What's new in IBM SPSS Statistics Subscription

September 2022 update

Analyze procedures

Linear OLS Alternatives

Elastic Net

Click Analyze > Regression > Linear OLS Alternatives > Elastic Net to obtain a Linear Elastic Net Regression analysis. The new Linear Elastic Net extension procedure uses the Python sklearn.linear_model.ElasticNet class to estimate regularized linear regression models for a dependent variable on one or more independent variables. Regularization combines L1 (Lasso) and L2 (Ridge) penalties. The extension includes optional modes to display trace plots for different values of alpha for a given L1 ratio, and to select the L1 ratio and alpha hyperparameter values based on cross validation. When a single model is fitted or cross validation is used to select the penalty ratio and/or alpha, a partition of holdout data can be used to estimate out-of-sample performance.

Lasso

Click Analyze > Regression > Linear OLS Alternatives > Lasso to obtain a Linear Lasso Regression analysis. The new Linear Lasso extension procedure uses the Python sklearn.linear_model.Lasso class to estimate L1 loss regularized linear regression models for a dependent variable on one or more independent variables, and includes optional modes to display trace plots and to select the alpha hyperparameter value based on cross validation. When a single model is fitted or cross validation is used to select alpha, a partition of holdout data can be used to estimate out-of-sample performance.

Ridge

Click Analyze > Regression > Linear OLS Alternatives > Ridge to obtain a Linear Ridge Regression analysis. The new Linear Ridge extension procedure uses the Python sklearn.linear_model.Ridge class to estimate L2 or squared loss regularized linear regression models for a dependent variable on one or more independent variables, and includes optional modes to display trace plots and to select the alpha hyperparameter value based on cross validation. When a single model is fitted or cross validation is used to select alpha, a partition of holdout data can be used to estimate out-of-sample performance.

Parametric Accelerated Failure Time (AFT) Models

Click Analyze > Survival > Parametric Accelerated Failure Time (AFT) Models to obtain a Parametric Accelerated Failure Time (AFT) Model analysis, which invokes the parametric survival models procedure with nonrecurrent life time data. Parametric survival models assume that survival time follows a known distribution, and this analysis fits accelerated failure time models with their model effects proportional with respect to survival time.

Pseudo-R² measures in Linear Mixed Models and Generalized Linear Mixed Models

Pseudo-R² measures and the intra-class correlation coefficient are now included in Linear Mixed Models and Generalized Linear Mixed Models output (when appropriate). The coefficient of determination R² is a commonly reported statistic because it represents the proportion of variance that is explained by a linear model. The intra-class correlation coefficient (ICC) is a related statistic that quantifies the proportion of variance that is explained by a grouping (random) factor in multilevel/ hierarchical data.

Command syntax

GENLINMIXED

The output now includes pseudo-R² measures and the intra-class correlation coefficient (when appropriate).

LINEAR_ELASTIC_NET

The new extension command uses the Python sklearn.linear_model.ElasticNet class to estimate regularized linear regression models for a dependent variable on one or more independent variables.

LINEAR_LASSO

The new extension command uses the Python sklearn.linear_model.Lasso class to estimate L1 loss regularized linear regression models for a dependent variable on one or more independent variables. The command includes optional modes to display trace plots and to select the alpha hyper-parameter value that is based on cross-validation.

LINEAR_RIDGE

The new extension command uses the Python sklearn.linear_model.Ridge class to estimate L2 or squared loss regularized linear regression models for a dependent variable on one or more independent variables. The command includes optional modes to display trace plots and to select the alpha hyper-parameter value that is based on cross-validation.

MIXED

The output now includes pseudo-R² measures and the intra-class correlation coefficient (when appropriate).

SURVREG AFT

The new command invokes the parametric survival models procedure with nonrecurrent life-time data.

Python and R upgrades

Python 3.10.4 and R 4.2.0 are a part of the IBM® SPSS® Statistics Subscription.

Select Cases - hidden cases

Unselected cases are no longer hidden in the Data Editor when a subset of cases is selected, and the unselected cases are not discarded. This represents a return to the behavior of the November 2020 update and earlier updates.

Violin plots

The Graph board Template Chooser includes a new violin plot, which is a hybrid of the box and kernel density plots. Violin plots show peaks in the data and are used to visualize the distribution of numerical data. Unlike a box plot that can show only summary statistics, violin plots depict summary statistics and the density of each variable.

Workbook mode enhancements

• Two new workbook toolbar items: Show/Hide all syntax windows and Clear all output.
• New button on the Status bar to switch between Classic (Output and Syntax) and Workbook modes.

Search enhancements

The Search feature now provides options for entering terms directly in a toolbar field and for viewing results in a drop-down pane.

November 2021 update

Analyze procedures

Kernel Ridge Regression

The new extension-based procedure uses the Python sklearn.kernel_ridge.KernelRidge class to estimate a kernel ridge regression of a dependent variable on one or more independent variables. The independent variables include model hyperparameters, or a selection of hyperparameter values, over a specified grid of values. Cross validation is achieved by using the sklearn.model_selection.GridSearchCV class.

Linear Mixed Models

A new output table for the procedure provides the marginal and conditional pseudo-R² measures. The table displays only in instances where appropriate.

Power Analysis procedures

The new precision feature computes the sample sizes that are required to estimate the population parameter with precision that is determined by user-specified confidence interval half-widths. The expected result produces the minimum sample size to ensure that the actual confidence interval half-widths do not exceed the desired values.

Note: The new feature is available for all Power Analysis procedures except Univariate Linear Regression.

Effect size as an input to the estimation of the power or sample size is now supported. The defined effect size value is passed to the intermediate step in the procedure and calculates the desired power or sample size. The following Power Analysis procedures support effect size as an input to the estimation of the power or sample:

• Power Analysis of One-Sample t Test
• Power Analysis of Paired-Samples t Test
• Power Analysis of Independent-Samples t Test
• Power Analysis of One-Way ANOVA
• Power Analysis of Univariate Linear Regression Test

Command syntax

OUTPUT CREATE

The new command provides options for creating custom tables, charts, and other output items from user-entered JSON or from an external *.json file. For more information, see OUTPUT CREATE.

