IBM SPSS AMOS 26 was released on April 9th, 2019. The two new features in version 26 are:

1. Automatically displayed output on the path diagram.

2. The ability to specify a model without drawing a path diagram via syntax.

IBM SPSS AMOS is found in the SPSS Statistics Premium commercial bundle, and is also available to purchase separately. Additional information can be found on the IBM Marketplace.

The release notes can be found here.

21 comments on"What’s New in SPSS AMOS 26"

  1. C. Arnold May 23, 2019

    Is that all? The list of missing features in relation to other SEM packages is getting longer from year to year.

    • Hi C. Arnold,

      Yes, those are the two features that were added to SPSS AMOS. We appreciate customer feedback so please let us know which areas you would like to see improved and key features you believe should be added.


      • C. Arnold June 02, 2019

        Hi Brent, Thanks for your reply. Here is a short list from my point of view, sorted by importance, of course also from my point of view.
        – ML with scaled se and test statistics (not only bootstrap for se)
        – Invariance tests based on scaled test statistics (Satorra/Bentler 2010)
        – Latent Moderated Structural Equations (Moosbrugger et al. 1997)
        – Estimator for categorial data
        – Multilevel SEM
        – NET procedure (Bentler/Satorra 2010)
        – […]

        • Hi C. Arnold,

          We greatly appreciate you taking the time to provide feature enhancement requests. We will bring these back to the team to evaluate further to improve our future products. If you have additional feedback regarding SPSS Statistics or SPSS AMOS, please let us know.

          SPSS Team

  2. Hi,
    I just wonder if AMOS 26 would work on SPSS 25 compatibly?

    • Hello,

      SPSS AMOS can be purchased separately from SPSS Statistics. Therefore, you could have AMOS 26 and SPSS Statistics 25.


  3. Hi, the lack of basics of SEM, like geting MIs even with missing data, categorical estimators and scaled, robust SEs and S_B Chi square for non- normal all data. are all available in competetors’ softwares. It seriously limits the usability of AMOS and has not been added for so long, that has been the reason why many have moved away from AMOS to other softwares.

  4. Peter Lugtig June 04, 2019

    This getting ridiculous. Both these features have actually been in AMOS for a long time (well admittedly, there is now one less button to press)..
    As a long time AMOS user, I would second C. Arnold’s opinion in that AMOS needs to start some serious and necessary updates. I used to love AMOS, especially in teaching, but am suffering more and more with every AMOS update.

    • Hi Peter,

      Thank you for your feedback. Which key features do you feel are missing from SPSS AMOS?

      SPSS Team

      • Peter Lugtig June 05, 2019

        1. categorical observed variables (ordinal and nominal)
        2. Categorical latent variables (nominal)
        3. Proper mediation module
        4. scaled s.e.
        5. multilevel module

        A basic strategy I think would simply look at what Mplus has done, and build the “popular” stuff from Mplus into Amos.

    • C. Arnold June 04, 2019

      Peter, thank you for supporting my view. I just don’t understand how a global company like IBM can market these “new” features and ignore the minimum requirements for “modern” sem. To be honest, in fact you can’t publish seriously with Amos anymore. Scaled se and test statistics are an absolute minimal feature and even this has been ignored for years. Unfortunately. I still hope that Amos will get a serious update very soon.

  5. Nathalie Zug June 12, 2019

    AMOS was taught at our university last year. I have to admit that using AMOS helped for a better understanding with Rstudio. But for that price it’s ridiculous that I can’t even use AMOS to evaluate my bachelor thesis.

    • Hi Nathalie,

      Student level pricing exists through the SPSS Statistics GradPack for AMOS. Have you already checked the pricing for the student AMOS version?

      SPSS Statistics Team

    • Ibrahim Nasser June 12, 2019

      I have to agree. I evaluated Amos for my thesis and found that I have a huge problem with missing and non-normal distributed data. First, I don’t get corrected test statistics. Second, I can’t even bootstrap with missing data. I don’t get modification indices either. I also switched to R.

      Best, Ibrahim

  6. tawanda Cleopas vera July 27, 2019


    I was wondering if the latest version of SPSS AMOS can be used to run a Dynamic SEM with financial time series. I was thinking of using it for investment portfolio optimisation

  7. It will be awesome if you can include the Satorra & Bentler correction for non-normal data.

  8. I have to say that I came to this forum excited about the possibility of seeing what changes have been made to AMOS in the newest version, and am sorely disappointed. I’ve been teaching with AMOS for several years. I also contribute video demonstrations to Youtube on the software and how folks can use its features. But I’ll be honest, the limitations (noted by other folks above) have led me to seriously reconsider teaching with it in the future, let alone using it or encouraging its use. Currently, the program is good for helping students gain an introduction to SEM, but it’s not useful for teaching students about the “state of the art” in SEM or for addressing more recent developments in “best practices”. It’s as if the program is still stuck in the mid 2000’s, while programs like MPLUS, LISREL, and Lavaan are light years ahead. As I see it, there is little point in purchasing AMOS as an add-on if the options do not fit with best practices or publishing norms in 2019. My recommendation is that the next iteration of AMOS seriously incorporate most of the recommendations suggested above [by the way, these are not “special features”; rather they are ground-floor necessities for effectively modeling data using SEM]. Otherwise I suspect fewer and fewer people will bother with the program. Sorry for the long post. But I am pretty disappointed on this.

  9. And how about modeling interactions in AMOS? Softwares like ADANCO and SMARt-PLS are light years ahead. AMOS must learn to reinvent itself and keep abreast of the competition. I do a lot of SEM training programs and am thinking of switching over to SMART-PLS from AMOS.

    • Modeling interactions is not so terribly difficult with Amos, assuming that assumptions hold. This has little in common with a completely different logic (PLS). Somewhere the demands on IBM have to be fair, don’t they? Only my two cents.

  10. Dear Amos team developer,

    Researchers are really puzzling with outdated software. From version 21 no major changes appeared. I agree with Arnold comments above. The team must implement these suggestions if you don’t want costumers to switch to other apps.

    Most importantly consider these:
    1. categorical observed variables (ordinal and nominal). Add more estimator techniques such as diagonally weighted least square, maximum likelihood estimation with robust standard errors.
    This is important because many researcher will refrain from using Amos if such estimators are unavailable.

    2. Make the default diagram output white and black please. The current fancy colours do not help in APA publication.

    3. Allow theta estimation. For example see this analytical approach which cannot be conducted in Amos, Exploratory Structural Equation Modeling

Join The Discussion

Your email address will not be published. Required fields are marked *