Re: Abstracts, evidence, and disappointment

[Mickey Rowe <rowe@psych.ucsb.edu>]

HP Mortimer (Mickey_Mortimer@email.msn.com) wrote:

> I know if I were to ever write an abstract I'd make damn sure it was
> accurate and based on valid research.  I guess I give others too
> much credit to do the same.

I fear that Mr. Mortimer is, as he wrote, being naive if I'm correctly
interpreting the tone of his message.  There are a variety of views
about what abstracts are or should be (HP Wagner's take on the issue
is a bit extreme in my experience, but he certainly has a valid
complaint wrt citation).  However, it seems to me grossly unfair for
HP Mortimer to cast aspersions on the intentions of abstract authors
when he hasn't yet been through the process of preparing for a
conference presentation.  Abstract submission deadlines are generally
months in advance of the conference, and when you write the abstract
your goal is to try to tell people what your presentation is going to
be.  Your goal for the presentation is to talk about the most exciting
thing you're working on right now.  As HP Farlow pointed out, research
isn't generally predictable.  Therefore, what you think when you write
the abstract is likely to be different from what you think six months
later and almost certainly different from what you will think when
you're confident enough to submit the work for peer review.

It's my understanding that things used to be different.  Abstracts
used to be more like peer-reviewed publications, and for many
conferences there is still a possibility that a submitted abstract
will be rejected based on an apparent lack of scientific merit.
However, the pace of life is quickening and scientific conferences are
not immune to this sociological change.  Read through the archives on
this list to see how impatient people are to get the latest and
greatest information.

On top of this is the slew of problems created by the rapid growth of
many meetings.  As the number of abstracts submitted climbs, time to
review individual abstracts goes down.  Conference organizers now
virtually rubber stamp any abstract they receive and spend most of
their time just trying to group the abstracts together so that the
meetings are organized thematically.  And even to do that job abstract
submission deadlines have had to be moved earlier with respect to
conference dates thus exacerbating the prediction problem I wrote
about above.

As a result, it is not uncommon to hear stories about conclusions
presented at a conference being the exact opposite of the conclusions
stated in the abstract.  HP Wagner is right to throw suspicion on
abstracts.  HP Mortimer is wrong to suggest that this is because
abstract authors aren't doing their damnedest to ensure accuracy and
validity.  Other scientists also want the latest and greatest, and if
you put into your abstracts only things you are sure will withstand
peer review (or even your own subsequent investigations) you'll likely
find less interest in your presentation.  Don't get me wrong here; I'm
not suggesting that scientists cut corners (or would be more
interesting if they did), but pushing the envelope when writing
abstracts is both understandable and common in contemporary science.

One reasonable way to view a conference presentation is that it is an
advertisement about what you're working on.  Then an abstract is an
advertisement goading people to come and get the whole sales pitch.
In terms of ethics, one of my colleagues has a rule he follows -- if
he sees something written in an abstract then he will not do research
on the same question for a year or until after the abstract authors
have published a peer-reviewed description of the research.  The
rationale here (not universally shared) is that you give other people
the chance to finish their work and get the credit for it.  If you're
interested in the same thing and they can't finish ahead of you
with over a year's head start then they should probably be doing
something else.

In any case, an abstract should be treated as snapshot of a
scientist's work in progress, not an end product.  That's my view on
abstracts...

-- 
Mickey Rowe     (rowe@psych.ucsb.edu)