OK, so I've been trying to follow all of this, and have to say, I'm finding
it all very interesting. I want to give a special shout-out to the cataloger
who have joined in; I (and, I think, much of code4lib) need this kind of
input on a much more regular basis than we've been getting it.

At the same time, I'm finding it hard to determine if we're converging on
"when trying to turn LCSH into reasonable facets, here's what you need to
do" or "when trying to turn LCSH into reasonable facets, you've haven't got
a freakin' prayer".  Can someone help me here?



On Sun, Apr 17, 2011 at 10:40 AM, Simon Spero <[log in to unmask]> wrote:

> On Fri, Apr 15, 2011 at 7:21 PM, Kelley McGrath <[log in to unmask]>
> wrote:
>
> >
> > It used to be that geographical subdivision was much more flexible and
> was
> > supposed to convey different meanings depending on where it occurred in
> the
> > string. Then there was some research showing that not only did users not
> > know how to interpret this, but catalogers did not understand these rules
> > and were constructing inconsistent headings.
> >
>
> The main study on this subject was the Michigan study performed/led by
> Karen
> Markey (some reports were written as Karen M. Drabenstott.  The final
> report
> of the project is available at
> http://deepblue.lib.umich.edu/handle/2027.42/57992 .  The work took place
> in
> the mid to late 90s, after  Airlie .
>
> This study had serious methodological problems; these became apparent
> during
> the course of the study, and were partly due to the results being so
> unexpected. Unfortunately, there have not been any follow up studies at
> scale that would correct for these methodological issues.  Some of the
> scoring approaches used by the Gleitmans for  "Phrase and Paraphrase" might
> be revealing.
>
> The most perplexing results were those that showed that measured
> understanding was lower when headings were displayed in the context of a
> bibliographic record rather than on their own. This indicates either a
> problem in the measurement process, or an either more fundamental problem
> with subdivided headings that may so negate the significant theoretical
> advantages of pre-coordination that the value of the whole practice is
> thrown in to doubt.
> (Incidentally, this year is the diamond anniversary of the  pre- v. post-
> debate)
>
> Touching on another topic, I believe that   the movement of geographical
> subdivisions to follow the right most geographically sub-dividable
> subdivision can sometimes be interrupted by the interposition of a $x
> topical subdivision, but I haven't determined whether this is a legacy
> exception (the ones that came to mind were related to subtopics of the US
> Civil War, which seems inevitable given that  the first elements are
> "United
> States--History--Civil War, 1861-1865--").
>
> Simon
>



-- 
Bill Dueber
Library Systems Programmer
University of Michigan Library