[metalang]

By far my biggest pain point is Microsoft Word. I've found that there are existing solutions for most steps in the research process (Git for version control, Rmarkdown/Quarto/Jupyter/Orgmode for integrating the code and the analyses, Zotero for citations, etc), but almost every research project that I've been a part of has eventually required the manuscript to end up in Word due to either journal requirements or some collaborators refusing to use other tools. At that point all the advantages of the other tools go out the window. It's easy for people to change tables or otherwise break the linkage between the manuscript and the analysis code, and you're stuck spending a ton of time auditing the paper and fiddling with the appearance of tables. Furthermore, once you're in Word you have to manually move any changes in the analysis into the manuscript rather than being able to rerun the code and run the output through Pandoc. I would love a tool that lets me work with collaborators in Word (including track changes) without losing the advantages of my preferred workflow.

[maxerickson]

I haven't used it enough to have an opinion about using it, but you can build a word template that you hand to pandoc when it is converting to word, and it will stuff all your content into the matching styles in the template. The table templates in word are annoying to use, but you can put nearly all of the formatting into the style and not have to mess with individual tables.

Mostly just throwing stuff out there in case you hadn't looked at it.

[metalang]

Thanks! I actually have not spent much time looking into that and will see if it helps.

[kzardar]

First of all, thank you for sharing! Much appreciated for the added insight.

How often do you tend to re-run analysis code and end up fiddling with tables in Word?

[metalang]

I have had to do it multiple times for every paper I’ve published. I do health economics research, which involves a lot of collaboration with physicians and people from other less quantitative fields (eg public health, implementation science, communications / media). From my experience those collaborators are unwilling to look at results until they look “finished” (ie are in the format that the target journal requires) and are part of a Word document that includes the working draft of the paper. Since they want to use Track Changes as version control, I can’t simply generate the tables again as a Word document and end up having to copy and paste them in, which is a huge time suck and runs the risk of making errors.

[kzardar]

What are your thoughts on using Google Docs to prototype this? Correct me if I'm wrong in my logic. You're looking for a way to link end results of your analysis to a 'Word' document that will auto-populate with each re-run. The physicians etc. want a 'clean' manuscript that mimics graphs in the format of the preferred paper + has version control.