# Release Notes

The changes in each new version of Q are listed below.

## 5.17.5.0

- Bug fixes.

## 5.17.4.0

- Bug fixes.

## 5.17.3.0

- Bug fixes.

## 5.17.2.0

- Bug fixes.

## 5.17.1.0

- Date period filter terms now include "within last" to allow selecting a time period to filter by in addition to "within last period"
- The Javascript formula option has been returned to the weighting dialog.
- Bug fixes.

## 5.16.5.0

- Bug fixes.

## Online Feature Release

- An update to the design-based Akaike Information Criterion (dAIC) for weighted Regression - Linear Regression models. The updated criterion accounts for sampling weights as before but computes the penalty using all regression parameters including the variance parameter. This update is live from Nov 1st and will align with the to be released version 4.3 of the R survey package on CRAN, https://r-survey.r-forge.r-project.org/.

## Online Feature Release

- Removed
**Automate > Browse Online Library > Data > Google trends**from the menu.

## 5.16.4.0

- Bug fix.

## 5.16.3.0

- Bug fixes.

## Online Feature Release

R version 4.3.1 has been released for Q users! Upgrades like this are important to keep all of the R-based features up to date, and in order for us to stay up to date with the latest security patches. As with all system upgrades, there can be small numerical differences in outputs due to changes in randomization and numerical precision. Q's advanced analysis tools have been thoroughly tested with this upgrade, and no differences of commercial relevance have been identified. If you notice a change that is concerning, please contact support@q-researchsoftware.com. Features in which you may notice small differences include:

- Dimension Reduction
- Regression
- Choice Modeling
- Max Diff

The new version of R is more conservative in some small ways to help R users to write better code. Changes to the way you write code are documented in this article. If you write custom R code in Q and have not yet had a chance to review this, then please do so.

Finally, the Brand Health Table may encounter an error in some cases. A tool has been added to Q to address this problem. Please see this article if you run into any problems with the Brand Health Table.

## 5.16.2.0

- Bug fixes.

## 5.16.1.0

- Added translate functionality to text coding dialog. Allows users to translate text responses to their chosen language via Google's API
- Revamped custom footers dialog
- Speed improvements in tables using banners
- Bug fixes.

## 5.15.3.0

- Bug fixes.

## 5.15.2.0

- Bug fixes.

## 5.15.1.0

- Improved table caption editing dialog
- Security improvements
- Bug fixes.

## Online Feature Release

- Expanded feature for
**Automate > Browse Online Library > Table Computations > Difference between pair of columns**and**Automate > Browse Online Library > Table Computations > Difference between pair of rows**. Indexed positions for Column or Rows can be specified. E.g. the Last Row - 2nd Last Row. Useful when a data update will add extra rows or columns to the table.

## Online Feature Release

- Expanded feature for automatically combining categorical data based off patterns (using CHAID) available in
**Automate > Browse Online Library > Create New Variables > Automatically Combine Categories > By Pattern (CHAID)**:- Adjacent Categories: Now possible to specify if some categories should be treated as unordered. E.g. "Don't know" category is allowed to merge with any other category while the other categories are only allowed to combine with adjacent categories.
- Adjacent Categories Unless Missing Value Code: Same behaviour as the Adjacent Categories feature except any categories which are coded with a value of NaN in the Value Atrributes with be allowed to combine with any other category.

## 5.14.2.0

- Bug fixes.

## 5.14.1.0

- The Advanced Statistical Testing Assumptions Dialog has been updated and new options added.
- Additional security options for your Q Customer Page
- Bug fixes.

## Online Feature Release

- New tools for automatically combining categorical data based off patterns (using CHAID) are now available in
**Automate > Browse Online Library > Create New Variables > Automatically Combine Categories > By Pattern (CHAID)**:- Any Categories: Categories are allowed to combine with any other categories.
- Adjacent Categories: Categories are allowed to combine with only adjacent categories.

## Online Feature Release

- New tools for automatically banding/binning numeric data are now available in
**Automate > Browse Online Library > Create New Variables > Automatically Combine Categories > By Value**: - A new option for identifying duplicate cases in the data set:
**Automate > Browse Online Library > Create New Variables > Duplicates** - Create a new variable to compare two date periods directly using
**Create New Variables - Contrast Periods**

## 5.13.1.0

- Added column comparison support for chart exporting to PowerPoint.
- More logical anchoring of items when sorting (e.g. SUM, NET, “Other”, “None”)
- The weight dialog has been redesigned to make it easier to use and more powerful. Have a look
**How to Configure a Weight from Variable(s) in Q**. The weights are now computed using R, and may occasionally result in minor differences in weight values when using multiple Adjustment Variable Sets. - We have introduced new visualizations for use in Q.
- Bug fixes.

## Online Feature Release

**Automate > Browse Online Library > Tables > Combine Tables**now allows you to override row and column names in the combined table.

## Online Feature Release

- Added the Propensity Weight extension button to return the Propensity Weights of a regression or machine learning output that is a binary classification model.

## 5.12.4.0

- Bug fixes.

## 5.12.3.0

- Bug fixes.

## Online Feature Release

- Machine Learning and Regression analyses, such as Regression - Linear Regression and Machine Learning - Random Forest, can now return very large outputs when working with big data sets. To enable, select the analysis and check
**Increase allowed output size**in the Object Inspector and increase the limit in the dropdown that appears.

## Online Feature Release

- Updated Correlation - Scatterplot Matrix to allow more charting options and include tooltips.
- Visualization - Column Chart and Visualization - Bar Chart with multiple series shown as grouped (or clustered) bars now allows spacing between bars to be adjusted.
- Create sentences or paragaphs to describe results with Data - Dynamic Text Box.

## 5.12.2.0

- Fixed a problem where exporting to PowerPoint to update a file created in an earlier version of Q did not update footers.
- Minor bug fixes.

## 5.12.1.0

- We have introduced a new method of exporting to PowerPoint, with many new features and improvements. The old method has been retained as a legacy option. Please read about the new exporting features at
**Exporting to PowerPoint**and**Automatic Updating of PowerPoint Documents**. - Bug fixes.

## Online Feature Release

- New feature Tables - Select Rows and Columns from Table, which can be used to easily select (remove) rows and columns from tables. Selection can be performed using check box controls in the Object Inspector, via an R Output supplying row/column labels, by specifying a range of indices, or using the first or last
**n**rows/columns for a specified**n**, with special functionality for working with tables that contain dates in their labels. - Correlation tables can now be supplied as inputs to Dimension Reduction - Correspondence Analysis of a Square Table and Dimension Reduction - Correspondence Analysis of a Table. In this case, 1 is added to each value in the table so that all elements are positive.

## Online Feature Release

- Percentage data in R tables is now represented in the same way as Q tables (that is, they are no longer stored as decimals).
- Column Chart visualizations has a new group of controls to add annotations that are separate from data labels and can be more flexibly positioned.
- Data label annotations in Column Chart, Bar Chart and Line Chart visualizations has new annotation types
*Caret - up*and*Caret - down*. - Features in the menus have been renamed. See Menu Changes (February 2021).
- Bug fixes for Radar Chart, including improved label placements. Axis and gridlines can now be shown in radar charts with small multiples.
- Annotations can now be added to Radar Chart. These can be attached to data labels or to the category labels.

## Online Feature Release

- Lots of Crosstabs (
**Create > Tables > Crosstabs**) now allows users to choose which statistics should be shown on the created crosstabs. - Improved controls for selecting "must include" and "mutually exclusive" alternatives in TURF - TURF Analysis. The constraints are now specified via comboboxes containing permitted options instead of dropboxes containing all variables in the data sets. This also permits "must include" and "mutually exclusive" alternatives to be specified for Pick Any - Compact questions.

