News for hackers. Stuff that matters.

jpatrick

> I don't understand why this article is so long.

It's a humor piece. "Shouts & Murmurs" is the New Yorker's humor section.

Re:Why I don't use ggplot2 on Thursday February 11, 2016 @09:50PM (#11084985)
I think intermediate-complexity graphics are where ggplot2 excels. If you want to do something like plot a bunch of small multiples, you'll have a much better time of it using facet_grid than wrangling with base graphics.

Once your graph reaches a certain level of complexity, though, or requires a certain degree of customization, I think base graphics regain the edge. At that point I prefer the level of control you get by drawing things from scratch with points(), lines(), text(), etc.

One of the co-authors of that Atlantic article is Richard Sander, who's the leading proponent of the mismatch hypothesis and routinely overstates the evidence for it. His original study is based on outcomes for law students who started law school in 1991 -- not necessarily applicable, I'd argue, to undergraduates starting college in 2015 -- and his methodology has been pretty much discredited anyway (see the amicus brief that doctorpangloss cites).

This is off-topic from the original post, but I think the whole mismatch controversy has been kind of a scary illustration of how one determined academic can shape the narrative around an issue: because he's a professor, and because media outlets trust professors (and like controversy), he can easily find a platform to represent the debate in a way that's favorable to his side (e.g. [0]). I don't think it's a crazy hypothesis or anything, and it's definitely worth debating, but I'd be skeptical of claims coming from someone who's staked so much of his career on its validity.

</rant>

[0] https://www.washingtonpost.com/news/volokh-conspiracy/wp/201...