writing a tutorial paper on intro to R^2 measures and their computation with performance package

[IndrajeetPatil]

No description provided.

[bwiernik]

For recent discussion #332 #334

[IndrajeetPatil]

Hi @mkshaw! 👋

Do you think you would be interested in helping with this?

Should be right down your alley 🙃

[mkshaw]

@IndrajeetPatil it does look like it's down my alley! I recently published this tutorial paper on R2s in MLMs using examples to show how to calculate and interpret output. It could be useful to have something similar for the various R2s available in performance: a top-level summary of all the measures available and generally how they differ/why we offer them, and then examples one-by-one showing their computation and interpretation.

What are your thoughts on this structure, and @bwiernik based on your discussions in #332 and #334 it seems like you're quite familiar with the measures already, would you be interested in collaborating on this?

[bwiernik]

Yes, that would be great! Do you think you could start by putting together an outline or a draft and starting a pull request?

[IndrajeetPatil]

@mkshaw Are you still interested in working on this? Thanks.

[mkshaw]

@IndrajeetPatil I am still interested. I have my dissertation proposal next week, then can flesh out the outline. What software do you recommend for collaborating on this paper? I usually just write in Word, but if I'm starting a pull request and it's larger-scale collaboration, I imagine something else might be prefereable.

[IndrajeetPatil]

@mkshaw Would you be comfortable collaborating on GitHub itself? You can have a look at this PR where we are collaborating to write a new paper.

As for software, we can use RMarkdown to generate a PDF (probably via rticles) we can use for journal submission.

[mkshaw]

@IndrajeetPatil Happy to do that, I just need some help getting started, haven't done that before. Looking at the first commit from that PR, did you write the outline in RMarkdown locally, export to pdf, then upload the Rmd, md, and pdf files manually? And then people edit the md from there?

[strengejacke]

I'd suggest using Word, since any revisions usually need to be submitted as "tracked changed" anyway, and "tracked changes" feature isn't really comfortable with markdown, even not for collaboration.

[mkshaw]

@strengejacke okay so you recommend pull the Word doc from GitHub, edit locally, push?

[strengejacke]

Yes, or even a Google docs file (or Word online). It depends on your workflow, pushing/pulling from GitHub is ok, too, I think.

[mkshaw]

Google Docs online works for me and is perhaps the easiest. @bwiernik we had talked about collaborating on this, do you have a preference?

[IndrajeetPatil]

This (and the links therein) can help you get started: https://easystats.github.io/performance/articles/r2.html

If we use a Word document, we could use Google Docs and collaborate there.

[bwiernik]

No preference

[mkshaw]

Does anyone know how to share a google doc file that has people request access to view? I don't want to create an anyone-can-edit link, but I don't have your email addresses to send specific access. It would be best to create a link you can click to request access, but that doesn't seem possible.

[mattansb]

If you share the url as-is (say, copied from the browser) it should have that effect.

[mkshaw]

Let me know if this works, barebones outline: https://docs.google.com/document/d/1AUWxoTyvRdDA-gjvnWNRlbh5DDZ6dYug/edit?rtpof=true