## 5.11.1.0

- You can now link Q to existing PowerPoint documents in PowerPoint so that it automatically updates”, https://wiki.q-researchsoftware.com/wiki/Exporting_to_PowerPoint.
- It is now possible for the user to change a Text Variable to a Categorical Variable in
**Variables and Questions > Variable Type** - Speed improvements
- Bug fixes

## Online Feature Release

- Added option to control the
**maximum number of ticks**shown on each axis in Area, Bar, Column, Line and Scatterplot visualizations. - Added Tables - Raw Data - Question which creates raw data tables from selected questions.

## Online Feature Release

- New QScript available for combining and merging tables: Tables - Combine Tables. For users familiar with Tables - Merge Two or More Tables, the merging is done in the same way, but unlike that Standard R item, the QScript works with any tables selected on the page and the join direction is determined automatically.
- Added
**Truncate cases when characters exceed**option to Text Analysis - Advanced - Principal Components Analysis (Text) so that cases in input text are truncated beyond a specified number of characters. - Added new menu item Visualization - Stacked Column Chart with Statistical Significance which allows a stacked column chart showing statistical significance in the same way as standard Q Tables (e.g. for z-Statistic and Column Comparisons) to be easily created.
- Updated the Regression - Legacy Driver Analysis - Table of Kruskal Importance Scores and Regression - Legacy Driver Analysis - Table of Shapley Importance Scores QScripts and Rules to compute a Regression - Legacy Driver Analysis - Table of Relative Weights (Relative Importance Analysis) instead when more than 27 independent variables are requested. Due to technical limitations, it is not possible to compute Kruskal and Shapley importance in those scenarios. In the moderate case where more than 15 but less than 27 independent variables are requested, the user is prompted with a suggestion to use Relative Importance instead since it typically gives very similar results and is a much faster algorithm while Kruskal or Shapley can be slow and take minutes or even hours to compute.
- Added Calculate - First and Calculate - Last to extract first and last rows and columns from tables and R data outputs.
- Classifier menu in the Automate > Browse Online Library has been replaced by the Machine Learning menu (all features that used to be in the Classifier menu are available in the Machine Learning menu).
- The mice package in R that is used to conduct Multiple Imputation has been updated from version 3.11.0 to 3.12.0. Some numerical differences in output will arise due to the new matching algorithm. The new algorithm however, is 50 to 600 times faster for the predictive mean matching imputation process and still retains the same statistical quality in the imputations.

## Online Feature Release

- Added new menu item Tables - Table of Differences for visualizing differences between two Tables. Statistically significant differences can be highlighted using arrows, cell shading, or boxes.
- A new QScript is available in Modify Tables or Plots - Rename Table Page Names and Titles with Table Name which allows easier import/export QPacks containing large numbers of tables between Q and Displayr.
- Allow pre-selection of Questions in Modify Data - Add Bottom K Category NETs (Bottom K Boxes) and Modify Data - Add Top K Category NETs (Top K Boxes). That is, any Questions that are selected before the script is executed will be used to construct the new NETs and no prompts about Questions selection presented. If no Questions are pre-selected then the user is given prompts to select Questions as previously.
- Added
**Truncate cases when characters exceed**option to Text Analysis - Automatic Categorization - Unstructured Text so that cases in input text are truncated beyond a specified number of characters.

## Online Feature Release

- The varimax rotation algorithm for Text Analysis - Advanced - Principal Components Analysis (Text) has been modified to perform better with larger data sets. This may result in slight differences in output. Also the return object has been trimmed to have a small memory footprint.
- An update has been made to Text Analysis - Advanced - Term Document Matrix so that it accounts for ngrams in the Text Analysis Setup item. The output is now in a sparse matrix format of class
`dgCMatrix`, which is more broadly used than the previous format. - The
**Suggest**feature which provides suggestions for categories in the text coding interface (accessed by right-clicking text variables in the**Variables and Questions Tab**and selecting one of the options under**Insert Variables > Code Text**) now runs faster for larger data sets (over 1000 cases). - The algorithm for the
**Sort by > Similarity to [category]**feature in the text coding interface (accessed by right-clicking text variables in the**Variables and Questions Tab**and selecting one of the options under**Insert Variables > Code Text**) has been updated, which may result in differences in the sorting of cases. - The underlying algorithm for Text Analysis - Automatic Categorization - Unstructured Text, Text Analysis - Advanced - Principal Components Analysis (Text) and text coding has been modified for better performance. This may result in differences in output such as generated categories, principal components and suggested categorizations.
- The algorithm for categorizing with an existing categorization in Text Analysis - Automatic Categorization - Unstructured Text has been updated. This may result in differences in categorization.
- Text Analysis - Advanced - Setup Text Analysis now shows a page of raw text in the output. The controls
**Page of raw text to show**and**Rows of raw text per page**have been added, allowing the page number and page size to be specified. Previously all the raw text was shown, potentially resulting in an error or unresponsive output for large datasets.

## Online Feature Release

- Fix made to Tables - Merge Two Tables and Tables - Merge Two or More Tables, so that merging a text and numeric table does not convert the numeric columns into text.
- The default behavior with unordered categorical variables for Dimension Reduction -- PCA, Scatterplot t-SNE, and MDS -- has changed to not create binary variables. To use the old default with new outputs, simply select the Check Box
**Create binary variables from unordered categories**. - Several features in the R Online Library, including all items in
**Automate > Browse Online Library > Dimension Reduction**,**Regression**,**Segments**, and**Visualization**now use the Value Attributes when converting Pick One and Pick One - Multi questions to numeric values. - Annotations can now be added to the data labels of Visualization - Line Chart based on other statistics in the input table.
- The
**minimum frequency**option for Text Analysis - Advanced - Setup Text Analysis now defaults to scale minimum frequency as a percentage of cases. There is still the option to specify the minimum frequency as a fixed number. - The underlying algorithm for Text Analysis - Automatic Categorization - Unstructured Text, Text_Analysis - Advanced - Principal Components Analysis (Text) and text coding has been modified to perform better with larger data sets. This may result in differences in output such as generated categories, principal components and suggested categorizations.

## Online Feature Release

- Visualization - Scatterplot has new option to
**use category labels instead of values**so that row or span labels can be used in the**X coordinates**,**Y coordinates**,**Sizes**or**Colors**variable.

## 5.10.1.0

- A new PowerPoint exporting option has been added at
**File > Export > To office > Fast PowerPoint Update**. Select an existing PowerPoint document. Q will then update the charts and tables to match the data in Q. - Speed improvements
- Bug fixes