POWER ONEWAY ANOVA

The CONTRAST subcommand's new HALFWIDTH keyword estimates the sample size based on specified confidence interval half-widths. For more information, see CONTRAST Subcommand (POWER ONEWAY ANOVA command).

The PARAMETERS subcommand's new ES and keyword specifies the effect size of the overall test, which is measured by either f or η². For more information, see PARAMETERS Subcommand (POWER ONEWAY ANOVA command).

The PLOT subcommand's new ES, ES_YAXIS, and ES_XAXIS keywords control the two-dimensional power by effect size chart, the three-dimensional power by total sample size (x-axis) and effect size (y-axis) chart, and the three-dimensional power by total sample size (y-axis) and effect size (x-axis) chart. For more information, see PLOT Subcommand (POWER ONEWAY ANOVA command).

POWER MEANS INDEPENDENT

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER MEANS INDEPENDENT command).

The PARAMETERS subcommand's new ES keyword specifies the effect size as an input to the estimation of the power or sample size. When the two independent groups for comparison are assumed to have unequal variances, the effect size of the independent-sample analysis is measured by the mean difference. For more information, see PARAMETERS Subcommand (POWER MEANS INDEPENDENT command).

POWER MEANS ONESAMPLE

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER MEANS ONESAMPLE command).

The PARAMETERS subcommand's new ES keyword specifies the effect size as an input to the estimation of the power or sample size. For more information, see PARAMETERS Subcommand (POWER MEANS ONESAMPLE command).

POWER MEANS RELATED

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER MEANS RELATED command).

The PARAMETERS subcommand's new ES keyword specifies the effect size as an input to the estimation of the power or sample size. For more information, see PARAMETERS Subcommand (POWER MEANS RELATED command).

POWER PARTIALCORR

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER PARTIALCORR command).

POWER PEARSON ONESAMPLE

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER PEARSON ONESAMPLE command).

POWER PROPORTIONS INDEPENDENT

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER PROPORTIONS INDEPENDENT command).

POWER PROPORTIONS ONESAMPLE

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER PROPORTIONS ONESAMPLE command).

POWER PROPORTIONS RELATED

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER PROPORTIONS RELATED command).

POWER SPEARMAN ONESAMPLE

The new PRECISION subcommand estimates the sample size based on specified confidence interval half-widths. For more information, see PRECISION Subcommand (POWER SPEARMAN ONESAMPLE command).

POWER UNIVARIATE LINEAR

The PARAMETERS subcommand's new ES keyword specifies the effect size value that is measured by f². For more information, see PARAMETERS Subcommand (POWER UNIVARIATE LINEAR command).

SAVE DATA COLLECTION

The command has been deprecated.

Pivot table enhancements

Most tables contain a mix of values, and applying a heat map to an entire table typically produces tables with wildly varying ranges. The pivot table editor now includes the menu option Color Scales, which provides heat map style settings that display selected table cells in varying colors based on the cell values. For more information, see Color scales.

Workbook mode enhancements

Context menu

The right-click context menu now provides options for cutting, copying, and pasting content, and for displaying the Style Output dialog, which specifies changes to make to selected output objects in the Workbook. For more information, see Style Output: Select.

Syntax paragraph error pane

Syntax-related error information now displays under the syntax paragraph.

Proxy settings

A proxy.ini configuration file is now installed with the product and provides options for manually configuring proxy settings. For more information, see Proxy configuration file.

Documentation

The "Exporting to Data Collection" topic has been removed as IBM SPSS Statistics no longer supports UNICOM Intelligence (formerly IBM SPSS Data Collection).

May 2021 update

Analyze procedures

Meta Analysis

Meta analysis is the analysis of the data obtained from a collection of studies that answer similar research questions. These studies are known as primary studies. Meta analysis uses statistical methods to produce an overall estimate of an effect, explore between-study heterogeneity, and investigate the impact of publication bias or, more generally, small-study effects on the final results.

The following Meta Analysis procedures are new in IBM SPSS Statistics Subscription.

Meta Analysis Continuous

Performs meta-analysis with continuous outcomes on raw data that are provided in the active dataset for the estimation of the effect size.

Meta Analysis Continuous Effect Size

Performs meta-analysis with continuous outcomes when the pre-calculated effect size data are provided in the active data set.

Meta Analysis Binary

Performs meta-analysis with binary outcomes on raw data that are provided in the active dataset for the estimation of the effect size.

Meta Analysis Binary Effect Size

Performs meta-analysis with binary outcomes when the pre-calculated effect size data are provided in the active data set.

Meta Analysis Regression

Performs meta-analysis regression.

General Linear Model (GLM) procedures

The General Linear Model (GLM) procedure user interfaces now provide a Compare simple main effects setting on their EM Means dialogs. The setting is enabled whenever the target list contains one or more product or interaction effects (for example, A*B, A*B*C). The setting supports the specification of comparisons among simple main effects, which are main effects nested within the levels of other factors.

One-Way ANOVA

The procedure now supports non-numeric categorical variables.

Power Analysis

The new Grid Values dialog provides options for specifying a range of POWER values for the purpose of viewing projected sample sizes in a grid format for every specified POWER range value.

The Grid Values dialog is available for every Power Analysis procedure when the Estimate sample size and Grid power values options are selected (click the Grid control to display the dialog).

Ratio Statistics

Price-Related Bias (PRB)

The procedure now supports the Price-Related Bias (PRB) dispersion method. PRB is an index of whether assessment to price ratios are systematically higher or lower for higher-priced properties. PRB regresses percentage differences in assessment ratios. The differences are drawn from the median ratio on the base 2 logarithms of value proxy measures. The proxy measures are calculated as the "averages of sales prices" and the "ratios of assessed values to the median ratio". The method also gives the percentage change in assessment ratios for a 100 percent change in value.

