I just wanted to give a little clarification here about abstracts (I was a database vendor for many years, and now I'm teaching in LIS for many years, so I've kind of been through the drill....)

First of all, any database of scholarly journal articles will only have author-supplied abstracts. That's just the way it works; so, if you're searching the serious stuff, you needn't ever worry about this, as it's the only way it's done .

For trade journals, it's a bit of a crap shoot, and depends on the journal itself. Many trades will take author-supplied abstracts, while others don't require it. If an author writes one, it is very rare that the journal would not print it verbatim. On the other hand, if you're talking popular publications, it is very rare that abstracts would even exist. When abstracts do exist in purely popular trades (like BusinessWeek, Aviation Week, etc.) they are virtually always written by the publisher.

Whenever abstracts appear to be almost identical to lead-ins in the text, this can be the choice of either the publisher or the author. (Note: many scholars - serious researchers - believe this is entirely appropriate, and while getting my Ph.D., I was actually told by some professors that it's supposed to be this way, though I've not done it myself whenever I've published. It is entirely a matter of style, and shouldn't be interpreted in a negative sort of way, as it's considered the right thing to do in certain academic circles.)

Lastly, and perhaps most to the point, there is no way, to my knowledge, to ever limit a database search to "author-supplied" vs. "publisher-supplied" abstracts. I've never seen it, and I know of no way it could technically be done, since database developers do not have access to that sort of information. I've built many of these databases, and absolutely never known of this ever to occur. However, I've also worked with many publishers, both as a researcher/writer and as an indexer, and was always told this isn't done.

So, "in closing" - HA! - ballpark guideline is to just keep reminding yourself while searching that abstracts are always author-supplied whenever you've got a scholarly journal, or any "serious" trade journal; and that they're likely to be publisher-supplied in popular and "lighter" trades.

-----Original Message-----
From: SERIALST: Serials in Libraries Discussion Forum
[mailto:SERIALST@LIST.UVM.EDU] On Behalf Of Steve Rossiter
Sent: Sunday, May 24, 2009 6:18 PM
To: SERIALST@LIST.UVM.EDU
Subject: [SERIALST] Whose abstract?

In doing searches in databases such as OmniFile (Wilson), OneFile (Gale /
Cengage), Business Source Premier (EBSCO), Research Library (ProQuest) you
will often find abstracts preceding the main body of the article.

At least in the case of ProQuest some of these abstracts / summaries are not
author supplied but merely a snippet of the first paragraph or some portion of
it and not very telling as to what the actual article is about.

I am curious to know if there is a listing out there somewhere, print or
online, which indicates whether publications come with author supplied
abstracts?

The reason I am asking is that in trying to determine which articles might be
of interest it would be easier to review a set of abstracts than having to
read the entire set of articles. Computer generated summaries really do not
work very for this purpose although I imagine most authors will try to
emphasize their main ideas in the lead paragraph.