## Online Feature Release

- Added the following extension buttons that appear in the Inputs tab when selecting a choice model experimental design: Balances and overlaps, Numeric design, Standard errors.
- Added the following extension button that appears in the Inputs tab when selecting a choice model output: Parameter statistics table.
- Added the following extension button that appears in the Inputs tab when selecting a latent class analysis choice model output: Class Parameters.
- Added the following extension buttons that appear in the Inputs tab when selecting a hierarchical Bayes choice model output: Posterior intervals plot, Trace plots.
- Added the following extension button that appears in the Inputs tab when selecting a correspondence analysis of a table: Quality table.
- Added the following extension button that appears in the Inputs tab when selecting a linear discriminant analysis: Coefficients of discriminant functions.
- Added the following extension button that appears in the Inputs tab when selecting a machine learning output: Prediction-accuracy table.
- Added the following extension button that appears in the Inputs tab when selecting a MaxDiff output: Parameter statistics table.
- Added the following extension buttons that appear in the Inputs tab when selecting a latent class MaxDiff output: Class parameters table, Class preference shares table.
- Added the following extension buttons that appear in the Inputs tab when selecting a hierarchical Bayes MaxDiff output: Posterior intervals plot, Trace plots.
- Added the following extension buttons that appear in the Inputs tab when selecting a regression model output (not all buttons apply to all regression types): Residual heteroscedasticity test, Multicollinearity (VIF) table, Residual normality (Shapiro-Wilk) test, Cook's distance plot, Cook's distance vs leverage plot, Influence index plot, Normal Q-Q plot, Residuals vs fitted plot, Residuals vs leverage plot, Scale-location plot, Prediction-accuracy table, Serial correlation (Durbin-Watson) test.
- QScripts in
**Automate > Browse Online Library > Create New Variables**will now use pre-selected questions (either in the*Variables and Questions*tab or from a table in the*Outputs*tab) as the input data to the QScript when appropriate. If no question is selected then the prompts will appear to select questions from a list in the same way as before. This change also occurs in**Automate > Browse Online Library > Filtering > Create Filters from Selected Data**and**Automate > Browse Online Library > Missing Data > Save Variable(s) > Complete Cases**.

## 5.9.7.0

- Bug fixes

## 5.9.6.0

- Bug fixes

## 5.9.5.0

- Bug fixes

## 5.9.4.0

- Bug fixes

## Online Feature Release

- Fixed error in handling of missing values when computing
*Percentages*of variables in Visualization - Number - Number, Visualization - Number - Number in Bar, Visualization - Number - Number in Donut, Visualization - Number - Number in Gauge, Visualization - Number - Number in Oval, Visualization - Number - Number in Rectangle and Visualization - Pictograph - Single Icon **Values Axis > Minimum value**now adjustable in Visualization - Radar Chart.- New relative importance output options of
**Jaccard Coefficient**and**Correlation**available in Regression - Driver Analysis (when**output**type is*Linear Regression*) and in Regression - Linear Regression - Added the following extension buttons that appear in the Inputs tab when selecting a hierarchical Bayes, latent class analysis or multinomial logit choice model output: Save individual-level coefficients, Save proportion of correct predictions, Save RLH, Save utilities (mean = 0), Save utilities (mean = 0, max range = 100), Save utilities (mean = 0, max range = 100), Save utilities (min = 0), Save utilities (min = 0, max range = 100), Save utilities (min = 0, mean range = 100).
- Added the following extension buttons that appear in the Inputs tab when selecting a latent class analysis choice model output: Save class membership probabilities, Save class membership.
- Added the following extension button that appears in the Inputs tab when selecting a principal components analysis, multiple correspondence analysis or t-SNE output: Save transformed variables.
- Added the following extension buttons that appear in the Inputs tab when selecting a linear discriminant analysis output: Save discriminant variables.
- Added the following extension buttons that appear in the Inputs tab when selecting a machine learning output: Save predicted values, Save predicted probabilities.
- Added the following extension buttons that appear in the Inputs tab when selecting a hierarchical Bayes, latent class analysis or multinomial logit MaxDiff output: Save preference shares, Save Sawtooth preference shares, Save zero-centered utilities, Save individual-level coefficients, Save proportion correct, Save root likelihood (RLH).
- Added the following extension buttons that appear in the Inputs tab when selecting a latent class analysis MaxDiff output: Save class membership probabilities, Save class membership.
- Added the following extension buttons that appear in the Inputs tab when selecting a regression output: Save fitted values, Save predicted values, Save residuals.
- Added the following extension button that appears in the Inputs tab when selecting a k-means cluster analysis output: Save segment membership.
- Added the following extension buttons that appear in the Inputs tab when selecting a text analysis output: Save categories, Save first category.
- Added the following extension buttons that appear in the Inputs tab when selecting a Setup Text Analysis output: Save tidied text, Save sentiment scores.
- Modified arrow symbols used in Segments - Segment Comparison Table to improve visibility.
- Random number changes in Machine Learning - Deep Learning outputs due to an update in Tensorflow to version 2.2.0.
- Fixed
`beta.draws`property in choice model outputs so that draws for numeric variables in`beta.draws`are correctly scaled. Also prefixed the term "Scaled" to numeric variable names in Choice Modeling - Diagnostic - Parameter Statistics (hierarchical Bayes only), Choice Modeling - Diagnostic - Posterior Intervals and Choice Modeling - Diagnostic - Trace Plots outputs as they are scaled by a multiplicative factor for modeling purposes. - Added the following extension button that appears in the Inputs tab when selecting a Principal Components Analysis or Principal Components Analysis (Text) output: Component Plot.
- Added the following extension button that appears in the Inputs tab when selecting a Choice Modelling output: Utilities Plot.
- A new QScript, Choice Modeling - Convert Alchemer (Survey Gizmo) Conjoint Data for Analysis, is available for automatically converting Conjoint data files from Survey Gizmo into a convenient format for use in Q's choice models (Hierarchical Bayes, Latent Class Analysis, and Multinomial Logit).
- Shortened names of variables generated by the
**Save Variable(s)**QScripts as well as various other QScripts that save variables with a name containing a random string. The names now simply consist of a short prefix and indices when there are multiple variables or naming conflicts with existing variables. - Added random seed option to Choice Model and MaxDiff Standard R outputs which allows for different random initial parameters for the model and different questions to be left out for cross-validation.
- Added "Number of starts" option to Choice Model and MaxDiff latent class analysis outputs which allows the model to be run multiple times with different initial parameters and the model with the best log-likelihood is selected as the final model.
- Added new controls to the relevant Text Analysis suite of functions to limit the number of levels and variables per case to save for the Text Analysis - Save Variable(s) - Categories and Text Analysis - Save Variable(s) - First Category output variables. These are two new controls that limit the number of unique levels in the categorical output variable and the number of saved variables per case. The new controls exist in the Text Analysis - Automatic Categorization - List of Items, Text Analysis - Automatic Categorization - Entity Extraction and Text Analysis - Advanced - Setup Text Analysis R outputs.

## 5.9.3.0

- Bug fixes

## 5.9.2.0

- Bug fixes

## 5.9.1.0

- This release includes major improvements and a new interface for Text Analysis, especially options under
**Create > Text Analysis > Advanced** - Users can now customize the labels used for quarters under
**Edit > Project Options > Customize > Output Text** - Speed improvements
- Bug fixes

## Online Feature Release

- Visualization - Line Chart now has the option to add data labels and markers to the end points of each series only.
- Regression - Driver Analysis added to the menu to facilitate quick driver analysis options and data stacking.
- Text Analysis - Advanced - Principal Components Analysis (Text) added allowing the option to analyze text data by conducting a PCA via a numeric text encoding. The correlation of the PCA scores with a type of document term matrix is used to help interpretation and a varimax type rotation is conducted by default to help facilitate the interpretation (the rotation can be not conducted if desired).
- The order of Row and column manipulations in Visualizations has been modified to make it easier to apply multiple manipulations. The operations are now performed in the following order: Hide empty rows/columns, Hide rows/columns with small sample sizes, Rows/Columns to show, Sort rows/columns, Order rows/columns by similarity, Rows/Columns to ignore, Number of rows/columns from top/left to show, Number of rows/columns from from bottom/right to show.
- Segments - Segment Comparison Table now has more options for shading significant values. These output options are now available in Segments - K-Means Cluster Analysis when the
**output**type is*Segment profiling table*. - Create New Variables - Translate added to the menu allowing translation of text questions to another language.
- New transformation QScripts are available in
**Create New Variables > Scale Within Case**and**Create New Variables > Scale Within Variable**. The QScripts formerly named Rank Within Variable, Rank Within Case, and Standardize Data by Case have been moved to these submenus. The new QScripts are 1) Standardize Within Case, which transforms data to have mean 0 and standard deviation 1, 2) Center Within Variable, which transforms data to have mean 0, and 3) Unit Interval Within Case, which transforms data to have minimum 0 and maximum 1.

