Data Science

What's New in SPSS Statistics 25 & Subscription

By Douglas Stauber posted Tue July 18, 2017 04:00 AM

  
[caption id="attachment_6901" align="alignnone" width="2850"]SPSS Statistics 25 Screenshot Bayesian Independent Sample Inference test found in SPSS Statistics 25.[/caption]

Today I am proud to announce the latest features coming for SPSS Statistics 25 and as a free update to all SPSS Statistics Subscribers. These features bring much desired capabilities in charting, new statistical tests, and enhancements to existing statistics which all originated from customer feedback. In fact, to select and prioritize these features I've met, spoken to, or gotten feedback from over 217 customers across 99 companies, 37 universities, and 22 countries. This feedback has led to a release built for 99% of our users and shows our commitment to our customers and the larger statistical and analytics community.


Here’s what you can do with the new features in version 25 and the next subscription update:


  • Execute new Bayesian statistics functions including regression, ANOVA, and t-tests.
    Bayesian statistics is becoming very popular, because it circumvents a lot of the misunderstandings brought by standard statistics. Instead of using a p-value to reject or fail to reject a null hypothesis, Bayesian places an uncertainty on parameters and captures all relevant information from observed data. Our approach to Bayesian statistics is unique because our Bayesian procedures are as easy to run as our standard statistical tests. In just a few clicks you can run Linear Regression, ANOVA, One-Sample, Pair-Sample, Independent-Sample T-tests, Binomial Proportion Inference, Poisson Distribution Analysis, Pairwise Pearson Correlation, and Loglinear models to test the independence of two categorical variables. Bayesian Statistics are found in the Advanced Statistics module for Stats 25 and to Subscribers of the Custom Tables & Advanced Statistics add-on.



  • Quickly create attractive, modern charts and edit them in Microsoft Office.
    Building modern, attractive, and detailed charts has never been easier. Our chart builder has been updated with the ability to create publication quality charts in just a few clicks. Now you can specify chart colors, titles, and templates as you're building the chart. And, our new default template ensures a great looking chart even without modifications. In addition, now if you wish, you can copy most charts as a Microsoft Graphic Object so you can edit titles, colors, styling, and even chart type right in Microsoft Word, PowerPoint, or Excel. Charting in SPSS has never been this easy. All these charting features are found in the Base editions.



  • Extend your advanced statistical analysis with updates to MIXED, GENLINMIXED, GLM, and UNIANOVA.
    We've responded to customer feedback by adding the most requested enhancements to a few of our most popular advanced statistics functions. The Mixed Linear Models (MIXED) & Generalized Linear Mixed Models (GENLINMIXED) procedures now provide random effects solution results (EBLUPs) and continuous time spatial covariance structures. Also, the General Linear Model (GLM) and UNIANOVA procedures have been enhanced with new features including profile plots with error bars, bar/line charts, an option to include grand mean, and an option to force the chart to include 0 on Y axis. GLM and UNIANOVA also now include new tests for heteroskedasticity and model specification, robust standard errors, and modified versions of Levene’s test. Most of these functions are found in the Advanced Statistics module for Stats 25 and to Subscribers of the Custom Tables & Advanced Statistics add-on, although UNIANOVA is found in the Base editions.


  • Write, edit, and format syntax faster with Syntax Editor shortcuts.
    Here are a set of features for those users that are primarily syntax writers. We’ve added a bunch of features (and associated keyboard shortcuts) to make writing, formatting, and editing syntax easier. Now with a simple keyboard shortcut you can join lines, duplicate lines, delete lines, remove empty lines, move lines up or down, and trim leading or trailing spaces. In addition, we’re introducing a new column editing mode which allows you to edit multiple lines at once. A killer feature here is the ability to copy data from the data editor (or Microsoft Excel) and paste “down” across multiple lines. It’s a great feature – once you try it, you will no longer be able to go back to the earlier syntax editor. These syntax features are found in the Base editions.


  • Plus more.
    We have a new merge UI that simplifies merging options, a simplified toolbar with some convenient buttons added, the ability to open and save Stata v14 files, as well as the ability to copy and paste data from the data editor with variable names or labels.








  • These features are now available in SPSS Statistics 25 and Subscription. In addition, the SPSS Statistics Subscription trial has been updated. Check out the IBM Marketplace for trial and purchase information.

    I highly recommend becoming a subscriber of SPSS Statistics so you can get these and all future updates with easy downloads, easy user and device management, and simple registration with IBMID instead of license keys. Consider becoming a Subscriber today!

    Big thank you to all customers, sponsored users, beta participants, designers, developers, statisticians, and business partners that have contributed their time to bringing these features to life.


    EDIT: Watch a webinar to see Chart Building and Bayesian features in action View now!

