R package for effect size calculations for psychology researchers

Dan Gerlanc writes:

I read your post the other day [now the other month, as our blog is on a bit of a delay] on helping psychologists do research and thought you might be interested in our R package, “bootES”, for robust effect size calculation and confidence interval estimation using resampling techniques. The package provides one function, ‘bootES’, that makes a variety of effect size calculations fairly straightforward for researchers with limited programming experience. The majority of the implemented are not available in R or SPSS without custom coding. Kris Kirby (Williams College) and I have published a paper in Behavioral Research Methods describing the methods and providing a tutorial on use of the package: http://bit.ly/YIM6VD. We hope that it’s useful to psychologists and other social science researchers!

I haven’t tried this out but it might be of interest for some of you.

10 thoughts on “R package for effect size calculations for psychology researchers

  1. bootES: Documentation has not a simple example, and it cites a paper that is not available. Why is the link to the tutorial not in the docs?

    I know that R documentation has strong limitations, but that’s not excuse not to use it within its limits.

  2. I just noted that the cited tutorial is behind a paywall, as gwern has already mentioned. bootES will be removed from my disk, for sure.

  3. It is not a paywall, it is a standard journal article. Not everyone has the resources to publish in open access journals. Yes, the cran reference manual is lacking but it takes guts for a psychologist (who probably it is not mainly quantitative) to get a package to cran.

      • I can see why they avoided JSS as psychologists don’t tend to read it (with a few exceptions, myself included) so BRM is a reasonable choice but a pre-print would be a good idea.

        As a psychologist I’d also argue that requesting the authors to post a pre-print is a better strategy than lambasting them for not having done so (at least if the goal is to promote open access).

    • That’s like saying for the NYT or WSJ, “It’s not a paywall, it is a standard newspaper”

      Yes, it’s a paywall. If you have to pay to read content it is a paywall. Notwithstanding that may be the standard model in academic publishing.

      Who’s saying you must publish open access? But post a pdf online and link to that! If your journal objects fight them. Or at least post a draft on arxiv and link to that! This isn’t novel, uncharted territory.

      It’s sort of inane to write a open-access package, then publicize it, expect people to show interest, and yet keep the primary documentation source behind a subscribers-only paywall!

    • Yes, the cran reference manual is lacking but it takes guts for a psychologist (who probably it is not mainly quantitative) to get a package to cran.

      Are you jokingm re psychology? Never heard of Spearman or Cattel? Perhaps even Galton? Never had a look at the Psycholgical Task View at CRAN(admitidly limited but revealing). Or as Andrew Gelman (paraphase) puts it ” Any new statistical area has already been done by psychologists 50 years ago”

      It is just sloppy work and resumably a bad link by mistake.

  4. ——————————-
    Thom wrote: “As a psychologist I’d also argue that requesting the authors to post a pre-print is a better strategy than lambasting them for not having done so (at least if the goal is to promote open access).”
    ——————————-
    Couldn’t agree more Thom. I think most of the posters are right about making the article accessible, but what’s the problem in asking politely first? This isn’t Facebook! If you’re not happy with the response you receive then go for it, but until then…

    And thanks Daniel for posting the new link.

Comments are closed.