## Online Feature Release

- Implemented the dummy variable adjustment technique for missing data in Regression - Linear Regression and the other regression types. It is accessible through the
**Missing data**dropbox for all relevant regression types. - Added
**Output language**control to Text Analysis - Automatic Categorization - Unstructured Text which allows categories and example text in the output to be shown in languages other than English. - Two new QScripts are available in
**Insert > Transform > Modifying Data**: Modifying Data - Add Top K Category NETs (Top K Boxes) and Modifying Data - Add Bottom K Category NETs (Bottom K Boxes), which allow for creating new NET categories in*Nominal*and*Nominal - Multi*variable sets.

## Online Feature Release

- Regression - Linear Regression and the other six regression types now have an automated outlier removal option. The new input control is called
**Automated Outlier Removal Percentage**and accepts a numeric value between 0 and 50 (including 0 but not 50). If a non-zero value is selected for this option then the regression model is fitted twice. The first regression model uses the entire dataset (after filters have been applied) and identifies the observations that generate the largest residuals. The user specified percent of data values that have the largest residuals are removed from the data. The regression model is refitted on this reduced dataset and output returned. If a zero-value is selected for this input control then no outlier removal is performed and a standard regression output for the entire (filtered) dataset is applied. - More formatting options for secondary data series has been added to Visualization - Column Chart
- Added support for text data in languages other than English in Text Analysis - Automatic Categorization - Unstructured Text through the new
**Source language**control. - Visualization - Scatterplot has been updated so that
*Row/column manipulations*are applied to the input table in the same way as other visualizations when*Input data contains y-values in multiple columns*is selected.

## Online Feature Release

- Annotations can now be added to the data labels of Visualization - Bar Chart and Visualization - Column Chart, based on other statistics in the input table.
- Visualization - Scatterplot now able to handle Regression output in the X coordinate and Y coordinate input area. The values in the plot are either the regression coefficients or importance scores respectively depending on the Regression settings.
- Text Analysis - Automatic Categorization - List of Items and Text Analysis - Advanced - Setup Text Analysis now default to counting frequency by number of respondents instead of occurrences in the input text.
- Added Marketing - Price Sensitivity Meter
- In Visualization outputs, the result when multiple
**Row/column manipulation**options are used has been changed.**Number of rows/columns from top/bottom/left/right**is now applied after**Rows/Columns to show**. For example, you can now specify a long list of variables to show using**Columns to show**, and then specify that only the first 3 of the columns selected should be shown, without using an intermediate table. - Visualization - Scatterplot now has the option of adding annotations.

## Online Feature Release

- Shapley Regression available for Regression - Linear Regression.
- Tick labels in Scatterplots with categorical axis is now ordered according to the order they appear in input table (instead of alphabetically).
- Regression outputs with imputated values may have small changes due to an upgrade to
`mice`version 3.7.0.

## 5.8.2.0

- Bug fix.

## Online Feature Release

- Raw text replacement and Category replacement options in Text Analysis - Automatic Categorization - List of Items has been replaced by the Required categories (variants can now be specified) and Split into categories option.
- Variant suggestions for Required categories are now shown in the diagnostics output for Text Analysis - Automatic Categorization - List of Items.
- Added support for the Text Analysis - Advanced - Save Variable(s) - First Category and Text Analysis - Advanced - Save Variable(s) - Categories QScripts for Text Analysis - Advanced - Setup Text Analysis output and Text Analysis - Automatic Categorization - Unstructured Text with a multiple response Existing Categorization.
- Added Segments - Segment Comparison Table.
- Some some changes in Machine Learning - Deep Learning outputs because of an update to use Tensorflow version 2.0.0

## 5.8.1.0

- The standard error shown on tables of correlations was computed using (1 - R2)/sqrt(n - 1). This is not correct. It now is computed using sqrt((1-R2)/(n-2)). The statistical tests were using the correct formula, so no change has occurred in terms of which results are flagged as significant.
- On summaries of numeric variables, the standard deviations and associated statistics (standard errors and statistical tests) had an error in them whereby they failed to take missing values into account. If a study has extremely high rates of missing data these may have caused non-trivial differences. This only affected summary tables; not any crosstabs.
- Errors in Independent Complex Samples t-Test / z-Test:
- Column Comparisons report statistically significant differences between values which are identically 100.
- Where the user has chosen to apply a z-Test to test means on this table (
**Edit > Project Options > Customize > Statistical Assumptions > Statistical Tests for categorical and numeric data > Means > z-Test**), but when using the 'alpha' button to test significance the output is providing a link to the t-Test version of the complex samples test for means.

- Bug fixes.

## 5.7.1.0

- Incremental improvements to data and R code handling to improve speed.
- Bug fixes.

## Online Feature Release

- Added Text Analysis - Automatic Categorization - Entity Extraction which automatically extracts Named Entities from a text variable.
- Added "Existing categorization" option to Text Analysis - Automatic Categorization - Unstructured Text which allows an existing categorization to be used by the categorization algorithm.
- Added support for identifying n-grams in Text Analysis - Advanced - Setup Text Analysis.
- Added Binary transformation for Numeric and Categorical data in Create New Variables - Binary Variable(s)
- Added Log based transformations in Create New Variables - Log Transform Variable(s)
- Added Square-root and power based transformations in Create New Variables - Square-root
- Added statistical rank transformations in Create New Variables - Rank Within Case and Create New Variables - Rank Within Variable
- Added support to Marketing - MaxDiff - Hierarchical Bayes, Marketing - MaxDiff - Latent Class Analysis, and Marketing - MaxDiff - Multinomial Logit for handling best and worst choices that are supplied as the displayed option (e.g. the values 1,2,..., up to the number of alternatives per question) instead of as the selected alternative.
- Changed handling of missing values in Pie Chart to be consistent with other visualizations.

## Online Feature Release

- Tables - Paste or Enter Table has new option to
*Autofit*table to fill the container (without scroll-bars). - Visualization - Column Chart has new option to add additional data series as lines with a second Y axis that can be customized separately from the primary Y axis.
- Added Text Analysis - Automatic Categorization - Unstructured Text which automatically codes a text variable containing unstructured text into categories.

## 5.6.5.0

- Bug fix

## Online Feature Release

- Tables - Crosstabs QScript now optionally orders the generated tables based on the most significant p-values and can also hide tables which don't contain p-values below a selected significance threshold.
- Diagnostics information is now shown with Text Analysis - Automatic Categorization - List of Items.

## 5.6.4.0

- Fixed bug in planned significance test results where all cells in the table including the NET are selected were not being shown for crosstabs of Number (blue question) and Pick One (brown question) in the case where there is missing data in the Pick One question. This also fixes an issue with Smart Tables where these crosstabs may have been incorrectly labeled as insignificant.
- Minor bug fixes

## Online Feature Release

- Marketing - Brand Health Table
- Visualization - Venn Diagram now supports the same color options as the other charts.
- Visualization - Bar Chart and Visualization - Column Chart can be shown using
**Multiple colors within a series**. - Additional ways to specify the color palette using
**Custom palette (color pickers)**and**Custom palette (R output)**. Sequential color palettes can also be used with**Values**, an optional numeric table to specify the colors used. - Dimension Reduction - Correspondence Analysis of a Table using a
*Bubble Chart*output now allows the bubbles to be colored according the values in a user-provided table. - Update of xgboost to version 0.90.0.1 may cause slight changes to Gradient Boosting outputs.