#25
#update
#syntax
#BusinessPartner
#statistics
#datavisualization
#Release
#chart
#SPSSStatistics
#major
#Visualization
#StatisticsSubscription
#Subscription
#datascience
#Regression
#spss
#bayesian
52 comments
7 views

Comments

Fri February 01, 2019 03:58 PM

hi Arun, you will need to reach out to IBM via this form for an official quote - https://www.ibm.com/account/reg/us-en/signup?formid=MAIL-analyticsstats

Fri February 01, 2019 03:57 PM

hi Arun - you will need to use this form to get an official quote from IBM https://www.ibm.com/account/reg/us-en/signup?formid=MAIL-analyticsstats

Sun January 27, 2019 05:24 AM

hi.
thanks for your grate effort.
unfortunately this new UI isn't accessible. for screen readers please consider accessible UI for latest updates.
thanks in advance.

Sat January 26, 2019 09:49 AM

I need Softwares like:
Specification:

1. SPSS Statistics V25 (latest) perpetual License Base + 14 modules (full pack) –Network version with one year AMC – 20 users with 2days onsite training – SOTP with Deployment Tool Analuo (20 users)

2. Mathematica Network License (latest version) with 1 years PS 10 users with installation and 2 days on site training.

Pls send quotation for this software.
Can you provide ?

Tue December 18, 2018 10:37 AM

Hello I am currently writing my master thesis. I am going to conduct a choice based conjoint analysis with a hierarchical bayes analysis. Is is the hierarchical bayes method possible with ssps?

Wed September 26, 2018 06:34 PM

Is there finally an Auto-Save feature in 25? I have lost many files because SPSS suddenly crashed and there was no backup.

Thu July 26, 2018 09:31 PM

I am on the V25 of Premium (standalone, one-year rental, academic). I need to review my annual license in a couple of weeks. This is about the time of year SPSS upgrades to the next version (ie V26). However I also know SPSS is is looking at a different UI via their subscription. So, was SPSS 25 the last year of an annual standalone update... or will something be coming out soon? Again, don't want to jump the gun if something new is coming out. Thanks to anyone that has insight.

Sun July 08, 2018 07:13 AM

Can you use a newer version of SPSS ie 25 to read files written in SPSS 22?

Wed June 20, 2018 06:30 AM

what is the exact utility of 'visual builder' ? the sample power which is included in the premium package, I was told the support for the same is discontinued so what is the use of it?

Mon May 14, 2018 08:21 PM

Hi - I'll have the team research - thanks.

Sat May 12, 2018 08:00 AM

Dear Doug
Hi
is it possible in the next release IBMSPSS can do DEA(Data envelopment analysis)
Thanks

Fri April 13, 2018 01:40 PM

I have contacted your Technical team for Exact logistic regression a few years ago, still its no in the package. I would love to see that.

Mon April 02, 2018 02:38 PM

Hi - We've added this feature as part of our new user interface.

Mon April 02, 2018 08:30 AM

In sass 24 I found annoying the lack of possibility of ordering the files according to date, type etcetera .
Can you add this feature. This will help retrieving whatever one is interested in.
Thanks

Wed January 03, 2018 09:31 PM

Once you create a chart in SPSS Statistics, you simply right click on the chart and "Copy As->Microsoft Office Graphic Object". Then you paste in Word. Note this works on most charts, but not 100% as SPSS supports more chart types than Office.

Wed January 03, 2018 05:41 PM

How do I go about editing a graph in word? which file extension should I use?

Fri December 01, 2017 04:37 PM

Hi Chris, it's something I'll have the team review for a future release.

Fri December 01, 2017 12:21 AM

Will Statistics be able to in the future? Does SPSS 22 interface or will I always need modeler?

Thu November 30, 2017 05:53 PM

Hi Chris - You cannot with SPSS Statistics 25, but SPSS Modeler 18.1 started supporting Cognos 11. So, not sure if you use Modeler alongside Statistics, but that is the best solution at this time.

Wed November 29, 2017 01:58 PM

Doug - Can I create a direct connection into Cognos 11 from SPSS Statistics 25? I have a user with SPSS Statistics 22 who is having problems connecting to Cognos to export data into SPSS 22 and thought maybe I need to upgrade her? Thanks!

Fri November 17, 2017 11:05 PM

We do not have a Bayes hierarchical method today, but it is something I will ask the team to look into. Thanks for the suggestion.

Fri November 17, 2017 07:54 PM

I like the addition of Bayes statistics but I'm looking to analyze Max-Diff data -- what about hierarchical Bayes?

Wed November 01, 2017 06:43 PM

Thank you for the comment Pedro. We will look into this.

Wed November 01, 2017 06:24 PM

My students have the same problem (in both Mac or windows). After doing Analyze>General Linear Models>Univariate>Save and clicking the Unstandardized for residuals, an error is produced:

"A program error has occurred: A procedure has attempted to add more variables to the file than it provided for in its call to OBINIT. The error was detected in a call to OBPINI. Please note the circumstances under which this error occurred, attempting to replicate it if possible, and then notify Technical Support.
Execution of this command stops."

No new column for residuals in created in the data view window and in some cases, SPSS crashed.

Best regards

Mon October 23, 2017 02:53 PM

Thomas - Thank you very much for your feedback!