Coefficient of Variation (COV)

The new COV dispersion method includes the median and mean-centered coefficients of variation and effectively replaces the Median Centered COV and Mean Centered COV dispersion methods. The median-centered coefficient of variation is the result of expressing the root mean squares of deviation from the median as a percentage of the median. The mean-centered coefficient of variation is the result of expressing the standard deviation as a percentage of the mean.

Command syntax

COXREG

The CONTRAST subcommand's DEVIATION keyword now defaults the refcat to the first category. For more information, see CONTRAST Subcommand (COXREG command).

LOGISTIC REGRESSION

The CONTRAST subcommand's DEVIATION keyword now defaults the refcat to the first category. For more information, see CONTRAST Subcommand (LOGISTIC REGRESSION command).

META BINARY command

The new command represents the meta-analysis procedure for binary outcomes when the raw data are provided in the active dataset for the estimation of the effect size. For more information, see META BINARY.

META ES BINARY command

The new command represents the meta-analysis procedure for binary outcomes when the pre-calculated effect size data are provided in the active data set. For more information, see META ES BINARY.

META CONTINUOUS command

The new command represents the meta-analysis procedure for continuous outcomes when the raw data are provided in the active dataset for the estimation of the effect size. For more information, see META CONTINUOUS.

META ES CONTINUOUS command

The new command represents the meta-analysis procedure for continuous outcomes when the pre-calculated effect size data are provided in the active data set. For more information, see META ES CONTINUOUS.

META REGRESSION command

The new command represents the meta-regression procedure. For more information, see META REGRESSION.

RATIO STATISTICS

• COV and PRB keywords added to the OUTFILE subcommand.
• COV, PRB, and N keywords added to the PRINT subcommand.

For more information, see RATIO STATISTICS.

Relationship Maps

Relationship maps are useful for determining how variables relate to each other by providing a visual representation of the connections and influences that each node and link has over each other. Relationship maps visually represent connections and influences through nodes and links. Nodes represent variables and variable categories; links represent the strength of influence between nodes. Larger nodes and thicker link lines represent stronger connections and influence. Smaller nodes and thinner link lines represent weaker connections and influence.

The Relationship maps feature is accessed through Graphs > Relationship Map...

R

R 4.2.0 is now part of the IBM SPSS Statistics. The R environment settings are defined in Edit > Options... > File Locations > R Location.

Python 3 and R programmability

Support for Python 3 and R has been enhanced by enabling an easily configurable virtual runtime environment.

The Python runtime environment is accessed by clicking the Python 3 IDLE (PythonGUI) (Windows) or Python 3 for SPSS Statistics (macOS) option in the product folder.

Windows

Start > IBM SPSS Statistics > Python 3 IDLE (PythonGUI)

macOS

> Applications > IBM SPSS Statistics > Python 3 for SPSS Statistics

Note: Python 2 is no longer officially supported. If you still need to run Python 2, refer to the Programmability SDK.

The R runtime environment is accessed by clicking the R x64 4.0.5 (Windows) or R for SPSS Statistics (macOS) option in the product folder.

Windows

Start > IBM SPSS Statistics > R x64 4.0.5

macOS

> Applications > IBM SPSS Statistics > R for SPSS Statistics

Installation and licensing

The product installer has been updated to provide the option registering either the subscription or licensed version of IBM SPSS Statistics.

Subscription

Requires an IBMid to activate and install the subscription-based version of the software. You must purchase IBM SPSS Statistics Subscription in order to activate the product via the subscription method.

Licensed

Requires an authorized user license or concurrent user license to activate the software. You must purchase an on-premise license for IBM SPSS Statistics in order to activate the product via a user license or concurrent user license.

For more information on the differences between the subscription and licensed versions, see Which IBM SPSS Statistics version is right for you?

Refer to the following introductory video for a brief overview on the installer updates:

Output enhancements

Workbooks

Viewing output in Workbook mode bridges the SPSS Statistics syntax editing ability with a notebook approach that provides an interactive method for running syntax and viewing the corresponding output. Workbook documents (*.spwb) consist of individual paragraphs. The paragraphs contain the output elements (syntax, tables, charts, and so on). Syntax paragraphs provide full syntax edit and run capabilities. Rich text paragraphs provide full rich-text editing capabilities.

Chart and table editor usability enhancements

Pivot table editor

The pivot table editor user interface now includes edit options slide-out pane on the right side of the dialog. The pane provides options for handling rows and columns, specifying text attributes, defining border parameters, specifying cells formats, and defining footnotes and table comments.

Installed extensions

Additional commonly used extensions are now automatically installed with the product. The installed extensions can be identified by the plus symbol next to their menu entry (for example, ).

Search enhancements

The search feature has been updated to now provide results for procedures, help topics, syntax reference, and case studies. The Search feature now searches all words/terms in each user interface dialog and help topic.

Refer to the following introductory video for a brief overview on the search enhancements:

Command syntax help has been updated to provide tool tips that provide syntax examples when hovering the cursor over commands and subcommands in the Syntax Editor.

Export output enhancements

Word Document (*.docx)

You can now export output to Microsoft Word (*.docx) format.

Text - Plain (*txt), Text - UTF8 (*txt), and Text - UTF16 (*txt)

The text export settings are now divided into three distinct options that provide different encoding methods.

Excel output

The Microsoft Excel export settings now provide options for creating both workbooks and worksheets.

Print preview

The File > Print Preview provides a PDF formatted preview version of the output.

Select Cases - hidden cases

By default, unselected cases are now hidden in the Data Editor when a subset of cases is selected, and the unselected cases are not discarded. Hidden cases are not picked up when rows are copied from the Data Editor.

You can select to display hidden cases by deselecting Edit > Hide excluded cases, or by right-clicking in the Data Editor and deselecting the Hide excluded cases option.

Chart Builder usability enhancements

The template controls in the Chart Appearance tab have been redesigned to streamline template selection options.

