You are on page 1of 28

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/237090970

The Impact of Presentation Format, Task Assignment, and Prior Knowledge on


Students' Comprehension of Multiple Online Documents

Article  in  Journal of Literacy Research · October 2007


DOI: 10.1080/10862960701675317

CITATIONS READS

90 106

2 authors:

Ludovic Le Bigot Jean-François Rouet


Université de Poitiers French National Centre for Scientific Research
75 PUBLICATIONS   583 CITATIONS    171 PUBLICATIONS   3,525 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Dialogue and emotions View project

Investigating the determinants of common ground information saliency View project

All content following this page was uploaded by Ludovic Le Bigot on 04 March 2014.

The user has requested enhancement of the downloaded file.


Journal ofhttp://jlr.sagepub.com/
Literacy Research

The Impact of Presentation Format, Task Assignment, and Prior Knowledge on Students' Comprehension
of Multiple Online Documents
Ludovic Le Bigot and Jean-François Rouet
Journal of Literacy Research 2007 39: 445
DOI: 10.1080/10862960701675317

The online version of this article can be found at:


http://jlr.sagepub.com/content/39/4/445

Published by:

http://www.sagepublications.com

On behalf of:
Literary Research Association

Additional services and information for Journal of Literacy Research can be found at:

Email Alerts: http://jlr.sagepub.com/cgi/alerts

Subscriptions: http://jlr.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://jlr.sagepub.com/content/39/4/445.refs.html

>> Version of Record - Dec 1, 2007

What is This?

Downloaded from jlr.sagepub.com by guest on October 11, 2013


JOURNAL OF LITERACY RESEARCH, 39(4), 445–470
Copyright © 2007, Lawrence Erlbaum Associates, Inc.

The Impact of Presentation Format, Task


Assignment, and Prior Knowledge on
Students’ Comprehension of Multiple
Online Documents

Ludovic Le Bigot and Jean-François Rouet


University of Poitiers and CNRS (France)

This study investigated the impact of prior knowledge, writing task, and hypertext
format on university students’ comprehension of multisource hypertext. Fifty-two
students categorized as having high or low prior knowledge studied hypertext con-
sisting of seven component texts on the topic of social influence. Some students
were given a summary writing task; others were asked to write arguments regard-
ing the topic. Also, some students experienced the component texts organized by
topic, and others experienced them introduced by source (i.e., author and date). De-
pendent measures included reading time and scores on a comprehension post-test
as well as number and nature of connectives and number and nature of idea units
included in students’ essays. High knowledge students spent less time than low
knowledge students and also did better on the comprehension post-test. Moreover,
the argument task led to more causal connectives and more transformed infor-
mation in students’ papers, whereas the summary task resulted in more temporal
connectives and paraphrases. Third, presentation of the material by source resulted
in better macrostructural comprehension on the post-test, whereas the topic format
resulted in better microstructural comprehension. Findings are discussed in terms
of development of hypertext materials and task construction, and suggestions are
made for further studies.

Electronic information resources are becoming increasingly widespread in schools


and universities, in the workplace, and indeed in society as a whole. Very often,
learners find themselves confronted by a mass of information that is more or

Correspondence should be addressed to Ludovic Le Bigot, Laboratoire LMDC, 99 avenue du


Recteur Pineau, 86000 Poitiers, France. E-mail: ludovic.le.bigot@univ-poitiers.fr

445
446 LE BIGOT AND ROUET

less appropriate for performing their activities. When searching for information
or a solution to a problem (a goal-driven activity) in an electronic documentary
environment, readers continually have to assess the relevance of the informa-
tion, constructing their own partial or complete representation of the items of
information contained in the documents. Within this context, research into pre-
sentation formats and their impact on student behavior is of major importance
to hypermedia design. More specifically, upon carrying out a particular task,
how do readers build their own representations of the information in terms of
its form and importance? How do they relate these items of information to each
other? How far do other variables, such as prior knowledge or task specificity,
affect the comprehension of hypermedia documents?
This study uses theories of text and document comprehension as a framework
for interpreting the effects of prior knowledge, task specificity, and sources on
students’ comprehension of instructional hypertext based on multiple documents.
Our purpose was to investigate the influence of hypertext node labeling in mul-
tiple source hypertext on the reader’s construction of a mental representation of
the contents. We review the role of rhetorical structure and content represen-
tations in the comprehension of single texts and hypertexts. Then, we discuss
the case of multiple source environments, both in print and online, and the
specific issue of representing source and content information in Web-like study
environments.

