1. Introduction
1.1 Background Information
The increase of computers in recent years has been contributed by
peoples' great demand for information. One of the most common places to
retrieve information is the internet. In order to make information
retrieval on the internet efficient, indexes, menus and links must be
carefully designed on each web page. Our experiment looks at the different
ways in which links can be arranged. There are two styles which a
web page designer can use to arrange his/her links, either emphasizing
breadth, where a large number of links appear on each page or
emphasizing depth, where a few links will appear per page, creating a tree
structure which is narrow.
We did not find any experiment published that looked at the
structure and arrangement of web links, but we were able to find a lot of
experiments comparing depth and breadth of menus. Since web pages are
designed in such a way that the menu of the pages are spread out within
the contents, we decided to follow the method of experimentation that
was employed in these menu experiments.
1.2 Other Experiments
Kiger(1984) did an experiment that gave users five modes of
varying menu designs. The menus were indexes to a database of some
information. Subjects participating in the experiment were asked to find
some piece of information in the database then to answer a questionnaire
regarding the experiment. The results of the experiment showed that, the
time and number of errors increased with the depth of the menu structure.
The participants ranked the menus with greater depth as the least
favorable.
An experiment by
Jacko and Salvendy (1996) showed that people consider deep menu
structures as complex. Their study showed that, as the depth of the menu
increases so does the users perception that the task of retrieving
information using the menu increases. This could be accredited to short
term memory load and the existing theory of task complexity.
The theory of task complexity in relation to hierarchical menu
designers is greatly discussed in Jacko and Salvendy (1996)s paper. The
paper referees Frese (1987) which indicates complexity is determined by
the number of decisions that have to be made and by the relationships
among
those decisions. With this understanding of complexity combined with the
results of their study it shows that as menu depth increases, complexity
may also increase. And since complexity increase response time and
decreases
accurate responses, minimizing complexity by creating less depth
structured trees might cause a better response time in searching databases
or the internet.
In his book,
The Psychology of Menu Selection
Kent
Norman suggests that the organization or naming of the indices or menus
plays an important part in the users response time. By using a well
organized menu structure, the response time can be reduced with much
broader menus.Norman also suggests that mixing depth and breadth might be
the best alternatives. Menus which are broader at the top or at the bottom
give the user a wide range of alternatives at the points of the searching
when the user will tend to make more mistakes. In selecting the wrong path
while using these menu design, the user will need to go back only one
level in the hierarchy in order to view the menu which will contain the
correct alternative. Response time will me reduced because, less time is
spent
searching for the correct paths. If the middle section is broad it
tends to hinder the search. It is in the
middle section that search needs to be focused so as to reduce the
probability of path errors.
Our study takes the design of these experiments on menu
structures and implements an internet experiment to show that just
like in menu structures, breadth will have an advantage over depth in
the design of web page links.
1.3 Commercial Systems
In commercial systems, there is no clear cut as to which design
people have implemented more frequently. We took a look at two different
commercial designers of web pages that were in the same industry . The
sites we chose were those of Brown
University and University of Maryland
. We did not ask our
subjects to test these web pages, but we test them ourselves. We selected
an information finding task and searched for the appropriate information
on
both sites. It took us 7 links two of them being errors to find the
information on
University of Maryland and only 2 links and with no errors with Brown. The
significant difference was due to the fact that Brown University's web
page had a lot
of breadth and very little depth. The fist page has 63 links. The other
pages have less
links, but you rarely go more than 3 links to reach the information you
need.
On the other hand, University of Maryland page had an average of 13
links
per
page and some information could be found as beep as 13 links. The
University of Maryland
pages though not very narrow, all the information is buried very deep in
the hierarchy,
which makes searching of information difficult. This observation showed
us that like
menu hierarchy, a large depth in a web design may lead to more errors and
a longer search
time on the internet.
Continue
Return to the Title Page for Depth vs Breadth in
the Arrangement of Web Links