Today is the one year anniversary of this blog. First of all, thanks to all the readers. And special thanks to commenters and guest posters. This seems like a good time to assess whether I have achieved my goals for the blog and to get suggestions on how I might proceed in year two.
GOALS. My goals in starting the blog were:
(1) To discuss random things that I happen to find interesting.
(2) To discuss ideas at the interface of statistics and machine learning.
(3) To post every other day.
Goal 1: Achieved.
Goal 2: Partially achieved.
Goal 3: Failed miserably. I was clearly too ambitious. I am lucky if I post once per week.
THE BEST AND WORST. Favorite post: flatland. I still think this is one of the coolest and deepest paradoxes in statistics.
Least Favorite Post: This post where I was dismissive of PAC learning. I think I was just in a bad mood.
LESSON LEARNED. Put “Bayes”, “Frequentist” or “p-value” in the title of a blog post and you get zillions of hits. Put some combination of them and get even more. If I really wanted to get a big readership I would just post exclusively about this stuff. But it would get boring pretty fast.
GOING FORWARD. I hope to keep posting about once per week. But I don’t have any plans to make any specific changes to the blog. I am, however, open to suggestions.
Any suggestions for making the blog more interesting or more fun?
Any suggestions for inducing more people to write comments?
Any topics you would like me to cover? (I already promised to do one on Simpson’s paradox).
Normal Deviate 
22 Comments
As an (until now) silent reader, I just want to thank you for your interesting blogging. As a non-statistician, I do not read all of your posts, but the posts on flatland and Stein’s example were highly appreciated. If you have something interesting to say about the axiom of countable additivity (and de Finetti’s questioning of it), at least I would be eager to hear.
Thanks
That’s a good suggestion for a post
Larry
Felicitations on your one-year blogging anniversary! I always look forward to your posts; what is lacking in quantity, you more than make up for in quality.
Suggestion for more fun: more intellectual disputes with Nobel Prize winners! (Or even Sveriges-Riksbank-Prize-in-Economic-Sciences-in-Memory-of-Alfred-Nobel winners.)
Perhaps I should do a whole series on disputes with Nobel prize winners.
A second (until now) silent reader. Very much appreciate your effort. As an applied economist, posts in any way related to (causal) inference are always in demand. A concrete suggestion: what do you think of the problem of multiple comparisons? Have you read Andrew Gelman’s paper on the topic (arguing it is not an issue)? I understand the basic problem, but given the way research is conducted in practice (i.e. test-running a zillion regressions before presenting results; and also looking for heterogeneous effects which is in fashion now in economics) I wonder if there is any way of getting to a workable solution except for papers that are explicitly data-mining (e.g. in genetic research).
Thanks
I did do one post on multiple testing:
But I could certainly do another
Maybe allow anonymous comments?
Not sure what you mean.
Can’t people already do that?
(Anyway,I am not a fan of anonymous posting.)
Thanks for the great blog and insights (and thanks to all of the commenters too – I have learned a lot from them as well). An every-other-day would have been too much – my productivity at work suffers enough once a week.
I would be interested in a post about other types of inference, specifically Fiducial and Likelihood. I don’t know enough (read: anything) about them to know how/why/when/if they are in use. I just hear about them in passing every once in a while.
Thanks again!
Fiducial and likelihood inference.
Will do
N.D. I always greatly appreciate your blog! For a recommendation, to enable some of us to be lazy and not try to search for a published work that contains the gem we just read in your post, it would be good to remember the boring business of references. Who can really cite a blogpost in published articles? In that connection, can you point me to a published paper of yours that includes the remarks on randomization (and causation) from your last post? Much gratitude.
I’ll try to be better about references.
Regarding randomization and causation: it’s in my book!
I’m an MSc student who for a long time was struggling to understand the differences and similarities between PAC/Vapnik-style learning theory and classical frequentist statistics. I’ve found your blog (and book) absolutely invaluable in this regard, and haven’t been able to find too many other places where the approaches are discussed together. Please do keep it up!
Congratulations! I enjoy reading your posts very much. Keep posting!
Larry, I would be interested in a contrast between the view that Rubin’s Bayesian bootstrap suggests the bootstrap is bogus and Efron’s take on it [Bayesian bootstrap] that it suggests the bootstrap is fundamental. (It came up here https://normaldeviate.wordpress.com/2013/03/19/shaking-the-bayesian-machine/#comment-7935. )
To me the bootstrap is obviously one of the most useful tools in statistics.
In my opinion, the results of Gine and Zinn make it all pretty clear
why/when it works.
I don’t see how anyone can say it is bogus unless they can find an error
in Gine/Zinn (which I doubt)
I’d appreciate math that displays in my RSS reader. I’m not sure if it’s possible without some fairly heavy work on your part.
See, for example: http://www.noamross.net/blog/2012/4/4/math-in-rss-feeds.html
I will look into it.
I use wordpress.com not wordpress.org
which makes the math more difficult.
As long as you’re taking requests……
Empirical Processes is a topic I wish I understood better. Even just what the basic theorems, tools, and goals are. And how it relates to more applied classical and modern statistics.
Some examples where saddlepoint approximations are useful would also be nice. The references I’ve seen are 20 years old or older, and I have a hard time relating to their goals. Or the lack of computational power they were trying to compensate for.
In general I feel like the statistical community has been slow to pick up variational bayesian methods. It’d be great if those techniques could get more attention. Or if someone could explain where their limitations lie and why sampling-based methods are preferable.
Also I’d ignore Mayo’s comment on citing blogposts. As platforms of publishing change what’s acceptable to reference needs to change as well. Plenty of papers used to, and still do, cite private communications with other specialists in the field. If that’s allowable then certainly blog posts are.
I came across your blog during a debate about p-values (ironic because of what you posted about hits). Interesting read. Anyways, I’ve always been into (a) controversial topics that may have a clear answer, but not consistent among statistical experts, and (b) modern topics and community events.
As some personal interests, I’d love to read some more about Empirical Processes, Influence Curves, and Functionals if you ever are inspired. Don’t know if that’s the audience you’re aiming towards however.
A little late but congratulations from me too! Love to read your stuff, and particularly how you can explain some ideas clearly in a few lines others write half a book about without reaching the same level of clarity!
An idea for a posting is perhaps: “Where it is useful to make obviously totally wrong model assumptions”. In case you have anything to say about that.
Thanks
Interesting idea.
I’ll have to see if I have anything useful to say about it though.
2 Trackbacks
[…] the role of Bayesian statistics has generated a lot more smoke than light. As Larry Wasserman at the Normal Deviate blog notes in his blog birthday self-reflection post, one of his lessons learned is that “Put […]
[…] After posting yesterday on why I’m reluctant to move out of my established Word/Endnote workflow, the post quickly became the most read in a single day on The Lab and Field (maybe I should have also put “Bayesian” or “frequentist” in the title!). […]