Hi all,
At 10:31 AM 10/12/00 -0400, Paul Leiberton wrote:
>It is frequently stated in intro texts books that chunking can increase
>the capacity of STM.
>
>But isn't there research that would support the idea that th size of the
>item in STM affects the capacity of STM? While 7+/-2 works fine with
>numbers of one syllable, if the numbers given to a subject were longer
>(e.g., 178, 438, etc) what can be retained in STM drops.
>
>One of my high school students pointed this out to me after viewing a
>video that contained a filme clip of a subject demonstrating the effect of
>word length on STM capacity. He wondered if the notion of chunking should
>be discarded.
It turns out this isn't an either-or situation. Chunking, by definition,
refers to
creating larger *meaningful* units of information. And, researchers like
Chase & Simon have demonstrated that some chunking strategies can
increase "STM" capacity dramatically for that particular kind of information
(e.g., the long-distance runner who was able to improve his digit span to
80+ by chunking the digits into meaningful running times; note that his span
for other materials, such as letters, remained 7+/- 2). So, while "178, 483"
might not improve memory relative to "4,7", chunks like "1776, 1492, 1812"
will.
But there is a tradeoff here. Baddeley and colleagues (and others), in their
research on the "phonological loop" component of working memory, have shown
that,
for lists of unrelated stimuli, people can effectively recall immediately
the amount that
they can articulate in about 2 sec. So, words that take longer to say aloud,
as Paul noted, lead to smaller word spans than do words that take shorter
to say
aloud (and this is true whether the information is presented auditorally or
visually. Interestingly, word "length," per se, is not the critical
variable here; single-
syllable words that differ in their pronunciation speed also show the
effect (e.g.,
"check" vs. "lathe").
So, it appears that meaningful grouping of information into units can work
to improve
immediate recall, while at the same time, groupings that are too large may
work to
impair recall. My hunch, however, is that effective chunking will usually
trump
articulation length. I'm not sure whether these two have ever been pitted
against
one another in a single experiment.
Best,
Mike
************************************************
Michael J. Kane
Department of Psychology
P.O. Box 26164
University of North Carolina at Greensboro
Greensboro, NC 27402-6164
email: [EMAIL PROTECTED]
phone: 336-256-1022
fax: 336-334-5066
