Icelandic gods and goddesses. Two years ago, students
began making some astonishingly simple-minded
statements about their topics, and claiming that they were
basing them on research they had done. When I asked them
to name their sources, many of them could not, and in
conversation it emerged that their source was the Web.
Because I wanted students to become more aware of how
they were using information, the next time I taught the
course I changed the assignment so that it directly involved
studying and comparing Web sites, and evaluating the
information found on them. From these assignments, I have
learned a great deal about how students use the Web, and
how they understand what they are examining when they
use Web materials. Students’ work on this assignment has
changed a great deal, as they now start by using materials
which are designed to train them as better Web users. Thus,
in this paper, I will report on observations made while
observing Web use by:
. middle school students gathering information
. students at Drexel gathering information
. students at Drexel whose task was to evaluate Web
sites, not just use them
. students at Drexel evaluating Web sites after being
trained to improve their Web use
I start from the assumption that material on the Web is
hypertext, and can be analyzed and evaluated according to
research done in the past on how users interact with
hypertext, before the Internet became the standard for all
our activities. I would note, however, that the Web is just
one instantiation of hypertext, and that as such it has
particular issues and hmits associated with it. Therefore, it
is with some caution that I would apply studies of users in
other hypertext contexts to Web situations.
To be sure, some researchers distinguish between small
hypertexts and large ones, but with the Web, we are dealing
with an information etwork that is not large but vasf and,
in addition, dynamic, so that it is difficult even for
systematic researchers (not to mention users) to describe
the nature of the universe being traversed. Spool et al., for
example, did a study of users’ interaction with nine web
sites, but had difficulties characterizing several of these
sites because two of them changed and one disappeared
from the Web in the course of their study.
In particular, the characteristics of the Web that seem most
problematic (for researchers, as for users) are its fluidity of
boundaries, its vastness, and its dynamic nature. Users
operating in what I would call a “closed” hypertext universe
don’t encounter some of the difficulties found in the “open”
context of Web use. Ways of examining closed hypertext
use don’t seem adequate to assess the effectiveness of Web
use. For example, Foltz has studied the use of hypertext
documents through the Kintsch model which measures how
well the user. builds a hierarchical structure based on work
with the document. This is not suitable for a Web-based
situation, where a user would then be building a hierarchical
structure encompassing a potentially infinite amount of
information. And yet the need for the user to build a
representation f the material he is traversing is real.
Work is beginning to be done on how users interact with the
Web, however. Spool et al. conducted astudy in which they
assigned a series of tasks of differing complexity to Web
users, and both monitored users’ behavior as they searched
and debriefed them when they were done. However, their
study was not a true investigation of the “open”
environment of the Web. They placed users at a specific site
and told them that they did not need to leave the site to
answer the question. Typical Web users have no such
anchor to incline them to remain at any particular site and
use the material found there. However, the users recruited
for the study were all familiar with the Web: their
experience ranged from two weeks to two years. It is likely
therefore that their behavior was conditioned by more
typical Web use and not by the artificial constraints of the
study.
The circumstances of my own observations, while far from
ideal, nonetheless involved looking at users in the open
environment of the Web. The users I observed were not
constrained as to location or site. In studying Olympia
School District students earching the Web as part of a
large project for computer literacy, I made the following
observations:
l Students deal with the overwhelming nature of Web
information by imposing arbitrary limits on what they
find. If their search yields no results, students will explain
that there was no material on the topic, without
considering that they may have misspelled the name, or
that their choice of syntax produced a less-than-successful
search. They rationalize the strategies in a pseudo-
scientific fashion, but the vastness of the Web makes it
unlikely that they will notice and learn from their mistakes
and become better searchers on their own. For example, I
watched amiddle school student look for information on
ancient Egypt. He typed his search terms into AIta Vista,
and scanned the results. He explained that the first page of
results is all that he needed to look at because the search
engine always puts the “best matches first.” The same
student also passed by a page on a sixth grade project on
Egypt. When asked why, he said “it tells what the
students did, not about ancient Egypt.” He was on the
quest for facts, not process. Another student, doing the
same search, used the search engine Excite and described
his results as “summan ‘es” of the sites, while in fact the
search engine was showing him the opening words of each
of the sites on its first page of results.
l Students often misunderstand what they are looking at
when they use the Web. They make mistakes because of
their own misinformation, as well as because the Web
leads them to unsuitable information. For example, when
266