Accessibility

The user interface now supports high contrast mode, which adjusts the background and text colors to make the application easier to read.

November 2020 update

Analyze procedures

Bivariate Correlations

The procedure has been updated to provide the option of suppressing the correlations table from the output. The procedure also now provides options for controlling the estimation of the confidence intervals.

Independent-Samples Proportions

The new procedure provides tests and confidence intervals for the difference in two independent binomial proportions. Output includes observed proportions, estimates of differences in population proportions, asymptotic standard errors of population differences under null and alternative hypotheses, specified test statistics with two-sided probabilities, and specified confidence intervals for differences in proportions.

One-Sample Proportions

The new procedure provides tests and confidence intervals for individual binomial proportions. Output includes the observed proportion, the estimate of the difference between the population proportion and the hypothesized population proportion, asymptotic standard errors under null and alternative hypotheses, specified test statistics with two-sided probabilities, and specified confidence intervals for proportions.

Paired-Samples Proportions

The new procedure provides tests and confidence intervals for the difference in two related or paired binomial proportions. Output includes observed proportions, estimates of differences in population proportions, asymptotic standard errors of population differences under null and alternative hypotheses, specified test statistics with two-sided probabilities, and specified confidence intervals for differences in proportions.

Reliability Analysis

The procedure has been updated to provide the Omega (McDonald’s Omega) model option. This model assumes that the model is uni-dimensional including a single factor with no local item dependence in the form of error covariances. The model implies that the covariance of the two different items is the product of their loadings.

Command enhancements

CORRELATIONS command

Added support for the NOMATRIX keyword in the PRINT subcommand. The keyword suppresses the correlations table from the output. For more information, see PRINT Subcommand (Correlations command).

Added support for the CI subcommand. The subcommand controls the estimation of the confidence intervals. For more information, see CI Subcommand (Correlations command).

MULTIPLE IMPUTATION command

Added support for specifying a single numeric parameter in the IMPUTE subcommand's SCALEMODEL keyword PMM method. The imputed value is based on the value defined for the closest randomly selected complete case from the closest (k) predictions, where (k) is a positive integer with a default value of 5. For more information, see IMPUTE Subcommand (MULTIPLE IMPUTATION command).

PORPORTIONS command

The new PROPORTIONS command computes tests and confidence intervals for binomial proportions or differences of proportions. Statistics are available for one-sample proportions (tested against a specified value), paired samples (different variables), or independent samples (different groups of cases).For more information, see PROPORTIONS, ONESAMPLE subcommand, PAIREDSAMPLES subcommand, and INDEPENDENTSAMPLES subcommand.

RELIABILITY command

Added support for the OMEGA keyword in the MODEL subcommand. The keyword provides an estimation of McDonald’s Omega to evaluate reliability. For more information, see MODEL Subcommand (RELIABILITY command).

Restore points

Restore points save data from active sessions that either quit unexpectedly (automatic recovery) or that you explicitly save. Each restore point is an SPSS Statistics session snapshot. Each restore point contains Data Editor, syntax, and output file information that was active at the time the session either quit unexpectedly or that you explicitly saved. Saved restore points remain in a backed-up state until you either restore or delete them.

Output enhancements

Export SVG charts

You can now export charts to Scalar Vector Graphics (*.svg) format.

Chart and table editor usability enhancements

• A Reset button was added to both the chart and table editors. The button resets the chart/table to its original configuration.
• The table editor toolbar has been split into Edit and Format toolbars.
• Increase Decimals and Decrease Decimals toolbar controls are now available. The controls allow you to increase or decrease the decimal place settings in tables.

APA style enhancements

• Footnotes and captions can now be double-spaced.
• Footnote alignment issues have been fixed.
• Table footnotes and captions can now be disabled.
• Chart spacing and alignment issue have been addressed.
• Small significance values can now be represented with "<0.001".

June 2020 update

Packaging

The Bootstrapping and Data Preparation features are now included in the IBM SPSS Statistics Base edition (Bootstrapping was previously included in Custom Tables and Advanced Statistics; Data Preparation was previously included in Sampling and Testing).

Auto-Recovery

Automatic recovery is designed to recover unsaved files and content in instances where the application quits unexpectedly. You can select to enable/disable the automatic recovery feature (the feature is enabled by default), select a time interval (in minutes) between saving files, and view or change the auto-recovery file location. For more information, see General options.

Upon relaunching SPSS Statistics after an unexpected exit, you are presented with an IBM SPSS Statistics error report, which allows you to enter information about your session prior to the unexpected exit. After leaving the exit report, you are presented with the Auto-Recovery dialog, which provides options for recovering prior session data or deleting the saved session data. For more information, see Auto-Recovery.

Privacy settings

The Options dialog now includes a Privacy tab the provides options for:

• Allow the SPSS Statistics application share information with IBM.
• Enable or disable SPSS Statistics from retrieving Welcome dialog content updates.
• Enable or disable SPSS Statistics from sending error reports to IBM.

For more information, see Privacy options.

Issue reporter

The Help menu now provides a Report Issue link that launches the IBM SPSS Statistics Issue Reporter dialog. The dialog allows you to enter information regarding any issues you may encounter when using the product. The information you enter is sent to IBM for use in improving the product.

Native macOS file selection dialogs

The file selection dialogs in the macOS version of SPSS Statistics have historically been heavily customized to accommodate specific SPSS Statistics file features. You now have the option of enabling native macOS file selection dialogs (via Edit > Options... > General > Windows > Display native macOS file dialogs). Native macOS file dialogs provide the following benefits:

• All of the benefits of native macOS file selection dialogs are available (for example, search, sidebar shortcuts, keyboard shortcuts, and so on).
• The SPSS Statistics file selection dialogs are consistent with other macOS file selection dialogs.