## 5.6.3.0

- Bug fixes.

## 5.6.2.0

- Addresses speed issues which occur when using double bit character sets (such as Chinese).
- Other bug fixes.

## Online Feature Release

**For details of new online features released up to June 2019, see here.**

## 5.6.1.0

- Corrects errors in importing data from MDD files.
- Prevents crashes caused by adding filters from different data files.
- Corrects an issue updating with new cases from an Excel file when the original file contained blank rows.
- Corrects an error which prevented updating from a datsa files with more variables than the original.
- Corrects rendering of headings in complex banners.
- Prevents an error when using text analysis setup on data which includes Unicode line separators.
- Prevents data being shifted when exporting data with double byte (including Chinese) characters to .csv files.
- Allows the merging of cases from larger data files without triggering out-of-memory errors.
- Retains the value atributes of an R variable when undo is used.
- Minor bug fixes.

## 5.5.4.0

- Bug fixes.

## 5.5.3.0

Features

- Support added for displaying multi column data frames for R Variables.

Bug fixes

- Fixed problems importing from Qualtrics if the default save directory does not exist.
- Corrected the size used for exported wordclouds.
- Minor bug fixes.

## 5.5.2.0

- Fixed an issue where the "Independent" option for column comparisons overlaps was not being observed.
- Fixed slowdown in Q when project contained hundreds of tables.
- Other bug fixes.

## 5.5.1.0

- When specifying a range of dates in the Date/Time settings for a Date variable, the minimum and maximum will now default to the start and end of the days in the range.
- If an Excel/CSV data set looks to contain tabular data (e.g. aggregated statistics) rather than raw data, the user will be given the option to instead add it as a table.
- If a filter or weight is applied to an R calculation that uses variables from different data sets, then there must be a relationship defined between those data sets. The filter/weight will be resized to suit the joined data. Note that we do not yet join the other input variables, but may in future.
- Fixed an issue where variables using Recode Rules would include original code values and labels when "Copy and Paste Variable(s) > Linked" was used.
- QScript can now export to Excel files.
- PowerPoint exports of R Outputs include transparency, so they will look good in templates.
- Bug Fixes

## 5.4.7.0

- Bug fixes

## 5.4.6.0

- Bug fixes

## 5.4.5.0

- Bug fixes

## 5.4.4.0

- Membership probabilities corrected in weighted Mixed-Mode Trees
- Changed data format for Excel charts to avoid double line breaks in axis labels.
- Added validation of variable names and questions in R variables created via QScripts.
- Fix an issue in legacy PCA observed when the number of observations is less than the number of components.
- Improve .QCodes file import
- Bug fixes

## 5.4.3.0

- Improve speed of cross-tab generation.
- Enabled table footer setting 'Include when columns compared'.
- In legacy regression, if predictors have missing values then predictions will also be missing.
- The maximum size of R output that can be exported to PowerPoint, or PDF is now limited to 6000 by 6000 pixels, to prevent exports being blank.
- Exported to PowerPoint Excel Charts are now updated when statistics shown change.
- Bug fixes for .MDD data file reading.
- Other bug fixes.

## 5.4.2.0

- Improve Qualtrics integration.
- Fixed an issue that prevented editing R data set code.
- It is possible to read a table value of R output in QScript using
`.table`

property. - Improve hypothesis testing of ranking questions.
- Add/Edit Table widget now support up to 1000 rows of data.
- Other bug fixes.

## 5.4.1.0

- Q can now load survey data directly from Qualtrics without requiring user to download *.sav data files first.
- The way that the JavaScript language evaluates 'if' statements without an 'else' has changed. This may affect some hand-coded JavaScript variables in old projects. Q should detect affected variables and mark them as in error, forcing you to fix them. See Changes to JavaScript if.
- Bug fixes.

Known Qualtrics import limitations

- Q does not currently support importing surveys with more than 1 million responses.
- Qualtrics Side-by-Side questions may have missing data for some multiple-response questions within them. In particular, responses in which the user has not selected a particular checkbox may be considered missing for the corresponding variable. Qualtrics has acknowledged this issue, and intends to fix it in future.

## 5.3.2.0

- Improved handling of categorical variables in MDD data sets.
- Bug fixes.
- Known issue. In rare cases the 'Undo' (Ctrl + Z) operation may not work after new filter is created from table cells.

## 5.3.1.0

- R calculations done in R Outputs are now allowed to continue for up to 1 day.
- Due to changes in the software on which Q is built, the order of rows may change in the following outputs:
- Matched data sets (which will affect results when using Data Fusion).
- Auto-coded categorical variables (which will result in a different ordering of some categories when Text variables have their Variable Type changed to Categorical).
- The list of subsets in the Homogeneous Subsets section in some significance tests.

- Q now requires .Net 4.5 to be installed.
- Bug fixes

## 5.2.7.0

- Bug fixes

## 5.2.6.0

- Bug fixes

## 5.2.5.0

- Fixed bug where rounding errors can result in the treatment of the median, percentiles and effective sample size to be inconsistent. Based on our sensitivity analysis we have confirmed that this only leads to changes in results in situations where the change is immaterial. In the case of the median, this issue arose on the rare occasion when the median occurred at the middle of two different values. Due to rounding errors, the median was randomly reported as either the higher or lower value, when it should have been the midpoint between the two values. In most cases when this does occur, the two values will be close and the resulting error will be small.
- Other bug fixes

## 5.2.4.0

- Improve import of MDD data files.

## 5.2.3.0

- Improve import of MDD data files.
- Other bug fixes

## 5.2.2.0

- Date variables can display time in charts and tables. Date variables can be aggregated using hours, minutes, seconds or milliseconds.
- Previously when a variable from a Pick Any question only had Missing Data and 1 as values, 1 would not be counted, resulting in 0% shown in the table. 1 is now counted and shown as 100% in the table.
- Q requires .Net 4.5.2 to be fully operational.
- It is now possible to hide tables, charts, R outputs and complete folders from export using
**Hide**from the context menu. - R Outputs returning scalar values and strings are now exported as text to PowerPoint and Word.
- Fixed bug where multiple comparisons correction in Table Options was ignored in Driver Analysis.
- Fixed question naming issues when importing MDD data files.
- Improved looped question detection when importing MDD data files.
- Fixed bug where an incorrect log-likelihood was reported when weights were applied in Segments > Latent Class Analysis.
- Other bug fixes.

## 5.1.2.0

- Check out the key new features: http://wiki.q-researchsoftware.com/wiki/Q5_Key_Features
- Some menu items have moved or changed names. See here for a list: http://wiki.q-researchsoftware.com/wiki/Changes_to_the_Menus_in_Q5
- Fixed a bug in the scaling of Eigenvalues shown in the technical outputs for the PCA Biplot. This does not change the information shown in the chart or relationships between the points in the chart.
- Formatted the intercepts from Ordered Logit to make them consistent with those from Regression - Ordered Logit, and fixed a bug in the saved probabilities from Ordered Logit.
- Other bug fixes.

## 5.1.1.0

- Data tables can now be manually entered or pasted in Q via
**Create > Tables > Enter Data**and**Paste Data**, as well as in Standard R items that accept tabular inputs. - Other improvements and bug fixes.

