Skip to content

Speaking frankly

Even within the realm of writing-about-statistics, there are things I can say in a blog that are much more difficult to include in an academic article. Blogging gives me freedom.

But I want to distinguish between two different sorts of frankness.

1. Obnoxiousness: In a blog I can write, “I hate X” as rudely as I’d like without needing to justify myself.

2. Openness: In a blog I can write about the limitations of my work. It’s a real challenge to discuss limitations in a scholarly article, as we’re always looking over our shoulder at what referees might think. Sure, sometimes I can get away with writing “Survey weighting is a mess,” but my impression is that most scholarly articles are relentlessly upbeat. Sort of like how a magazine article typically will have a theme and just plug it over and over. In a blog we can more easily admit uncertainty.

Overall, I think blogs are more celebrated for feature 1 above (the freedom to say what you really feel, to be rude, partisan, and politically incorrect), but I think feature 2 (the freedom to express uncertainty) is important too.


  1. Pat Burns says:

    Another bit of frankness that is possible in blogs is the ability to write about things that don’t work. Knowing what doesn’t work is very valuable but it’s close to impossible to get that into a journal.

  2. K? O'Rourke says:

    Perhaps speaking about something – without having to do a lot of work to get to say it in a publication.

    Just noticed this paper – Hidden Dangers of Specifying Noninformative Priors. SEAMAN, SEAMAN Jr., and STAMEY. The American Statistician, May 2012, Vol. 66, No. 2

    I do think the following two points are very sensible but not emphasized enough.

    “(1) Always examine induced priors on functions of interest. Simulation is usually the easiest way to see such a prior.

    (2) Always compare the posterior to its corresponding prior, induced or otherwise. Ideally, this should be done graphically.”

    Now it may be related to the other work that the authors are involved in (and getting publications may well be important to them), but coming up with the eight pages and locating and reading all those references does seem like a lot of work to get those ideas out in that journal. I’ll try to remember to check the citation count in a couple years.