Christian has a fun post about the rise of the B-word (Bayesian). “Bayesian ” kills “frequentist.”

Well, how about the other B-word, “Bootstrap.” Look at this Google-trends plot:

The bootstrap demolishes Bayes!

Actually, Christian’s post was tongue-in-cheek. As he points out, “frequentist” is *… not a qualification used by frequentists to describe their methods. In other words (!), “frequentist” does not occur very often in frequentist papers.*

But all joking aside, that does raise an interesting question. Why do Bayesians put the word “Bayesian” in the title of their papers? For example, you might a see a paper with a title like

“A Bayesian Analysis of Respiratory Diseases in Children”

but you would be unlikely to see a paper with a title like

“A Frequentist Analysis of Respiratory Diseases in Children.”

In fact, I think you are doing a disservice to Bayesian inference if you include “Bayesian” in the title. Allow me to explain.

The great Bayesian statistician Dennis Lindley argued strongly against creating a Bayesian journal. He argued that if Bayesian inference is to be successful and become part of the mainstream of statistics, then it should not be treated as novel. Having a Bayesian journal comes across as defensive. Be bold and publish your papers in our best journals, he argued. In other words, if you really believe in the power of Bayesian statistics, then remove the word Bayesian and just think of it as statistics.

I think the same argument applies to paper titles. If you think Bayesian inference is the right way to analyze respiratory diseases in children, then write a paper entitled:

“A Statistical Analysis of Respiratory Diseases in Children.”

Qualifying the title with the word “Bayesian” suggests that there is something novel or weird about using Bayes. If you believe in Bayes, have the courage to leave it out of the title.

### Like this:

Like Loading...

*Related*

## 14 Comments

bootstrap is rising because it’s the name of a javascript library, which is becoming the dominant interpretation of the query.

http://www.google.com/trends/explore#q=bootstrap,%20twitter%20bootstrap

I see hockey sticks everywhere…

“bootstrap” is a “sleek, intuitive, and powerful front-end framework for faster and easier web development” (from http://twitter.github.com/bootstrap/ ). no wonder that it is taking off like crazy :)

PS: yes, I agree with Dennis Lindley. However, to get you paper published (particular in non-statistical journals) having “Bayes” in the title is a plus and makes perfect sense: it signals that your method is modern and cutting edge. This is true also for key phrases such as “nonparametric aproach”, “machine learning approach” etc. so it is not confined to Bayes.

Ah, Larry….!

Following one comment on my post, I found on Ngram that many more books (not papers) use posterior distributions than Bayesian or Bayes… So maybe the authors do refrain from putting so much stress on the fact that they are Bayesian. And maybe all this does not make much sense given that it is only analyses books, rather than papers…

interesting!

another interesting Ngram search is:

confidence interval, posterior distribution

People should have listened to Lindley, if he said this*. The prevailing attitude seems to be that Bayesians need to have their own everything: ISBN, Bayesian retreats, journals….and in fact, so far as I could tell, the JSM does not list methodology in general but only Bayesian. Then of course there’s the constant Bayesian cheerleading or “tribal drums” as someone on Gelman’s blog put it once. I cannot imagine frequentist error statisticians doing this…

*If he really did believe in his Bayesian methods, one wonders why he felt the need to write the letter about Higgs physicists practicing “bad science” because they used p values.

“Tribal Drums”?!?!?! Nice! We are kind of “Savages”, anyway!

The Foundations of Statistics

Mayo: “the JSM does not list methodology in general but only Bayesian”. Not true. JSM specifically has sessions on “General Methodology”, go look at the sponsors. It also lists hundreds of sessions on general areas of methodology; e.g. methods for doing nonparametric work, or survey statistics, or data mining, where use of the B word does not determine relevance.

“Constant Bayesian cheerleading” was a fact of life years ago – when many in the field would disdain Bayesian methods. Not true today; try going to an ISBA meeting to see how little cheerleading there is. And if you can’t imagine other statisticians “cheerleading”, check out e.g. Oscar Kempthorne’s [failed] attempt to shout down Lindley and Smith. There are plenty of one-sided commentaries out there.

\begin{flamebait}

On the other hand, “Bayesian” in the title makes such papers easier to ignore.

\end{flamebait}

I kid, I kid! Couldn’t help myself…

Larry:

I agree. This reminds me of a foolish thing that Leo Breiman once wrote, claiming in 1997 that Bayesian applied statistics was nearly nonexistent, based on the evidence that there were almost zero papers that included the words “data” and “Bayes” or “Bayesian” in their keywords.

He somehow had no idea that a paper with titles such as “Physiological pharmacokinetic analysis using population modeling and informative prior distributions” or “Estimating the electoral consequences of legislative redistricting” could represent applied work. After all, we didn’t have “data” in our title or our keywords. How applied could a paper be, if it doesn’t include “data” as a keyword???

I’m attending the MIT Sloan Sports Analytics Conference. The research papers compute posterior estimates. But I haven’t seen the term ‘Bayesian’ once.

I agree that the comparison to “A Frequentist Analysis of…” makes putting “Bayesian” in the title of papers seem strange in comparison. But I think in some cases “Bayesian” in titles can be seen as referring more to method or tool than to a philosophy of probability. That’s especially true in machine-learning papers, where “Bayesian methods” are seen as a particular toolbox. In that case we might compare it to papers that put things like “neural network” or “Support Vector Machine” or “boosting” in their titles. Those may all still be giving a particular technique more prominence in the paper’s labeling than it deserves, but it at least makes “Bayesian methods” no more strange than other machine-learning methods.

I completely agree that if we’re talking about purely applied papers “Bayesian” should be replaced by “statistical”. But if a novel method is being proposed I think it can still be relevant that the method is Bayesian.

## One Trackback

[…] here (Xi’an’s Og)and here (Normal […]