## 5.0.3.0

- Machine Learning - Random Forest
- Analysis of Variance - One-Way ANOVA
- Analysis of Variance - One-Way MANOVA
- Fixed bug with Spearman’s Correlation where Spearman ranks were being calculated without omitting the cases with missing values in
*any*of the two input variables. The impact of this bug should be minimal and proportional to the amount of missing values in the data set. We have observed that the bug typically resulted in a difference in the reported Spearman correlation at the 4th decimal place compared to the correct value. As a result, is unlikely that this bug would have affected Spearman correlation significance test results and any inferences made from them. - Improved Qualtrics and Survey Gizmo data file detection
- Fixed a bug where by the word "mean" was appearing in the place of "proportion" in some test outputs.
- Other bug fixes and improvements

## 5.0.2.0

- Bug fixes

## 5.0.1.0

- Check out the key new features: http://wiki.q-researchsoftware.com/wiki/Q5_Key_Features

## 4.11.15.0

- Column Comparisons in Statistics - Below are now computed in Pick One - Multi questions using any available data (previously they were filtered to use only data for respondents with complete data). The previous results can be replicated by filtering.

## 4.11.14.0

- Bug fixes

## 4.11.13.0

- Bug fixes

## 4.11.12.0

- Bug fixes

## 4.11.11.0

- Check out the key new features: http://wiki.q-researchsoftware.com/wiki/Q4.11_-_Key_Features
- Columns can now be compared across column spans.
- Bug fixes

## 4.11.10.0

- Bug fixes

## 4.11.9.0

- Bug fixes

## 4.11.8.0

- Bug fixes

## 4.11.6.0

- Bug fixes

## 4.11.5.0

- Bug fixes.

## 4.11.4.0

- Bug fixes.

## 4.11.3.0

- Bug fixes.

## 4.11.2.0

- Export to Excel (fast) now works without Office installed.
- New Table Context Menu items allow you to modify questions without leaving the table, including renaming the Question and changing its Question Type.

## 4.11.1.0

- Banners can be modified directly on tables by dragging and dropping column spans.
- Individual banner points can be added directly on tables using new concise Logic formulas.
- New Ranking plot visualizes changes in ranks as a series of colored ribbons.

## 4.10.6.0

- Bug fixes.

## 4.10.5.0

- Significance arrows and column comparison letters are now included in Excel spreadsheets and PowerPoint chart labels.
- Bug fixes.

## 4.10.4.0

- Bug fixes.

## 4.10.3.0

- Bug fixes.

## 4.10.2.0

- New Word Clouds visualize categorical or open-ended responses and allow simple coding.
- New options have been added for specifying
*design effects*, which allow you to: (a) conduct statistical tests with weighted data using un-weighted sample sizes, and (b) supply design effects that are incremental to those computed automatically to take the weighting into account. As a consequence of these changes, Population statistics (i.e., weighted sample sizes) no longer change when design effects are changed. Note that some results will change if you have previously specified design effects and then selected an option not using design effects (as now, all options will work with design effects). See Weights, Effective Sample Size and Design Effects. *Dependent*versions of z- and t-tests have been added for proportions and means: Complex Samples Dependent Z-Test - Comparing a Sub-Group Proportion to Total, Complex Samples Dependent T-Test - Comparing a Sub-Group Proportion to Total, Complex Samples Dependent Z-Test - Comparing a Sub-Group Mean to Total, and Complex Samples Dependent T-Test - Comparing a Sub-Group Mean to Total.- Confidence Intervals.
- Added an editor for QScripts. Significantly updated the JavaScript Variable expression editor and Rule script editor. See Editing JavaScript in Q.
- Banners now update automatically whenever their input questions update. Banners can now be edited to change which questions are included.
- Time Series Analysis settings can now be changed for multiple rows at once.
- The Data Difference Warning table now has a
**Copy**button for pasting into Excel. - A new search box to help you find Q features and documentation.
- A new Export to Excel (fast) option. Better exported sheet names and the option to export each group to a sheet.

## 4.9.7.0

- Bug fixes.

## 4.9.6.0

- Bug fixes.

## 4.9.5.0

- Bug fixes.

## 4.9.4.0

- Bug fixes.

## 4.9.3.0

- Bug fixes.

## 4.9.2.0

- Bug fixes.

## 4.9.1.0

- Q is faster when working with large data files (100,000+ cases).
- New Find and Replace feature allows you to search everything in your project: tables, variables, value attributes, raw data and more.
- Improved importing of CSV and Excel data files.

## 4.8.9.0

- Bug fixes.

## 4.8.8.0

- Driver (Importance) Analysis can now analyze questions from multiple data files.
- Bug fixes.

## 4.8.6.0

- The tables created by Automate > Browse Online Library > Driver (Importance) Analysis can now be crosstabbed against Pick One, Pick Any and Date questions.
- TURF is now available from Automate > Browse Online Library, allowing you to select products and maximum portfolio size.

## 4.8.5.0

- Bug fixes.

## 4.8.4.0

- Bug fixes.

## 4.8.3.0

- Rules let you automatically highlight important cells and reformat tables.
- Significance of table cells is now calculated by Q prior to Rules. This allows Rules to format tables based on Corrected p, and to modify arrows and font colors.
- Driver (Importance) Analysis is available from Automate > Browse Online Library.
- Tables of Pick Any - Compact questions now automatically remove respondents who were not asked the question.

## 4.8.2.0

- Bug fixes.

## 4.8.1.0

- Bug fixes.

## 4.7.5.0

- Bug fixes.

## 4.7.4.0

- Bug fixes.

## 4.7.3.0

- Bug fixes.

## 4.7.2.0

- Project Templates let you set default statistical assumptions, table styles, colors, etc. for new projects.
- Bug fixes.

## 4.7.1.0

- The QScript Online Library provides pre-written QScripts to perform common tasks. It is available in the Tools menu, QScripts, Online Library.
- Use Excel spreadsheets as raw data.
- Now much easier to control whether/how significance is shown for any table/plot. Look for the new "Show significance" toolbar.
- Plots can now show column comparisons.
- Plots can show significance arrows without also needing to show labels. Most plots now show significance by default.

## 4.6.6.0

- Bug fixes.

## 4.6.5.0

- Bug fixes.

## 4.6.4.0

- Bug fixes.

## 4.6.3.0

- Tables with no data will no longer stop Microsoft Office exports mid-stream.
- During text coding, codes may now have the same name so long as they do not have the same parent code.
- Fixed inconsistent spacing between letters on plots exported as pictures to PowerPoint or Word.
- Axis labels on plots can be turned off.
- Bug fixes.

## 4.6.2.0

- Text questions can now be crosstabbed against Pick One, Pick Any, Pick Any - Compact or Date questions.
- Excel table exports can now be formatted using Excel's table styles, within Q.
- Bug fixes.

## 4.5.5.0

- Bug fixes.

## 4.5.4.0

- Q now works with R3.0.1.
- Bug fixes.

## 4.5.3.0

- Bug fixes.

## 4.5.2.0

- Bug fixes.

## 4.5.1.0

- You can specify the gaps between tables when exporting all tables to a single Excel worksheet.

## 4.4.4.0

- Bug fixes.

## 4.4.3.0

- Bug fixes.

## 4.4.2.0

- Bug fixes.

## 4.4.1.0

- Help | Interpret This Table prepares a summary of some of the salient aspects of a table.
- Basic Tables has a new checkbox that allows you to add all crosstabs into a single group (instead of adding one group for each Key question's crosstabs).

## 4.4.0.0

- Export to native PowerPoint tables. These look much better than the old tables, and will work with PowerPoint table styles.

