Request for Technical Support: Issues with Meta-Analysis Module in JAMOVI Version 2.6.26

[arshishad]

Dear JAMOVI Development Team,

Greetings from Team MERIT INDIA.

We are a dedicated academic team engaged in statistical training and research capacity building across India primarily with JAMOVI Software. Our lead resource person is Dr. Pragya Kumar, Professor at AIIMS Patna. Over the past five years, we have conducted more than 100 online and offline workshops and webinars on JAMOVI, training over 18,000 participants from diverse academic and professional backgrounds. All our training materials and recorded sessions are made publicly accessible through our YouTube channel to support open learning.

Recently, we expanded our training portfolio to include Meta-Analysis using JAMOVI, which has been well-received by our academic community. While our first workshop (October 2024) on meta-analysis ran successfully using JAMOVI, we are currently facing technical challenges with the latest version 2.6.26 (Solid) during our second workshop.

Issue 1: Error in Dichotomous Meta-Analysis Model
Participants using JAMOVI v2.6.26 are encountering an error in the Meta-Analysis → Dichotomous Model module. The error message appears as:

“The condition has length > 1”

This issue occurs even when using valid datasets and has been replicated on both the desktop and cloud versions of JAMOVI. We suspect this is a bug introduced in the latest update, as the same dataset works correctly in the previous version 2.3.28, which we are now using as a temporary workaround.

Our next hands-on session is scheduled for Saturday, 29th March 2025, with over 80 registered participants, some of whom are already using the latest version.

Issue 2: Categorical Moderators in Meta-Analysis
While the drop-down menu in the Meta-Analysis module suggests support for categorical moderators, the current interface allows only the inclusion of continuous variables in the "Moderator" (covariate) field. We would greatly appreciate clarification on whether this feature is under development or if it can be activated through a different workflow.

We kindly request your support in resolving these two issues at the earliest, as we are also preparing for our third national Meta-Analysis workshop in October 2025, where we hope to demonstrate JAMOVI's capabilities without the need for workarounds.

[jonathon]

hi,

MAJOR is a community project, and not something that the jamovi project itself is responsible for. Further, the original author of the module hasn't been active in some time. Having said that, I think this is something we can take a look at fixing.

The error message you describe is not a bug in jamovi, but rather a change in R's behaviour. R, at the time when MAJOR was first written, permitted the use of vectors greater than length 1 in 'if statements'. They changed this behaviour in a more recent version of R, and when we updated the version of R we ship with jamovi, that problem surfaced.

Anyway, Maurizio is going to take a look at issue 1. Issue 2 I'm not sure about, let's wait and see what Maurizio says.

Kind regards

Jonathon

[arshishad]

Hi Jonathon,

Thank you for the clarification. We have informed the participants about this limitation. Fortunately, we had version 2.3.28 available and shared it with them, so they were able to complete the analysis.

Meta-analysis is quite popular among our participants, so this module is very important for us. We really hope Maurizio is able to find a solution.

We are also trying to understand the coding part, but it’s quite difficult for us as we come from a completely different background.

Thanks again for your support.

Shamshad

[MAgojam]

We really hope Maurizio is able to find a solution.

Dear Shamshad and Team MERIT INDIA,

Thank you for bringing these important issues with the Meta-Analysis MAJOR module to our attention. I have read your request and Jonathon's helpful clarification regarding the community nature of this module and the underlying changes in R.

Please be assured that I cannot ignore this call for help, especially given the significant work you are doing in India to promote statistical training and research using jamovi. Your efforts in conducting numerous workshops and making resources publicly available are truly commendable, and the popularity of meta-analysis among your participants underscores the importance of this module.

As Jonathon mentioned, I am aware that the MAJOR module has not been actively maintained by its original author for some time. I have personally worked on adjustments to this module in the past, and I am committed to addressing the immediate error you are encountering in the Dichotomous Meta-Analysis Model. I will investigate the "The condition has length > 1" error related to the R update and aim to provide a fix for this in the near future.

Regarding the second issue concerning categorical moderators, I understand this is an important feature for meta-analysis. While a quick resolution might not be immediately possible, I will certainly explore the underlying code and the capabilities of the metafor library (which currently sits at version 4.6-0 in the R snapshot used by the compiler, a significant update from the v3.9 likely used by the current MAJOR module).

Please understand that a more comprehensive review and enhancement of the MAJOR module to fully align with the latest metafor library and potentially introduce new features, such as proper support for categorical moderators, will require more time and effort. This will involve a deeper dive into the code and potentially significant modifications. However, ensuring the module remains functional and relevant within the jamovi ecosystem is a priority for me.

Thank you for your patience and understanding. I will keep the community updated on my progress in addressing these issues. In the meantime, I appreciate you sharing the workaround using the older version of jamovi with your workshop participants.

Best regards,

Maurizio
https://www.jamovi.org/about.html

[arshishad]

That's great to hear..
Atleast you can fix status quo...

Regarding Categorical moderator we currently find a way using filter option and selecting only desired studies at a time..Thanks to your team also...We can see the effect of jamovi...The way we shifted from spss to it wayback..then our entire institute and many more thereafter...with only one word from everyone's mouth "It's a MAGIC"..

Our core aim is to reduce the fear of data analysis in our region and give everyone the first feeling that YES I CAN DO ANALYSIS...with this confidence our major work get over...Rest they keep learning...


Soon we are coming with all experience to your team...To fix everything in best possible way..

Love you all.
Team MERIT INDIA