Journal of the American Statistical Association Francisco J. Samaniego -
Editor-Elect - Theory & Methods Department of Statistics, University of
California, Davis

Instructions For Referees

General remarks:

1)Please return the (electronic or paper) postcard to the Associate
Editor promptly.

2)Referee report.  When your review of the manuscript is finished,
return  completed form to the associate editor, along with a copy of
the manuscript if it has been marked. You may, if you wish, submit your
report on a separate page (or separate pages) with the title "Referee's
Report on JASA Ms. # ", followed by the ms number, at the top.

3)Manuscript.  Keep a copy of the manuscript.  If it is revised, you may
be asked to compare the revision with the original.  If it is rejected,
the editors may need further amplification of your report.  If you have
written comments in the manuscript, send this annotated copy to the
Associate Editor with your report.


JASA highly values your services as a referee and depends on you to
help maintain the quality of the journal and get good manuscripts
into publishable form. Thank you for the time and effort you expend as
a referee.

Criterion for Acceptance and Rejection.  Keeping in mind that we want
JASA to continue to be a premier publication, we should be somewhat tough.
The current average paper length is about 25 double-spaced typed pages (or
about eight printed pages).  Mathematical correctness is not sufficient
grounds for publication.  Generally, we are looking to publish papers
that are important and novel extensions to the literature.  However,
if a manuscript really breaks new ground, it might well be a valuable
contribution even if it can't be thought of as definitive.  Of course
we prefer an ambitious paper to one that makes minor improvements.
For more about what we are looking for, see "Information for Authors"
in the front of the JASA.

What to look for.  Most JASA papers contain mathematical developments
that need to be checked.  Everyone would agree that papers published
in JASA should be, at a minimum, correct technically.  Please check
the technical content as carefully as possible, posing questions to
the authors on points that are doubtful or unclear.  In addition to
commenting on the mathematical issues that arise, a referee should
step back from the manuscript's details and reflect on the questions of
novelty, creativity, completeness and significance. Does the paper really
take a step forward?   Is it sufficiently ambitious (or does it leave
too many unanswered questions)?  Does it propose something that's clearly
better than standard practice (at least in some specific circumstance)?
Is it likely to have a positive impact on research or practice in the
area to which it contributes?  Comments on these points are especially
helpful in assessing the paper's suitability for publication in JASA.

Confidentiality.  We never disclose reviewers' names to authors even
when reviewers explicitly permit such disclosures.  You should regard
the submitted manuscript as a confidential document.

Timeliness of Reviews.  We would like reviews be completed within four
to six weeks, with appropriate allowances made for refereeing the longer
or more complex papers.  If you need more time than that suggested by the
Associate Editor, please let the Associate Editor know when you'd be able
to complete the review.  You should aim at writing as comprehensive a
review as is possible given the operative time constraints. But short
reviews confirming the paper's importance or identifying some of the
paper's flaws are also useful. You may find it helpful to read Leon
Gleser's fine article on refereeing in The American Statistician,
November 1986, Vol. 4, No. 4, pp. 310-312.

Courtesy to Authors.  Referee reports are usually sent unedited to
authors.  While many manuscripts will turn out to be unsuitable for
publication in JASA, the authors have nevertheless spent great effort to
produce them.  It is always a good idea to acknowledge whatever merit the
manuscript may contain.  Sweeping criticisms should be avoided in favor
of particular criticisms.  Severity of a review may cripple enthusiasm
and future research productivity, especially with younger authors.

Copies.  Please be sure to retain copies of the manuscript and your
referee's report before sending them to the associate editor. Keeping
your detailed notes on the manuscript may also prove useful if and when
a revised paper is examined.

Recommendation to Publish.  Please give detailed comments.  Most accepted
manuscripts will require some revision by the authors.  Suggested
revisions should not be extensive and should be principally those
you feel are obligatory.  You may suggest discretionary revisions,
but make that clear in your review.  The Editor and Associate Editor
often check a revised manuscript, but you may also be asked to do so.
In that case, refer to your original report to make sure obligatory
changes have been made. This would not be a good time to raise new points,
but if an technical error surfaces during the review of a revised paper,
it should certainly be mentioned.

Recommendation not to Publish.  Give one or two cogent reasons for
rejections, even though you may find much that is wrong.  About 75%
of submitted manuscripts are not published, and most of these are
rejected by the editors and reviewers after the initial submission.
The manuscripts sent to you for review have been seen by the editor and
associate editor, but you should not regard this as an endorsement.

Comments Sent to the Author.  Usually your detailed comments will be sent
you verbatim to the authors. It's best if these comments do not include
any specific recommendations regarding publication.  It is preferable that
you make your recommendation on acceptance or rejection of the paper in
a separate note to the Associate Editor. The recommendation section of
the Referee Report, and any personal references, are erased before it
is sent to the authors.  Anonymity is always preserved.