JOSS review: author list

[kevinrue]

To date, the statistics of repository contributors report that @ibecav has made 200 commits to the repository adding 9,644 lines and removing 6,479 lines.

See:

• https://github.com/IndrajeetPatil/ggstatsplot/graphs/contributors
• https://github.com/IndrajeetPatil/ggstatsplot/commits?author=ibecav

This seems worthy of adding the contributor at least as a package contributor in the DESCRIPTION file (ctb, see https://cran.r-project.org/doc/manuals/R-exts.html#The-DESCRIPTION-file), unless they deserve authorship (aut), in which case that should also be reflected as authorship in the published article.

For openjournals/joss-reviews#3167

[IndrajeetPatil]

Yes, it is true that @ibecav contributed during the initial phase to ggstatsplot's development, but things have changed dramatically since then. He was indeed listed as a co-author initially but was later removed (and is not be listed as a co-author on this paper and package) for the following reasons:

1. ggstatsplot-universe

Although ggstatsplot used to be a single package when he contributed (March-June 2019), it was then split into multiple smaller R packages (ipmisc, pairwiseComparisons, statsExpressions, broomExtra, and tidyBF) for development and maintenance convenience. As a result, many of his contributions were dispersed across these packages, and were further outright removed/replaced during major code refactoring (see Point 3 below) that took place since he stopped contributing (June 2019).

2. Size of contributions

You mention that he has added 9,644 lines and removing 6,479 lines.

Across the 6 packages (ggstatsplot, ipmisc, pairwiseComparisons, statsExpressions, broomExtra, and tidyBF) that supported or supports ggstatsplot package, the tally of total contributions (mostly from me) to GitHub repositories have been following:
(on 25.05.2021 at 14:48 Berlin time)

Package	Additions	Deletions
ggstatsplot	831,008 ++	721,826 --
statsExpressions	152,527 ++	136,548 --
pairwiseComparisons	48,815 ++	42,133 --
ipmisc	28,642 ++	24,920 --
tidyBF	29,874 ++	23,316 --
broomExtra	62,476 ++	55,195 --
Total	1,153,342 ++	1,003,938 --

So, currently, his contributions count for 0.84% (9,644 / 1,101,935) and 0.65% (6,479 / 926,217) of all codebase that supported or supports ggstatsplot package. This is not even counting that this is an ongoing project and these contributions would continue to dilute further.

Therefore, I do not consider these contributions significant enough for co-authorship on the paper.

3. Refactoring

Since he stopped contributing in June 2019, the entire codebase of ggstatsplot, which was over 12,000 lines of code at the time was slowly, painfully, and painstakingly refactored to the current codebase of only about 2,000 lines of code. This was a major refactoring that removed/replaced most of his contributions (e.g., a function to compute confidence intervals for Cramer's V, using ez and MBESS for ANOVA, etc.) with other packages/ functions (effectsize::cramers_v, afex::ez_aov and effectsize::eta_squared, etc.). Some other functionality he had added to this package (ggbarstats) is now also re-implemented in functions from his own personal package (https://cran.r-project.org/web/packages/CGPfunctions/vignettes/Using-PlotXTabs2.html).

[kevinrue]

Thanks for the thorough historical view of the package evolution.

I thought as well that authorship was probably too much. I mostly wondered leaned toward the status of contributor (ctb), which would not affect authorship in the paper itself.

Having said that, I agree that the overall proportion of contribution, and the fact that he is ceased to contribute to the software development effort a while ago, is enough justification for me.

Thanks again for taking the time to explain.