Analyze procedures

Bivariate Correlations

The Show only the lower triangle setting was added to main dialog. When the setting is enabled only the correlation matrix table's lower triangle is presented in the output. When not selected, the full correlation matrix table is presented in the output. The setting was introduced to allow table output to adhere to APA style guidelines. For more information, see Bivariate Correlations.

Crosstabs

The Create APA style table settings was added to Cell Display dialog. The setting produces a table that adheres to APA style guidelines. For more information, see Crosstabs cell display.

Frequencies

The Create APA style tables settings was added to main dialog. The setting produces tables that adhere to APA style guidelines. For more information, see Frequencies.

Power Analysis

Power analysis plays a pivotal role in a study plan, design, and conduction. The calculation of power is usually before any sample data have been collected, except possibly from a small pilot study. The precise estimation of the power may tell investigators how likely it is that a statistically significant difference will be detected based on a finite sample size under a true alternative hypothesis. If the power is too low, there is little chance of detecting a significant difference, and non-significant results are likely even if real differences truly exist. The new procedures are grouped as follows.

Means

One-Sample T-Test

In one-sample analysis, the observed data are collected as a single random sample. It is assumed that the sample data independently and identically follow a normal distribution with a fixed mean and variance, and draws statistical inference about the mean parameter. For more information, see Power Analysis of One-Sample T Test.

Independent-Samples T-Test

In independent-samples analysis, the observed data contain two independent samples. It is assumed that the data in each sample independently and identically follow a normal distribution with a fixed mean and variance, and draws statistical inference about the difference of the two means. For more information, see Power Analysis of Independent-Samples T Test.

Paired-Samples T-Test

In paired-samples analysis, the observed data contain two paired and correlated samples, and each case has two measurements. It is assumed that the data in each sample independently and identically follow a normal distribution with a fixed mean and variance, and draws statistical inference about the difference of the two means. For more information, see Power Analysis of Paired-Samples T Test.

One-Way ANOVA

Analysis of variance (ANOVA) is a statistical method of estimating the means of several populations which are often assumed to be normally distributed. The One-way ANOVA, a common type of ANOVA, is an extension of the two-sample t-test. The procedure provides approaches for estimating the power for two types of hypothesis to compare the multiple group means, the overall test, and the test with specified contrasts. The over test focuses on the null hypothesis that all group means are equal. The test with specified contrasts breaks down the overall ANOVA hypotheses into smaller but more describable and useful pieces of the means. For more information, see Power Analysis of One-Way ANOVA.

Proportions

One-Sample Binomial Test

The binomial distribution is based on a sequence of Bernoulli trials. It can be used to model those experiments including a fixed number of total trials that are assumed to be independent of each other. Each trial leads to a dichotomous result, with the same probability for a "successful" outcome.

The one-sample binomial test makes statistical inference about the proportion parameter by comparing it with a hypothesized value. The methods for estimating the power for such a test are either the normal approximation or the binomial enumeration. For more information, see Power Analysis of One-Sample Binomial Test.

Related-Samples Binomial Test

The binomial distribution is based on a sequence of Bernoulli trials. It can be used to model those experiments including a fixed number of total trials that are assumed to be independent of each other. Each trial leads to a dichotomous result, with the same probability for a "successful" outcome.

The related-sample binomial estimates the power of McNemar's test to compare two proportion parameters based on the matched pair subjects sampled from two related binomial populations. For more information, see Power Analysis of Related-Sample Binomial Test.

Independent-Samples Binomial Test

The binomial distribution is based on a sequence of Bernoulli trials. It can be used to model those experiments including a fixed number of total trials that are assumed to be independent of each other. Each trial leads to a dichotomous result, with the same probability for a "successful" outcome.

The independent-sample binomial test compares two independent proportion parameters. For more information, see Power Analysis of Independent-Sample Binomial Test.

Correlations

Pearson Product-Moment

Pearson's product-moment correlation coefficient measures the strength of linear association between two scale random variables that are assumed to follow a bivariate normal distribution. By convention, it is a dimensionless quantity and obtained by standardizing the covariance between two continuous variables, thereby ranging between -1 and 1.

The test uses Fisher's asymptotic method to estimate the power for the one-sample Pearson correlation. For more information, see Power Analysis of One-Sample Pearson Correlation Test.

Spearman Rank-Order

Spearman rank-order correlation coefficient is a rank-based nonparametric statistic to measure the monotonic relationship between two variables that are usually censored and not normally distributed. The Spearman rank-order correlation is equal to the Pearson correlation between the rank values of the two variables, thereby also ranging between -1 and 1. Detecting the power of the Spearman rank correlation test is an important topic in the analysis of hydrological time series data.

The test uses Fisher's asymptotic method to estimate the power for the one-sample Spearman rank-order correlation. For more information, see Power Analysis of One-Sample Spearman Correlation Test.

Partial

Partial correlation can be explained as the association between two random variables after eliminating the effect of another or several other variables. It is a useful measurement in the presence of confounding. Similar to the Pearson correlation coefficient, partial correlation coefficient is also a dimensionless quantity ranging between -1 and 1.

The test uses Fisher's asymptotic method to estimate the power for the one-sample Pearson correlation. For more information, see Power Analysis of Partial Pearson Correlation Test.

Regression

Univariate Linear

Univariate linear regression is a basic and standard statistical approach in which researchers use the values of several variables to explain or predict values of a scale outcome.

The test invokes power analysis for the type III F-test in univariate linear regression. For more information, see Power Analysis of Univariate Linear Regression Test.

Command enhancements

CORRELATIONS command

Added support for the FULL, LOWER, and LNODIAG keywords in the PRINT subcommand. The keywords control the display of the correlation matrix table's lower triangle or the full correlation matrix table. The keywords were introduced to allow table output to adhere to APA style guidelines. For more information, see PRINT Subcommand (Correlations command).

MATRIX-END MATRIX command

• The NCDF.BETA cumulative distribution function is now supported.
• Probability density functions are now supported (they were previously only supported by the COMPUTE command).
• Tail probability functions are now supported (they were previously only supported by the COMPUTE command).
• Random variable functions are now supported (they were previously only supported by the COMPUTE command).

