: Model´s Chi² result
Feature: New model fit measures: d_ULS and d_G exact tests of model fit
Feature: Bootstrap results of new fit measures
Feature: Refurbished the HTML results report
Improved: Project Explorer offers shortcuts for recently opened workspaces
Improved: Construct reliability overview table in the results report
Improved: SRMR results for mixed models with composites and common factors
Improved: Excel results report (look & feel and memory consumption)
Improved: Internal results report (e.g., colors, display of total effects, switch between group-specific results)
Improved: IPMA: Additional importance-performance map for standardized total effects
Improved: FIMIX-PLS: Settings, outputs and reports
Improved: Option to rename variables
Improved: Indicator overview view: Variable number included
Improved: Shortcuts on certain actions for the Windows OS
Improved: Reworked licensing dialog
Improved: Software activation process
Improved: Bootstrapping Performance
Fixed: Rendering problems on high resolution displays with a custom DPI setup greater than 100% on Windows
Fixed: MICOM: Results report for single-item constructs
Fixed: Undo action for 'Apply Theme'
Fixed: Use of alignment and match actions does no longr clear displayed calculation results
Fixed: Problems with empirical weights in permutation algorithm
Fixed: Documentation could not be displayed on computers with curious browser setups.

Version 3.2.3, released 2015-10-06

Improved: Rendering of tables. Increased line height for better readability.
Fixed: Improved: backward compatibility for existing workspaces and data.
Fixed: A link in the new Excel export dialog.

Version 3.2.2, released 2015-10-01

Feature: NEW ALGORITHM: Permutation test for multi-group analysis
Feature: MICOM test for measurement invariance (included in the permutation test results)
Feature: Median, standard deviation, skewness, kurtosis and correlations for indicators (shown in data and indicator views)
Feature: Copy & paste options in the dataview allow for an easy export of descriptive statistics and raw data
Feature: Additional operators ("does not equal", "or") for a more flexible definition of data groups
Feature: Additional chart for HTMT in the PLS report with threshold visualization
Feature: Show p- and t-values in combination with path coefficients or outer weights as results in the path model
Feature: Additional option to hide specific results in the path model (e.g., outer model)
Feature: Additional results for PLS-POS (Weighted R², total and indirect effects)
Feature: Define outer model weighting options (Mode A, Mode B, Equal Weights/Sumscores, Pre-Defined Weights)
Improved: Calculation speed (all algorithms 2 times faster now)
Improved: FIMIX-PLS: segments are numbered in descending order of the segment size (segment 1 always is the largest)
Improved: FIMIX-PLS: reduced list of quality criteria (focus on relevant criteria only)
Improved: PLS-MGA reports all results for interaction & quadratic effects, now
Improved: Missing value identification: Empty cells are recognized as missing values
Improved: Columns with text in the dataset can be used as grouping variable
Improved: Excel Export: new options let you customize your Excel output
Improved: Excel Export: memory usage
Improved: Excel Export: changed file extension to *.xlsx
Improved: Tooltips for large column names in the results view
Improved: Image-Export: Remove focus when exporting models as image
Improved: Error messages and highlighting for invalid path models
Improved: "All calculations are finished" message appears when calculation is complete
Fixed: SRMR calculation of factor models with interaction and/or quadratic terms
Fixed: HTMT bootstrap error in path models with purely formative indicators
Fixed: Change of custom color themes does no longer modify the default theme
Fixed: Inconsistencies of labels in different algorithms
Fixed: Calculation of inner model residuals
Fixed: PLS-MGA confidence intervals
Fixed: "Check for Updates" option has been enabled
Fixed: Image export no longer omits file extensions (png, svg) under Windows 8