CONTENT REPRESENTATION AND HYPERTEXT


COMPREHENSION

Navigating hypertext (i.e., networks of text pages connected through electronic


links) involves a complex set of cognitive processes that are partly similar to
those involved in the comprehension of printed texts (Dillon, 1994; Foltz, 1996;
Rouet, 2006). Text comprehension theories (Kintsch, 1998; Zwaan & Singer,
2003) assert that readers of written language construct a multilayered men-
tal representation of the text contents. Such a representation typically includes
a propositional hierarchy of semantic propositions that captures the text’s lit-
eral semantic contents, or textbase, and a combination of text contents and
the reader’s relevant prior knowledge, or situation model (van Dijk & Kintsch,
1983). The textbase and the situation model include micro- and macroproposi-
tions. The latter play a structuring role, in that they both condense and connect
other propositions. Kintsch (1998) suggested that two types of factors influence
the construction of an integrated mental representation: the rhetorical structure
of the text (i.e., the type of text schema or macrostructure) and the reader’s
prior knowledge, goals, and strategy. Thus, texts that present “good” or standard
rhetorical structures are better understood by readers (Kintsch & Yarbrough,
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 447

1982), and readers who possess more prior knowledge can compensate for the
lack of explicit information in the text (McNamara, Kintsch, Songer, & Kintsch,
1996). Contextual factors, such as specific tasks or study directions, can also
influence readers’ organization and elaboration of information into their memory
representation (Rouet & Vidal-Abarca, 2002; Winne & Hadwin, 1998).
The usability of hypertext is also strongly related to its rhetorical structure
and the use of organizing devices, for example, explicit menus or content maps
(Charney, 1994). Hypertext readers tend to move along the paths that help them
maintain a sense of global coherence (Foltz, 1996). They rely strongly on cues
that signal the overall structure of the information network, the user’s previous
selections, and his or her current location in the network (Chen, Fan, & Macredie,
2006; Dee-Lucas & Larkin, 1995; McDonald & Stevenson, 1998; Nilsson &
Mayer, 2002). Content representations such as menus, tables of contents, and
concept maps have been found to facilitate hypertext readers’ navigation and
orientation (Britt, Rouet, & Perfetti, 1996; McDonald & Stevenson, 1998; Rouet,
2003).
There is evidence that students respond differently to various kinds of content
representation devices, as a function of their perception of task demands and
their level of prior knowledge. For instance, Dee-Lucas and Larkin (1995) found
that high school students’ exploration and revision of a physics text presented
electronically was made easier when hierarchical content diagrams were included
compared to a simple topic list or a mere series of text pages. Moreover, student’s
recall of section titles was best in the hierarchical condition, and they judged
it to be easier to understand than a printed text. The benefits of a hierarchical
content representation did not appear, however, in a second study where the
students were given a more explicit summarization task, as if the higher task
demands would overcast the benefits of the improved content representation.
Furthermore, Hofman and van Oostendorp (1999) asked college students with
either high or low prior knowledge to study a hypertext on the topic of sun rays.
Compared to a networked concept map, an alphabetical topic list enhanced
low prior knowledge student comprehension. The authors concluded that simple
presentations such as content lists were of greater assistance to readers with a
low level of knowledge than complex ones. Potelle and Rouet (2003) found that,
compared with a list and a network map conditions, a hierarchical content map
significantly improved macrostructural comprehension of a hypertext on social
influence by college students with low knowledge in social psychology. No effect
of the presentation format was recorded for readers with high prior knowledge.
Potelle and Rouet concluded that simpler metaphors for organizing contents
may fit better the needs of low prior knowledge students, whereas students with
higher prior knowledge might be less sensitive to content organization devices.
In short, structured content representations generally are found to promote
students’ integration and understanding of information in hypertext provided
448 LE BIGOT AND ROUET

they are tailored to the readers’ level of knowledge (e.g., simpler diagrams
for students with lower levels of prior knowledge). These organizers facilitate
students’ identification of global semantic links across information units that
make up the hypertext.