For more information, see MATRIX-END MATRIX.

NONPAR CORR command

Added support for the FULL, LOWER, and LNODIAG keywords in the PRINT subcommand. The keywords control the display of the correlation matrix table's lower triangle or the full correlation matrix table. The keywords were introduced to allow table output to adhere to APA style guidelines. For more information, see PRINT Subcommand (NONPAR CORR command).

NPTESTS command

CRITERIA subcommand

The SEED keyword is now supported. The keyword resets the random seed used for the Monte Carlo sampling.

ONESAMPLE subcommand

The KOLMOGOROV_SMIRNOV keyword now supports the following Lilliefors test for Monte Carlo sampling settings:

NSAMPLES keyword

Resets the number of replicates used by the Lilliefors test for Monte Carlo sampling.

MC_CILEVEL keyword

Resets the confidence interval level that is estimated by the Kolmogorov-Smirnov test.

SIMULATION keyword

Controls whether the Monte Carlo simulation will be used to conduct the Lilliefors test for Normal distribution when the parameters are not specified.

POISSON keyword

The SAMPLE setting has been removed from the POISSON keyword.

For more information, see NPTESTS.

NPAR TESTS command

KS_SIM subcommand

The KS_SIM subcommand is now supported. KS_SIM (KOLMOGOROV-SMIRNOV simulation) controls the parameters for the Monte Carlo simulation for Normal, Uniform, and Exponential distributions. The new subcommand supports the following Lilliefors test for Monte Carlo sampling keywords:

CIN keyword

Resets the estimated confidence interval level used by the Kolmogorov-Smirnov test (using the Monte Carlo simulations).

SAMPLES keyword

Resets the number of replicates used by the Lilliefors test for Monte Carlo sampling.

NONORMAL keyword

When specified, results will not include the Monte Carlo sampling for Normal distribution.

K-S subcommand

POISSON=varlist is no longer supported.

For more information, see NPAR TESTS.

OMS

• FORMAT=REPORTHTML and FORMAT=REPORTMHT deprecated from the DESTINATION subcommand. The subcommand syntax has been mapped to the HTML subcommand.
• REPORTTITLE keyword deprecated from the DESTINATION subcommand.

For more information, see OMS.

ONEWAY command

The CRITERIA and ES subcommands are now supported by the ONEWAY command:

CRITERIA subcommand

The optional subcommand controls the significance level to estimate the confidence intervals.

ES subcommand

The optional subcommand controls the effect size estimation by providing keywords for controlling the effect size calculation for the overall test, and controlling the calculation of the contrast test effect size.

For more information, see ONEWAY.

OUTPUT EXPORT

Support for the REPORT subcommand has been deprecated. The REPORT subcommand syntax has been mapped to the HTML subcommand. For more information, see OUTPUT EXPORT.

OUTPUT MODIFY

• Added support for the PIVOT keyword in the TABLES subcommand. The keyword pivots the specified row dimension to the specified column dimension. Any existing column dimensions are incremented outwards. The keyword was introduced to allow table output to adhere to APA style guidelines.
• Added support for the HIDE and UNGROUP keywords in the TABLECELLS subcommand. HIDE suppresses the selected row or column; UNGROUP deletes the selected row or column group header. The keywords were introduced to allow table output to adhere to APA style guidelines.
• Added support for the PARENT and CHILD options for the SELECTCONDITION keyword in the TABLECELLS subcommand. Both options specify primary and secondary string conditions to apply changes within the area of the table specified by the SELECT keyword.
• Added support for the VALID, TOTAL, MISSING, CUMULATIVEPERCENT, and VALIDPERCENT options for the SELECTCONDITION keyword in the TABLECELLS subcommand.

For more information, see OUTPUT MODIFY.

OUTPUT SAVE

The TYPE subcommand's SPW option has been deprecated. For more information, see OUTPUT SAVE.

POWER ONEWAY ANOVA command

The new command estimates the power for two types of hypothesis to compare the multiple group means, the overall test, and the test with specified contrasts. The over test focuses on the null hypothesis that all group means are equal. The test with specified contrasts breaks down the overall ANOVA hypotheses into smaller but more describable and useful pieces of the means. For more information, see POWER ONEWAY ANOVA.

POWER MEANS INDEPENDENT command

The new command invokes power analysis for the independent sample t-test to draw statistical inference about the difference of the two means. For more information, see POWER MEANS INDEPENDENT.

POWER MEANS ONESAMPLE command

The new command invokes power analysis for the one sample t-test to draw statistical inference about the mean parameter. For more information, see POWER MEANS ONESAMPLE.

POWER MEANS RELATED command

The new command invokes power analysis for the related sample t-test to draw statistical inference about the difference of the two means. For more information, see POWER MEANS RELATED.

POWER PARTIALCORR command

The new command invokes the power analysis for the one-sample partial correlation test. Partial correlation can be explained as the association between two random variables after eliminating the effect of another or several other variables. It is a useful measurement in the presence of confounding. For more information, see POWER PARTIALCORR.

POWER PEARSON ONESAMPLE command

The new command invokes the power analysis for the one-sample Pearson correlation test. Pearson product-moment correlation coefficient measures the strength of linear association between two scale random variables, which are assumed to follow a bivariate normal distribution. For more information, see POWER PEARSON ONESAMPLE.

POWER PROPORTIONS INDEPENDENT command

The new command invokes the power analysis for the independent-sample binomial test to compare two independent proportion parameters. For more information, see POWER PROPORTIONS INDEPENDENT.

POWER PROPORTIONS ONESAMPLE command

The new command invokes power analysis for the one-sample binomial test to make statistical inference about the proportion parameter by comparing it with a hypothesized value. For more information, see POWER PROPORTIONS ONESAMPLE.

POWER PROPORTIONS RELATED command

The new command invokes power analysis for the related-sample binomial test (or McNemar's test) to compare two proportion parameters based on the matched pair subjects sampled from two related binomial populations. For more information, see POWER PROPORTIONS RELATED.