## 4.3.21.0

- Bug fixes.

## 4.3.20.0

- Bug fixes.

## 4.3.19.0

- Bug fixes.

## 4.3.18.0

- Time series charts allow you to select the statistic to be plotted.
- Significance tests on Number questions now test against an expected value of 0.0 (previously no result was shown).
- Choice of significance tests for grid questions (e.g., Nonparametric, z-tests, Quantum and Survey reporter).

## 4.3.17.0

- The multiple comparison correction for Column Comparisons can now be applied to the whole table, or, within each row and span. (Previously it was only applied within each row and span.)
- Symbols used in Column Comparisons (e.g., upper case for significance levels than 0.001 and lower case for between 0.001 and 0.05 are now computed using the Corrected p. Previously they were computed using p. Note that this does not change which column comparisons are marked as significant, it only changes the letters used. Also note that this change only affects users on the beta, as the different column letters are not available in older versions of Q.
- Improved warnings when Experiment questions have been setup incorrectly.
- Modified list of Multivariate Tests (no change in the tests, but modified names of the tests).
- The Independent Complex Samples t-Test - Comparing Two Means is now the default test employed with weighted data (previously a version assuming equal variance was applied in situations with small sample sizes).
- The Independent Samples t-Test - Comparing Two Coefficients for an Experiment and Ranking questions now uses the effective sample size when computing the degrees of freedom (previously it took into account the effective number of observations).

## 4.3.16.0

- Online wiki-style documentation, integrating the current Help, Q Reference Manual and How Tos [1].
- Statistical Assumptions setting for
**Proportions**now determines the choice of statistical test on SUMMARY tables containing proportions . - Where samples are related (i.e., matched or paired), the
*Dependent Samples t-Test*has been replaced by the Paired Z-Test of Proportions. Note that the dependent tests were only ever available in beta versions of Q. Dependent sample tests are still available by selection within Statistical Assumptions (see Related Samples Tests - Comparing Two Proportions and Related Samples Tests - Comparing Two Means ). - Where a Pick Any question has only two categories a related sample test is used (see Related Samples Tests - Comparing Two Proportions).
- Correction of bugs for tests involving
**Quantum**and**Survey Reporter**testing assumptions. In particular, an Effective Base n statistics were computed only for columns. Note: these tests were only ever available in the beta. - Fixed problem with Column n statistic in multiple data file crosstabs involving
**Ranking**questions.

## 4.3.15.0

- Fixed problem with three-dimensional crosstabs with Date questions in the brown dropdown.
- Fixed problem with Column Comparisons in three-dimensional crosstabs.

## 4.3.14.0

- Three-dimensional and four-dimensional tables; crosstabs between Grid and other questions are now possible.
- Automatic setup of data files from SurveyMonkey.
- Total Unduplicated Reach and Frequency (TURF).
- User can choose algorithm for Correlations in Statistical Assumptions.
- Additional statistical tests added to Statistical Assumptions including Quantum and Survey Reporter's style
*dependent*tests (see Related Samples Tests - Comparing Two Proportions and Related Samples Tests - Comparing Two Means). - Nonparametric Tests can be selected on most tables. (Not available on Pick Any - Grid, Experiment and Ranking questions, although the ranking question tests are arguably nonparametric).
- Modification to defaults used for indicating non-tests and non-significant comparisons in Column Comparisons.
- Change to computation of statistical tests in SUMMARY tables of Number - Multi questions. Previously, each category was compared to the average of all the categories (excluding NETs). Now, each cell is compared to the average of the other cells (excluding NETs).
- Additional multiple comparison corrections have been programmed for Column Comparisons, including: Fisher LSD, Duncan, Newman-Kuels (S-N-K), Tukey HSD and Dunnett (see ANOVA-Type Tests - Comparing Three or More Groups ).
- The
**Bonferroni**correction is no longer available as a for cell comparisons in Statistical Assumptions. - Change to null hypothesis of ANOVA when selecting cells in an appropriate table (the revised null hypothesis is shown when running Planned Tests Of Statistical Significances.
*VIF*in Regression is now computed using the weighted correlation matrix.- Additional Mapping algorithms, including Multidimensional Scaling (MDS) and Principal Components Analysis Biplot.

Please note: a number of the new features described here were available in some earlier 4.3 releases.

## 4.3.12.0

- Pie charts.
- JavaScript variables now have the ability to reference variables in other data files.

## 4.3.9.0

- When (for example) you restrict a weekly Date question to end on a Wednesday then calculated weeks will go from Thursday to Wednesday. Similarly for months and years. Previously Q moved periods to match natural boundaries (Sunday, end of month, etc) in this case.

## 4.3.8.0

- Q can read SPSS Data Collection Data Files (.mdd + .ddf).
- New and improved JavaScript variable editor.
- Edit | Table JavaScript - allows you to script the output of a table or plot, which can generate heatmaps, custom significance highlighting, and more.

## 4.3.7.0

- New weighting facility - enter targets as percentages, handles new waves of trackers, automatic rim weighting (combines multiple weight targets), enter targets that vary by month.

## 4.3.5.0

- Fetching raw data from SQL databases.

## 4.3.1.0

- Find relationships between variables: right click and choose Trace Dependencies

## 4.3.0.0

- New type of output: Text. Allows you to enter formatted text and bullet lists to be displayed in PowerPoint slides or an online dashboard.
- Significance test highlighting and column comparisons are now customizable: arrow size, text color, text size and column names can be customized at specific p-value cutoffs.
- QScripts are new Office-like macros for Q, allowing you to modify variables and questions and customize your tables and plots using JavaScript.

## 4.2.20.0

## 4.2.19.0

## 4.2.14.0

- Improved varimax rotation algorithm in Principal Components Analysis.
- Fixed bug in significance testing for trend plots.
- Changed treatment of dates in Complicated Binary Variables. Missing values are now always returned as false. Previously treated as less than all other dates.

## 4.2.7.0

- Q can update the data in exports to PowerPoint and Excel, leaving your slides and changes intact.

## 4.2.6.0

- Many new plot types.
- Multiple data files can be loaded into a Q project.
- Time series can show two series from different data files (e.g. advertising spend and customer satisfaction).
- Upload a Q project to an online dashboard via File | Share as Dashboard. (Dashboards purchased separately.)
- The names of statistics can now be customised (e.g. 'NPS' instead of 'Average'); see the Output Text tab in either Project Options or Table Options.
- Statistical assumptions (including multiple comparisons method of significance testing) can now be set per table.
- Sorting by statistics now works on multiple tables at once.
- Office 2007 and later: titles are now included in exported charts. You can turn them off using templates.
- Improved support for dates and times in Excel-style and JavaScript formulas.
- Ability to copy selected cells from Q to Excel.
- Custom text may be entered into table footers.
- Coded questions now show spans and sub-NETs for parent codes in the codeframe.
- Text variables which are converted to Categorical variables by Q will now update seamlessly with new text responses.

## 4.1.8.0

- Shift+F3 changes case of selected labels similar to Word.

## 4.1.4.0

- New question type Pick Any - Compact.
- New Tables Tab feature - View | Hide Empty Rows/Columns.
- May edit time series analysis (smoothing) options in Date crosstabs.
- Ability to check out transferable licenses from Q.
- List of deleted cases that can be individually restored via the Data Tab.
- Trees can now handle all categorical and numeric variables.
- Improvements in planned significance tests.
- Pairwise comparisons between different columns in a table.
- Improvements in the way that sample size is presented.
- Moving averages on statistics in tables.

