Hi, 

I'm a reporter for Science magazine doing a story about the increasing use
of Bayesian approaches in many fields. I'm interested in hearing from
anyone on this list on two issues:

1) Thoughts on why the Bayesian approach is growing in popularity, and
examples of particularly important, exciting, or easy to understand
applications that demonstrate a major difference from frequentist
approaches. Also, any examples or thoughts on whether the old
frequentist/Bayesian debates are increasing, dying out, or moving to new
ground.

2) I've heard many times that "Bayesians can come up with different answers
to the same problems." I'd like to throw out a "problem" and a data set
that will be used to illustrate my story.

Here is the question: How many papers with "Bayes" or some variant in the
title or abstract will be published this year by the 6,000 science and
social science publications tracked by the Institute for Scientific
Information in Philadelphia.

The data: Here is the information I already have from ISS:

Year    # Bayes papers
1991    606
1992    631
1993    705
1994    862
1995    865
1996    1028
1997    1139
1998    1158
1999*   1110 (*thru 9/30/99)

Here is what else I know: 
In 1991, there were 695,688 papers total in the sciences only.
In 1998, there were 954,761 papers total in the sciences only.

>From 91 to 98, there was a 37.2% increase overall in the number of science
papers, but about a 91% increase in the Bayes papers (I haven't checked
this math, got it from Institute for Scientific Information in Philly).

So, the question is, what will the number of Bayes papers for 1999 be?

I'm also interested in hearing how frequentists might answer this problem.

I need the answers by early next week (the week of 10/11) or sooner.

Thanks! Any and all other comments/questions welcome.



David A. Malakoff
Reporter, Science Magazine (www.sciencemag.org)
1200 New York Avenue NW
Washington DC 20005
(202) 326-6446
(fax) 371-9227
Email: [EMAIL PROTECTED]
Alternate email: [EMAIL PROTECTED]

Reply via email to