POWER SPEARMAN ONESAMPLE command

The new command invokes the power analysis for the one-sample Spearman rank-order correlation test. Spearman rank-order correlation coefficient is a rank-based nonparametric statistic to measure the monotonic relationship between two variables that are usually censored and not normally distributed. For more information, see POWER SPEARMAN ONESAMPLE.

POWER UNIVARIATE LINEAR command

The new command invokes power analysis for the type III F-test in univariate linear regression. Univariate linear regression is a basic and standard statistical approach in which researchers use the values of several variables to explain or predict values of a scale outcome. For more information, see POWER UNIVARIATE LINEAR.

QUANTILE REGRESSION command

CRITERIA subcommand

The QUANTILE keyword now provides support for a grid of quantiles (connected by the keywords TO and BY). The quantile grid can be mixed with other quantiles, and can be placed anywhere. For more information, see CRITERIA Subcommand (QUANTILE REGRESSION command).

T-TEST command

The ES subcommand is now supported:

ES subcommand

The optional subcommand controls the effect size estimation by providing keywords for controlling the printing of the effect size calculation for the overall test, and controlling how the standardizer is computed in estimating the Cohen's d and Hedges' correction for each variable pair (only for Paired-Samples T Test). For more information, see ES Subcommand (T_TEST command).

WEIGHTED KAPPA command

Cohen’s kappa statistic is broadly used in cross-classification as a measure of agreement between two observed raters. It is an appropriate index of agreement when ratings are nominal scales with no order structure. The new WEIGHTED KAPPA command is an important generalization of the kappa statistic that measures the agreement of two ordinal subjects with identical categories. For more information, see WEIGHTED KAPPA.

Charting enhancements

Chart Builder has been updated to include the following features/enhancements.

Bubble charts

Bubble charts display categories in groups as non-hierarchical packed circles. The size of each circle (bubble) is proportional to its value. Bubble charts are useful for comparing relationships in data.

High resolution chart export options

When the None (Graphics only) option is selected as the document type in the Export Output dialog, the default file type is now set to Production Ready Postscript (*.eps), which is a high resolution image format.

When the None (Graphics only) option is selected as the document type in the Export Output dialog,you can now select the Scalable Vector Graphics (*.svg) format, which is a high resolution image format.

Chart templates

• The Edit > Options > Charts dialog now includes a Samples Settings section that provides preview settings for any selected chart template. The dialog dynamically updates the preview chart images based on the specified settings.
• Chart Builder's Chart Appearance tab now provides options for selecting chart templates. You can choose to use settings that are defined in Edit > Options > Charts, select a chart template that is installed with IBM SPSS Statistics, or select a chart template from another location. For more information, see Chart Appearance Settings.

Default chart colors

The default chart colors have been changed to a blue theme.

Chart Builder > Chart Appearance tab

The tab now allows you to directly select different chart template files.

Chart Editor

You can now increase/decrease font sizes directly in the editor.

Legends and titles

You can now move chart images and titles directly in the output.

SPSS Web Reports and Cognos Active Reports

Support for both SPSS Web Reports and Cognos Active Reports have been deprecated.

Font size selection

You can now manually change the font size in the following locations:

• Edit > Options... > Viewer
• File > Page Attributes... > Font
• Pivot Table Editor (via the Formatting Toolbar)

The font Size lists provide a set of predefined sizes, but you can manually enter other, supported size values.

Search enhancements

The Search feature has been updated to provide results that include:

• Menu dialogs
• Help topics
• Case studies
• Syntax reference

Clicking a search result will take you directly to the relevant procedure dialog, help topic, case study, or syntax reference topic.

November 2019 update

Analyze procedures

ROC Analysis

The CLASSIFIER keyword was added to the PRINT subcommand. The keyword controls the display of the Classifier Evaluation Metrics table in the output. The table shows how well a classification model fits the data compared to a random assignment. For more information, see ROC Analysis: Display.

Performance enhancements

• Memory consumption has been improved when performing transformations.
• Application start time is now improved on Microsoft Windows machines.
• Improved support for importing Cognos BI data into the application.
• Support for the Microsoft Access database with the Office 2016 drivers.

June 2019 update

User interface

Welcome screen

The Welcome screen layout was enhanced and URLs were updated.

Data submission

Added opt-in feature to allow users to help improve SPSS Statistics with usage reports.

Give feedback prompt

The 'Give Feedback' prompt was optimized to not trigger as often.

Startup screen

The startup splash screen includes improved hover text and displays consistent product names.

About dialog

The About dialog now identifies, and easily allows copying of, the product version.

Licensing

Validation and response time

License validation performance and response time have been improved.

Bug fixes

Fixed Custom Dialog installation issue caused when help file is not installed properly.

Fixed Fleiss' Kappa confidence interval computation issue.

Fixed save syntax document issue caused by highlighted syntax in the syntax editor.

Fixed an issue where the machine temporarily stops responding during application startup.

Fixed a system resource consumption issue during application startup.

Fixed Quantile Regression Estimated Parameters issue caused when plots are too wide to display.

Various fixes to improve security and reliability.

April 2019 update

Analyze procedures

Quantile Regression

Models the relationship between a set of predictor (independent) variables and specific percentiles (or "quantiles") of a target (dependent) variable, most often the median.

Quantile regression makes no assumptions about the distribution of the target variable, tends to resist the influence of outlying observations, and is widely used for researching in industries such as ecology, healthcare, and financial economics.

ROC Analysis

Assesses the accuracy of model predictions by plotting sensitivity versus (1-specificity) of a classification test (as the threshold varies over an entire range of diagnostic test results). ROC Analysis supports the inference regarding a single AUC, precision-recall (PR) curves, and provides options for comparing two ROC curves that are generated from either independent groups or paired subjects.