Mon October 23, 2017 02:43 PM

We also need SPSS to add p-values for eta coefficients under crosstabs. Also some point bisereal pb's would be nice to have. You can always get to them with ANOVA but its harder... R does them on the same run.

Fri October 20, 2017 07:19 PM

Thank you for your question! I have just run the procedure on V25 and it has generated the new columns and added to the data view. I encourage you to reach out to support on our new support page: https://www.ibm.com/analytics/us/en/spss/spss-support/

Fri October 20, 2017 06:43 PM

Doug,

For version 25, when running a 2way ANOVA I am unable to get the Predicted Values/Residuals. Is there a work around for this?

Currently, clicking Analyze>General Linear Models>Univariate>Save and clicking the Unstandardized for predicted values, as well as Unstandardized and Studentized for residuals, it does not place the 3 new columns in the data view, as previous versions would. This data is needed when spiting files to look at Normality and Outliers, which I cannot do from version 25.

Any information you can provide would be helpful

Thu October 12, 2017 11:42 AM

Raking in SPSS is done with an extension command. On the menu, go to Extensions->Extension Hub. Search for "Rake" then select Get Extension. It will install on the Data menu.

Wed October 11, 2017 10:52 PM

I downloaded the trial and don't see raking

Tue October 10, 2017 02:23 PM

Hi Doug

I'm trying to create a bar chart (using % less than function) and no bars are visible on the output! I can't figure out how to make them visible.

Rich

Wed September 13, 2017 03:46 PM

David- See the link at the bottom of the blog. We have demos of the new Chartbuilder and Bayesian.

Fri September 01, 2017 08:01 AM

I would like to have a improvements. Like Lack-Of-Fit test to linear and quadratic regression. Also I would like to have Mandel fitting test.

Mon August 28, 2017 07:31 PM

Hi Jennie - I consider this a bug. Thanks for bringing this to our attention. I've asked the team to review and get this fixed in the next fixpack.

Mon August 28, 2017 01:34 PM

Hi Doug!
We often use stacked bar (in Chart builder) but when we want to transpose stacked bar in "Basic Elements" to a horizontal bar, the labels (or colours) in the bars go in a different direction than the labels (colours) in the legend. How can I solve this problem? There is a substantial demand of using this cind of horizontal bars.

kind regards from Jennie

Tue August 22, 2017 06:40 PM

Hi Matthew - Today I asked the team review this request. Thanks for the suggestion.

Tue August 22, 2017 03:36 PM

Are there any plans to allow the split file function (more specifically multiple imputation) to work with Generalized Linear Mixed Models?

Mon August 21, 2017 02:28 PM

We are listening! Thanks for the suggestions.

Sat August 19, 2017 01:02 PM

The new syntax features seem to meet some of the improvement suggestions I made in the SPSSX Discussion Group (http://spssx-discussion.1045642.n5.nabble.com/Making-SPSS-better-td5732642.html). Looking forward to version 25 and hope we get it soon. Thanks!

Tue August 15, 2017 01:00 AM

Yes, I will record a video of the most popular procedures and update this page with the links.

Mon August 14, 2017 09:40 PM

Do you plan to have a youtube demo uploaded so we can share the link?

Mon August 14, 2017 03:24 PM

Great feedback, thanks.

Sun August 13, 2017 08:37 PM

Hi Doug,

The choice of activation function in the multilayer perceptron model does not include rectified linear unit, and the maximum number of hidden layers is limited to 2. I hope a new version of SPSS is more flexible to design neural network analysis. Also, I will be surprised if SPSS can handle convolutional neural network for image recognition in the future.

Thanks,

Koji

Thu August 10, 2017 11:33 PM

Hi Koji - What kind of changes are you interested in?

Thu August 10, 2017 10:34 PM

Any change or progress in neural network analysis?

Sun August 06, 2017 08:57 PM

Great! We need it!

Fri August 04, 2017 05:57 PM

I hope It won't take another 25 years. :) Regards

Fri August 04, 2017 05:25 PM

Hi Juan - I will examine this for a future release. Thanks for bringing my attention to the need.

Fri August 04, 2017 12:44 PM

update:
Any possibility of including NONLINEAR mixed- effect regression modeling? This question was filed to the development team when version 19 came out.
Following is the link where this question was raised.
http://www-01.ibm.com/support/docview.wss?uid=swg21474925

Stata included this capability in the newly version 15. Also in Minitab V18.1
Thank you for your feedback

Fri August 04, 2017 12:30 PM

Any possibility of including NONLINEAR mixed- effect regression modeling? This question was filed to the development team when version 19 came out. Thx.

Wed July 26, 2017 05:31 PM

Hi Carolyn - There are no enhancements in that particular product area, but I am sensitive to the issue. In fact, I'm considering making a whole bunch of updates to our regression procedures. I plan to have the community suggest and vote on regression enhancements. Keep visiting this community for more info on that.

Wed July 26, 2017 04:13 PM

What has IBM done about running conditional logistic regression models (not the work around using Cox Regression)?