Social Science Statistics Blog

I think that the gold standard for posts on this site should both: (1) highlight an interesting statistical technique and (2) have application to an interest social science problem. Of course, many posts may have only one or the other and be relevant and interesting.

I was inspired to make these recommendations based on a post by Andrew Gelman called "A nearly generic referee report."

Personally, I also have a tolerance / interest in a limited number of posts (say less than 15%) of a more personal nature (e.g., what's going on at Harvard, musing about current research, etc.).

Amy,

Go ahead and blog on any academic topic you feel like. A “statistic” is a function of a sample, which should give you plenty of room to treat the topic as related to social science statistics. Posts that pass Michael Weiksner’s gold standard test would certainly seem to excite more interest and comment. But can any of us say that posts -- from essays to quips -- that fail won’t be as interesting. I expect that readers of these pages can quickly discern whether a post is going to be of interest to them. Maybe you will say something about cognitive science that will send me off looking at alternative time series structures with temperature dependent variables. OK, not likely but now and then this old fossil reads a bit about your field. (If you want to write about the problems in making a non-experimental confidence interval around an unconditional prediction from a artificial neural net, I will be very interested even though I will almost certainly never use the results.)

I am interested in the “science” of statistics, but I don’t read these pages for it. I read it for the “art” of statistics. Gary King’s post on partisan symmetry directs one to a lot of the science. I am reading some of it now, not that I give much more than a hoot or two about said science (Nothing against Prof. King, whom I admire, but his field ain’t mine). I am reading them because there is some art -- tricks, judgments, comparisons, warnings or what-have-you -- within.

So my standard is much weaker than Michael’s: if it is interesting to the poster (and not overly self-indulgent), maybe it is interesting to the reader. Keep up the good work.

Marty

Thanks to both of your for sharing your thoughts; one of the interesting things (for me, at least) about blogging is that -- unlike, say, a scientific conference or a classroom -- it's not clear who the audience is or what their goals and interests are. Definitely my main criteria for blogging about something is whether I find it interesting, but there's plenty of scope within that (rather wide) topic to slant more to one end or another; so your feedback is great.

I especially like the point about the "art" rather than the "science" of statistics: it's definitely something that an informal outlet like a blog (or a conversation) is better equipped to get at, but I hadn't thought about it in that way before. And, Michael, your "gold standard" is worth keeping in mind -- I don't know how frequently I'll attain it, but it makes a great goal! :)

1) Get the technical details correct. For example, I do not think that e-mail notification works. At least, it does not work for me. I also have trouble with the "emember Me?" option. Clicking Yes does nothing.

2) Be more interactive. Unless there is a spam problem, there is little reason not to allow comments to appear immediately.

3) Add more bloggers. Why not issue a general invitation to anyone associated with IQSS to join. You may be surprised at how many people out there want to try this whole blogging thing.

4) Add more content. 3) will help with this, of course, but the blog does not have enough material, I think. Shorter links to news items would be welcome. I think Marginal Revolution and/or Crooked Timber serve as useful templates.

5) Be more topical. One of the big news items today involves Harvard's decision to get rid of Early Action in admissions. There are a lot of interesting statistical issues associated with this decision. The IQSS blog should be the place were people go to talk about such topics.

Then again, I made many of these suggestions a year ago, so perhaps I am an outlier . . .