Bayesian Statistics

One-way Repeated Measures ANOVA

This new procedure measures one factor from the same subject at each distinct time point or condition, and allows subjects to be crossed within the levels. It is assumed that each subject has a single observation for each time point or condition (as such, the subject-treatment interaction is not accounted for).

One Sample Binomial enhancements

The procedure provides options for executing Bayesian one-sample inference on Binomial distribution. The parameter of interest is π, which denotes the probability of success in a fixed number of trials that may lead to either success or failure. Note that each trial is independent of each other, and the probability π remains the same in each trial. A binomial random variable can be seen as the sum of a fixed number of independent Bernoulli trials.

One Sample Poisson enhancements

The procedure provides options for executing Bayesian one-sample inference on Poisson distribution. Poisson distribution, a useful model for rare events, assumes that within small time intervals, the probability of an event to occur is proportional to the length of waiting time. A conjugate prior within the Gamma distribution family is used when drawing Bayesian statistical inference on Poisson distribution.

Reliability Analysis

The procedure had been updated to provide options for Fleiss' Multiple Rater Kappa statistics that assess the interrater agreement to determine the reliability among the various raters. A higher agreement provides more confidence in the ratings reflecting the true circumstance. The Fleiss' Multiple Rater Kappa options are available in the Reliability Analysis: Statistics dialog.

Command enhancements

GENLINMIXED command

• New Covariance Type structures ARH1 & CSH, Random Effects. The CSH and ARH1 options were added to the /RANDOM subcommand (keyword COVARIANCE_TYPE).
• New Covariance Type structures ARH1 & CSH, Repeated Effects. The CSH and ARH1 options were added to the /DATA_STRUCTURE subcommand (keyword COVARIANCE_TYPE).
• Kenward - Roger Degree of Freedom method. The KENWARD_ROGER option was added to the /BUILD_OPTIONS subcommand (keyword DF_METHOD).
• Kronecker Covariance types. The options UN_AR1, UN_CS, UN_UN were added to the /DATA_STRUCTURE subcommand (keyword COVARIANCE_TYPE).
• New KRONECKER_MEASURES keyword. The keyword is used for specifying a list of variables for the /DATA_STRUCTURE subcommand. The keyword should be used only when COVARIANCE_TYPE is one of three Kronecker types. The rules for KRONECKER_MEASURES are the same as for REPEATED_MEASURES. When both specifications are in effect, they may or may not have common fields, but cannot be exactly the same (regardless of whether they are in the same order).

MIXED command

• DFMETHOD keyword introduced on the CRITERIA subcommand.
• KRONECKER keyword added to the REPEATED subcommand. The keyword should be used only when COVTYPE is one of three following Kronecker types.
• UN_AR1, UN_CS, and UN_UN options added to the COVTYPE keyword on the REPEATED subcommand.

December 2018 update

New "Give Feedback" option in the Help menu

Allows users to submit product feedback.

Updated SPSS support link and SPSS Statistics Subscription services terms

Deprecated support for the .NET plugin

The following issues have been addressed

• Fixed Set Customer Table Look as default and other Table Look issues.
• Fixed a MIXED procedure issue that caused it to hang with certain excel data types.
• Fixed a Statistics engine crash issue when executing Explore Analysis on mm:ss timestamp variable.
• Fixed an error when adding values to a password protected file.
• Fixed a Pivot Table issue that caused incorrect rounded decimal digital values to be displayed.
• Fixed an issue where Hiragana characters could not be entered in the Syntax Editor.
• Fixed the silent installation error when using the -log flag.
• Fixed Python application bundle startup failure caused by spaces in the installation path.
• Fixed various issues to improve security and reliability.

November 2017 update

MATRIX-END MATRIX command

SPSS Statistics now supports to following MATRIX command enhancements:

• Long variable names (up to 64 bytes) can be used to name a matrix or vector name (such as COMPUTE, CALL, PRINT, READ, WRITE, GET, SAVE, MGET, MSAVE, DISPLAY, RELEASE, and so on).
• Variable names that are included in a vector or matrix object are truncated to 8 bytes. This is because the matrix/vector structure is an array of numbers, and each number can match a string only up to 8 bytes. Long names (up to 64 bytes) are supported only when explicitly specified.
• Long variable names are supported in GET and SAVE commands when explicitly specified on the /VARIABLES subcommand (and when specified on the /STRINGS subcommand for the SAVE command). Variable names for GET and SAVE commands are truncated to 8 bytes when they are referenced through a vector in the /NAMES subcommand.
• The GET, SAVE, MGET, or MSAVE statements support both dataset references and physical file specifications.
• MATRIX-END MATRIX now supports statistical functions that were previously only supported by the COMPUTE command (for example IDF.CHISQ, CDF.NORMAL, NCDF.F, and so on).

August 2017 update

Bayesian statistics

SPSS Statistics now supports Bayesian statistics. Bayesian inference is a method of statistical inference in which Baye's theorem is used to update the probability for a hypothesis as more information becomes available. The following Bayesian statistics are supported:

• One Sample and Pair Sample t-tests
• One Sample Binomial Proportion tests
• One Sample Poisson Distribution Analysis
• Related Samples
• Independent Samples t-tests
• Pairwise Correlation (Pearson)
• Linear Regression
• One-way ANOVA
• Loglinear Regression

Output Viewer "Copy as" enhancement

You can now right-click a selected object in the Output Viewer and select Edit > Copy as to copy to the most popular formats (for example, All, Image, or Microsoft Office Graphic Object). Selecting Edit > Copy copies All.

Licensing

The SPSS Statistics licensing process has been replaced by your IBM account, also known as IBMid. An IBMid provides access to all IBM's applications (to which you are licensed), communities, and support channels. For more information, see Logging on and downloading updates.

When you first open IBM SPSS Statistics Subscription, you're prompted to log on with your IBMid. If you don't yet have an IBMid, follow the on-screen instructions.

The SPSS Statistics licensing options have been simplified. Previous versions provided 14 separate licensing options; those options have been consolidated down to 4 options.