## 4.1.3.0

- Q runs in 64-bit mode on 64-bit editions of Windows; can handle bigger tables, bigger data files.
- Analysis outputs improved - may now copy and paste output tables to Excel.
- Merge Pick One and Pick Any questions by right-clicking in the Variables tab and selecting Insert Ready-Made Formula(s) | Merge Questions.
- You may Undelete categories from a table while still retaining any merging or netting already done.
- The method for computing the number of degrees of freedom in ranking questions has been modified; please e-mail Q Support for more information.
- Automatic tests of significance have been improved for Date questions; please e-mail Q Support for more information.

## 4.1.2.0

- You can drag variables, questions and tables between open Q projects. Not all types of variables and tables are supported.

## 4.1.1.0

- JavaScript variables are faster. (This includes banner questions.)
- Q can export to XtabML, used by E-Tabs.
- Statistical tests for the whole table may be exported to Excel (click File | Export to Office, then Advanced button).
- The Excel-style formula functions SEARCH, FIND and REPLACE have been added.
- New hotkey to lock and unlock tables (Ctrl + L).
- Coding: text responses can be sorted by number of responses.
- Exporting raw variables to Excel is much faster.

## 4.0.22.0

- Multiple correspondence analysis available by clicking Create | Map.

## 4.0.19.0

- Q can now use data from Triple-S (.sss) files.

## 4.0.18.0

- The new Q Online Training system is now available from the Help menu.

## 4.0.17.0

- Ability to automatically update license when renewal invoices are paid.
- No requirement to be Administrator when activating or deactivating license.
- Modified significance tests for Pick Any - Grid and Number - Grid questions (previously known as Two Dimensional questions); they now produce significance tests of NET and SUM.

## 4.0.16.0

- A new software licence agreement which contains a clause that protects both parties in regards to confidential information, so that you can send bug reports to us containing your client's data, and it will be kept in confidence.

## 4.0.0.0

- New question types: Experiment, Ranking, Date.
- Tools | Save Data as SPSS .sav File
- Tools | Convert .sps File to .sav
- Tools | Merge Data Files | Add New Cases
- Tools | Merge Data Files | Add New Variables
- Toggle multiple table statistics by holding down Ctrl.
- Segmentation (or classification) trees, in the Create menu.
- Time series plots, in the Create menu.
- Use <filter> and <weight> in table or group names, and they will be replaced with the actual filter and weight labels upon exporting.
- Spans included in Excel exports of tables.
- Copy and Paste Question(s) in the Variables and Questions tab.

## 3.3.8.0

- An important bug fix: in some tables, the 'Sample scaling constant' from 'Analysis | Statistical Assumptions...' was ignored.

## 3.3.0.0

- Shortcut (Ctrl + I) to select all variables in a question in the Variables and Questions tab.

## 3.2.6.0

- The way chart templates work has changed. Go to User Options and reconfigure them.
- Q works with Office 2007 chart templates.
- Chart templates can be set for individual tables.
- Effective sample size now computed for each cell in a table. Effective sample size shown at the bottom of the page is the largest of these.

## 3.2.0.0

- Medians, quartiles and trimmed averages.

## 3.1.1.0

- Support for Money-type variables and symbols.
- Extra missing values identified in SPSS files are now understood.
- Improved automatic setup of two-dimensional questions.
- Turn statistics on or off for multiple tables at once in the Report tree.
- Choose if and where SUMMARY/frequency tables get output from Basic Tables.
- New Variables and Questions Tab operation: Combine (Binary) Variables.
- Ability to find variables in Data Tab (shortcut: F5 key).

## 3.0.0.0

- A Data tab to view, edit and sort raw data, and to create variables using Excel formulas (e.g., =Q2+Q7).
- Multiple filters on a single table.
- Table styles, which control the appearance of tables on screen and when exported.
- Advanced regression, including an expert system to select the appropriate test.
- Latent class analysis, which forms segments using both numeric and categorical data.
- Rim weighting.
- Tidy categories in a table by sorting categories and Search/Replace...
- Q projects can be read into R.
- Convert an unstacked SPSS data file into a stacked SPSS data file.
- New categories can be added to a banner question.
- Spans can be used to show different groups of categories in a question (e.g., within a banner question).
- Additional variables can be pasted from Excel.

## 2.3.3.0

- Percentages in Q are exported to Excel as formatted percentages, when you display % signs in a table in Q. Please check any automated workbooks that use these values.

## 2.3.1.0

- Better handling of large data files.

## 2.0.2.0

- Row and column statistics.
- Exporting to Microsoft Word.
- Improved report print-outs/PDFs, with scaled bar graphs and cell colouring.
- Separate control of decimal places for each data series.
- Recoding ranges of variable values using rules.
- Cover pages on printed/PDF reports.
- Table of contents in printed/PDF reports and optionally in Excel exports.
- Customisable PowerPoint exports.

## 1.5.1.0

- Improved Cluster Analysis algorithm.

## 1.5.0.0

- Basic Tables feature, available from 'Advanced | Basic Tables...'

## 1.4.0.0

- Ability to import/export codes in your coded variables (via the right-click context menu in the code editor).
- Ability to 'tag' questions so they may be used as filters, weights, or be hidden from the blue and brown dropdowns.

## 1.3.0.0

- Ability to view simple statistics of all variables via View | Show Statistics.
- Ability to add Comments to variables in the Variables and Questions tab (via right-click menu).
- Coding module, now available to all licensed users.
- Moon plot option for Correspondence Analysis.
- Drag and dropping of variables in the Variables and Questions tab.

## 1.1.0.0

- Ability to export raw data using 'File | Export Selected Variables to Excel...'

## 1.0.2736.33301

- Customisation of print/PDF report fonts.
- Ability to export project and data file as a package to clients using 'File | Send Package'.

## 1.0.2712.17650

- Ability to control the layout of two-dimensional questions.
- Ability to rename variable labels using search and replace.

## 1.0.2693.19516

- Ability to provide templates for exporting charts to Microsoft Office.

## 1.0.2677.29899

- Ability to customize default Office chart types (available through 'Edit | Default Export Options...').
- Enhancements to Office exporting.

## 1.0.2657.20305

- Enhanced Office chart appearance.
- Ability to 'link' variables (right-click on a variable and select 'Construct New Variable(s) | Linked...').
- Faster start-up time.
- Web version check, available through 'Help | Check for New Version'.
- Locking of Q project files during editing - prevent other people editing your project at the same time.
- Ability to tell Q what to do with new variables from a data file.

## 1.0.2603.22342

- Compatibility with Windows Vista.

## 1.0.2586.20118

- License information is now shown in the About dialog.

## 1.0.2585.29173

- Currently selected filter is now included in new filter when you press the "Create filter variable" toolbar button.
- New icons for menus.

## 1.0.2573.27532

- New icons for all types of variables and questions.

## 1.0.2545.21451

- New Smart Tables feature, available from the Advanced menu.
- Type-ahead search boxes - added to the Variables and Questions list, as well as the Binary and Computed variable constructors.
- Improved appearance of toolbars.

## 1.0.2532.29985

- Added Computed Text Variables, so you can combine multiple text variables into one by joining together their responses (see Help for an example).

## 1.0.2526.17385

- Added ability to create filters from selected table cells (click on the filter toolbar button).
- Added ability to import SPSS for Mac files.

## 1.0.2508.20679

- "Create Banner Question" added; available from the Advanced menu.

## 1.0.2490.20515

- Improved handling of changes to data files.
- Online help added.

## 1.0.2477.32121

- Updated manuals, tutorials and reference guides.
- Improved validation when constructing variables.

Further reading: Key Driver Analysis Software