Re: [Phys-L] refereeing

[John Denker <jsd@av8n.com>]

On 09/10/2013 05:20 AM, Carl Mungan wrote:
> Suppose you worked in a large research group. Suppose a subset of
> members of the group (not including you) wrote a paper. Suppose the
> journal asked you to review the paper (with no mention in the request
> acknowledging they want your opinion even though you're part of the
> research group). Suppose it might not be casually obvious you're part
> of the group because you have a different address. Suppose other
> members of the group found out you were sent the paper to review
> (because you foolishly let it slip out). What would you do? 

Short version:  Tell the editor what's going on.
 *) If you think you can still give a fair review, upholding
  the best interests of the journal and its readers, say so.
 *) If not, say so.
 *) The editor might want to take you off the job anyway,
  just to avoid the /appearance/ of conflict of interest.
  It's an unusual case, involving two mistakes:
   -- Sending the manuscript to the same institution is not
    the usual practice;  even an anonymous reviewer might be
    tempted to give his buddies a boost.
   -- The breakdown of anonymity amplifies the downside potential.

Either way, I wouldn't worry about it too much. Anonymity is overrated.
The objective is to get a fair review.  Sometimes anonymity helps 
maintain fairness, but that's only in cases where the referee needs 
to be protected from browbeating.  If the paper is accepted, or if 
it is rejected for well-documented objective reasons, the issue does 
not arise.  I've done plenty of non-anonymous reviewing, and there 
have only been a few half-hearted and one serious attempt at browbeating.
These had no effect on me.

The worst case is if the paper is correct, but so unclear and/or
unimportant that it is not suitable for publication.  Clarity and
importance are judgment calls, partly subjective.  Reasonable 
persons could have different opinions.

Even so, the referee's opinion is the one that counts.  Hans Bethe 
said "Never argue with the referee.  He's representing the readership 
of the journal.  If the referee is too %@#= stupid to understand the 
paper, the readership will be too %@#= stupid also."

IMHO he overstated it a little bit, but only a little.  So I would
expect the authors to /want/ you to review it.  If there's something
wrong with it, wouldn't they rather find out before publication
than after????

In the product-development world, testing proceeds in stages:
 a) "Alpha" testing is done in-house.
 b) "Beta" testing exposes the product to a smallish number
  of outsiders
 c) Then there is mass deployment

The same thing goes for papers:
 a) You get your friends to critique a draft.
 b) Then you send it to the journal, and they arrange for a
  couple of formal reviews.
 c) Then there is mass publication.

I'd rather do an alpha-review than a beta-review any day.  That's
because at the alpha stage you can make constructive suggestions.
Most journals insist that their reviewers not do that;  they're
supposed to be evaluators, not collaborators.

You can blur the distinction by sending out lots of preprints.  On
more than one occasion I've gotten feedback from somebody who was
obviously reviewing my paper for the journal, but to maintain the
pretense of anonymity he pretended he was asking about the preprint.

========

There have been times in my life when I spent a huge part of my
work-week talking to people who wanted to bounce ideas off me.
Sometimes half-baked ideas, sometimes highly polished manuscripts.

I remember one paper where I said it's all fine, except you need
to make a slight change in the title.  You need to add the word 
"Not", because the paper pretty much proves the opposite of what 
you set out to prove.  That provoked a lot of yelling and screaming
and throwing things, all day every day for two weeks, but finally
they realized that indeed the data was good and the analysis was
good and the unexpected result was actually /more/ interesting
than the expected result, so they changed a few words and everybody
was happy.  They wound up winning some sort of "best in conference"
award.

=========

Bottom line:  I wouldn't worry about it too much.  Anybody with 
any sense would /want/ you to review the thing.  But let the 
editor know.  Let him make the call.