Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 23 down vote favorite
8

I have a manuscript on a bootstrap method for testing hypotheses of one mean, and I would like to send it for publication, but I have a moral dilemma. I have signed on to the protest against Elsevier for their unethical business practices, and reading up on the whole issue really made me question the ethics of other for-profit academic journals. I would therefore like to publish in a journal which is non-profit, preferably open-source, even though I understand that such journals aren't yet considered on par with the more established journals as far as prestige is concerned. Fortunately, since I already have tenure, that's not a big consideration for me.

I would appreciate any recommendations.

share|improve this question
2  
Larry Wasserman suggests foregoing referees altogether! stat.cmu.edu/~larry/Peer-Review.pdf – shabbychef May 9 '12 at 17:31
Welcome to the site, Anna. I have seen and appreciated your questions and answers on MO. I've even intended to answer/comment on one of your questions, but have not (yet) found the time. – cardinal May 9 '12 at 17:45
1  
I think publishers charge for online articles because there is an expense to publishing that they have a right to recoupe. Many journals give away abstracts and a select set of articles. But shifting from paper to electronic does not mean that they are no longer entitled to recoupe losses and earn a profit for their service which includes editorial expenses I think. – Michael Chernick May 9 '12 at 19:01
1  
I can't recommend anything, but wikiepdia has a huge list of relevant links over at en.wikipedia.org/wiki/List_of_open_access_projects – naught101 May 10 '12 at 1:14
2  
@shabbychef: foregoing referees, yes, but Wasserman isn't really arguing for getting rid of peer-review generally, just opening it up. Also, he notes that such a solution isn't necessarily good for medicine, I would argue that it's also potentially a bad idea for controversial sciences like climate science (where there's a political incentive to write misleading papers). – naught101 May 10 '12 at 1:22
show 1 more comment

3 Answers

up vote 16 down vote accepted

A rather lengthy list can be found at the Directory of Open Access Journals. Using the search term statistics yielded a list of 80 open-source journals as of the time of this post.

I have had good experiences and success in the past with the IMS and Bernoulli society co-sponsored open-source journals, particularly

  1. Electronic Journal of Statistics
  2. Electronic Journal of Probability
  3. Electronic Communications in Probability

The flagship IMS journals (AOS, AOAS, AOP, AOAP) offer a pseudo open-source arrangement where access to current articles is restricted but one can download new (forthcoming) articles for free. See the Annals of Applied Statistics "Next Issues" page for an example. Posting official versions to arXiv is also permitted by these journals, though the statistical research community does not yet appear to use this resource as much as others.

Some other journals have gone to an online-access model recently, including, e.g., Sankhya.

share|improve this answer
Thanks, this is really helpful! – Anna Varvak May 9 '12 at 19:57
@AnnaVarvak: I'm glad you found it useful and hope you to see you continue to participate on this site. Cheers. :) – cardinal May 9 '12 at 22:22
@Rob: Thanks for catching that typo in the first link. I thought I had tested them all, but apparently not! – cardinal May 10 '12 at 13:48
@cardinal A pleasure :) – Rob May 10 '12 at 13:55
up vote 6 down vote

In case your method is somewhat implemented, Journal of Statistical Software is a pretty nice option -- they put emphasis on reproducibility and availability of methods and algorithms.

share|improve this answer
up vote 1 down vote

To add to what cardinal has said the journal Statistics in Biopharmaceutical Research is a purely online journal but you do have to subscribe. Like what cardinal says about Annals of Applied Statistics this journal does give selected issues or some individual articles out for free. It is published by Taylor and Francis. I am curious about your article. In your question you state that you want to publish an article about using a bootstrap method to test a hypothesis about a population mean. This is a well studied topic. What is it about your paper that makes it original?

share|improve this answer
1  
So what are your assumptions then? Is this an iid sample from some unknown distribution with a finite mean? Are you testing a null hypothesis that the mean is a fixed value a versus the alternative that it is different from a? Are you then doing something better than inverting a bootstrap confidence interval for the mean? – Michael Chernick May 9 '12 at 20:28
3  
Maybe this thread should be moved to chat... – jbowman May 9 '12 at 22:14
1  
@AnnaVarvak: You are quite likely to already be aware of it, but the method you describe sounds roughly similar to some work within the last few years by various authors in the empirical-likelihood literature. There are a couple of approaches in which "artificial" data points are added to the sample in a strategic way in order to get better small-sample performance. There, if I recall, the inference is on a $p$-dimensional mean. But, on the surface, it seems reasonable to believe a similar approach might work for the bootstrap. – cardinal May 9 '12 at 22:34
1  
This was originally suggested by Jiahua Chen under the name of "adjusted empirical likelihood". Search for it, there should be more than one paper by now; the original contribution, I believe, came out in JCGS, stat.ubc.ca/~jhchen/paper/JCGS08.pdf. There's a saying the empirical likelihood literature that every empirical likelihood method has its bootstrap analogue (e.g., there are block empirical likelihoods for time series that correspond to block bootstraps), but I never could really get this analogy. – StasK May 10 '12 at 3:04
4  
Quick reminder. A suggestion is generally made to move long series of threads to dedicated chat rooms. The system is supposed to handle this automatically. When this is not the case, you can create one (you only need 100 rep) or use Ten Fold directly. This allows you to continue the discussion at your leisure. – chl♦ May 10 '12 at 8:32
show 12 more comments

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.