Version 3.2.1, released 2015-05-05

Feature: Style of path model can now be customized using colors, borders and font sizes (Professional version only).
Feature: New toolbar for easier access to modeling and styling options.
Feature: New 'Duplicate' action for easier cloning of project files.
Feature: New sample model "European Customer Satisfaction Index (ECSI)".
Feature: New 'Install Sample Models' action for supplementary re-installation of the sample models.
Feature: Algorithm settings are now displayed in all reports.
Feature: An alternative label can be maintained for all variables that make use of line breaks.
Feature: TAB key can be used to jump between latent variables of a model.
Feature: Bootstrapping results of the SRMR criterion.
Feature: Residual correlation results.
Feature: New PLSc quality criterion Rho_A.
Feature: Export a path model to a vector based graphic format (.svg).
Improved: Importance-performance map: chart legend entries are sorted.
Improved: Removed yellow message box stating that the 'Student Version' is used.
Improved: Settings of existing moderating and quadratic effects may be changed (double click the variable).
Improved: Introduced tooltip for moderating and quadratic effects that shows all selected settings.
Improved: Changed default look and feel of moderating and quadratic effects and simplified the default name of new effects.
Improved: Default position of interaction effects is now shifted when multiple effects are create for the same latent variable.
Improved: Excel reports now support more than 256 columns.
Improved: Improved: startup time of algorithms when many data groups were defined but not used.
Improved: Clean-up of menu structures (main menu and context menus) and toolbars.
Improved: Complete bootstrapping is now default instead of basic bootstrapping.
Improved: Row colors in result tables keep alternating correctly, even if the table is sorted by a certain column.
Improved: Connections to the one and only indicator of a single item are rendered without arrow.
Improved: Header of confidence interval results.
Improved: Importance-performance map analysis (IPMA) results representation.
Fixed: Blue links could not be clicked on Mac OSX 10.10.3 (Yosemite).
Fixed: Rendering error after import of a CSV file that had an error in the first line (e.g. columns with the same name).
Fixed: HTMT results computation and display issues.
Fixed: PLSc computation problems (interaction terms and "connect all option".
Fixed: Running the double bootstrapping routine on multiple processors
Fixed: Duplicate naming issue in case of renaming an object.
Fixed: PLSc results output issues (e.g., R^2 values) with "connect all option".
Fixed: Problems with less than 10 Bootstrapping resamples.
Fixed: Loadings of formative and single item constructs in PLSc.

Version 3.2.0, released 2015-01-20

Feature: Graphical representation of bootstrap results (i.e. histograms of paths)
Feature: Simple slope interaction plots (moderation)
Improved: Speed of bootstrap routine.
Improved: Separate installation of JRE no longer necessary on Windows machines.
Improved: Calculation for the whole data set may now be skipped, when data groups are analyzed.
Improved: Variables available for group generation are sorted alphabetically.
Improved: Moving or hiding variables after a calculation does no longer clear the calculation results.
Improved: Newly imported data files will be activated by default.
Improved: If a model file is copied, the new model will be opened immediately.
Improved: Calculation results view does no longer scroll to top if results are switched.
Improved: Prefix for generated groups is now optional.
Fixed: Unexpected sign changes of indirect and total effects.
Fixed: Interaction effect computation problems (moderation).
Fixed: Computation of the unstandardized latent variable scores (IPMA).

Version 3.1.9, released 2015-01-20

Fixed: Licensing server connection problems.
Improved: Minor typo corrections.

Version 3.1.8, released 2015-01-18

Improved: Added Java version check to the licensing dialog.

Version 3.1.7, released 2015-01-02

Feature: Support for weighting vector.
Improved: Introduced student version without registration.
Improved: Introduced 30day trial.
Improved: Introduced licensing server support.
Improved: Change file extension for R export to '.rData'.
Fixed: Issues with R² in interaction effect models.
Fixed: Some report table rows were hidden on Mac OSX Yosemite.

Version 3.1.6, released 2014-11-01

Feature: Automatic Update (still experimental).
Feature: Validation errors of data files are now displayed.
Improved: Calculation results explorer now adapts better to the available screen size.
Improved: Easier copy & paste of files in the project explorer.
Fixed: Issues with studentized confidence intervals in bootstrapping procedure.
Fixed: Calculation of indirect and total effects.
Fixed: Bootstrapping procedure for model consisting of formative models only.
Fixed: Bootstrapping procedure for model consisting of single indicator models only.

Version 3.1.5, released 2014-09-03

Fixed: The 'Snap to Grid' function runs, now.
Fixed: Issues with the HTMT criteria.

Version 3.1.4, released 2014-09-01

Improved: Reworked licensing dialog.
Improved: Minor label changes in data file editor.
Improved: Cells in Excel export are formatted as 'numeric' and may now be used for further calculations without reformatting.
Fixed: Problems when sorting tables in reports.
Fixed: A bug that occurred on CSV export of a single table which had been sorted by path relations.
Fixed: An errors that occurred on shutdown.
Fixed: Problems with “Pairwise Deletion” method to treat missing values.
Fixed: Issues with the Blindfolding procedure.
Fixed: A bug that damaged model files on certain Windows machines.

Version 3.1.3, released 2014-07-26

Feature: You can now decide what should happen after a calculation (e.g. open the full report)
Feature: Export a full report to R.
Feature: Export single matrix to R.
Improved: The result navigation now resides in its own view.
Fixed: A "widget is disposed" exception occured during shutdown on some computers.
Fixed: Contents of the "About" dialog where cropped on some Windows computers.
Fixed: About-Dialog now contains a link to the Terms and Conditions.
Fixed: The initial licensing dialog is now resizable.
Fixed: The amount of RAM allocated to the application on startup was to high for some computers.
Fixed: Although not needed, POS algorithm stopped with error when no data groups where defined a priori.
Fixed: Sorting indicators by the number of missing values in the indicator view caused an error.
Fixed: After setup of a new missing value for the data file, the window was not refreshed properly.
Fixed: Improved: caption for calculation runs.