COMPREHENSION STRATEGIES IN MULTIPLE


SOURCE HYPERTEXT

As the use of the Internet and multiple document environments become more
common in education and training contexts, students have to use sophisticated
strategies to respond to the specific demands of Web-based learning tasks. Some
theorists consider that current definitions of literacy should be expanded to in-
clude those environments and the specific demands they create. For instance,
Leu, Kinzer, Coiro, and Cammack (2004) proposed that:

New literacies include the skills, strategies, and disposition that allow us to use the
Internet and other ICTs effectively to identify important questions, locate informa-
tion, critically evaluate the usefulness of that information, synthesize information
to answer those questions, and then communicate the answers to others (p. 1572;
See also Coiro, 2003; Reinking, 1998).

Particularly relevant to Information and Communication Technologies (ICT) are


the integration and transformation of source information, a complex skill that
combines advanced comprehension strategies and writing abilities (Jacobson &
Spiro, 1995; Segev-Miller, 2004; Spivey & King, 1989).
Regarding cognitive processes underlying multiple document comprehension,
it has been suggested that specific integration mechanisms may be involved (Britt
et al., 1996; Foltz, 1996; Rouet, 2006; Wineburg, 1991). Unlike information
from single texts, information from multiple texts cannot always be integrated
on the basis of factual coherence. Instead, the texts must be compared and
organized using more global rhetorical predicates, especially when they present
incomplete or discrepant information. Such a situation is typical of both print and
electronic document comprehension (Wiley, 2001). Integration of information
across heterogeneous online documents draws on the reader’s prior experience
in the content area. Early empirical studies found that experienced students use
evaluation and integration heuristics distinct from novices (Rouet, Favart, Britt,
& Perfetti, 1997; Wineburg, 1991). More specifically, expert readers of multiple
documents tend to rely on the features of information sources (e.g., author, date,
type of text) when making decisions as to what documents to read next (Rieh,
2002), when reflecting about the documents at hand (Strømsø & Bråten, 2002),
or when assessing the documents in retrospect.
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 449

Perfetti, Rouet, and Britt (1999) have proposed a theoretical framework to ac-
count for the cognitive processes involved in multiple document comprehension.
According to their document model framework, readers of complex nonlinear
texts build a mental representation that includes two main components: the in-
tertext model and the situations model. The intertext model may be described
as a network of text nodes connected by intertext predicates. Each text node
contains a set of variables, which describe the rhetorical and pragmatic char-
acteristics of the texts: author, date, rhetorical goals, topic or main argument,
and so forth. Intertext predicates, such as “corroborates,” “opposes,” “supports,”
and so forth, describe the overall relationships between the texts. The situations
model may include aspects of the situation described in only one source, as-
pects common to several sources, and even aspects that are contradictory across
sources. Therefore, contrary to single-text situation models, the mental represen-
tations that emerge from reading multiple documents do not have to be internally
coherent. Instead, contradictory aspects of these situations may be held in an in-
tegrated memory representation through rhetorical predicates such as “according
to,” “from,” “based on,” and so forth, which allow the reader to relate critical
items of information to the relevant sources. The overall representation resulting
from the integration of the intertext model and the situations model is called a
document model.
Even though the document model framework originated from studies of
printed documents, it applies to reading hypertext or other Web-based sources.
Readers of online sources must often relate information through complex rhetor-
ical links as opposed to simpler, coherence-based links. According to this ap-
proach, readers of multiple online documents should devote specific attention to
source information when navigating and relating hypertext nodes. The empiri-
cal evidence shows, however, that inexperienced students learning from Web-
based materials tend to ignore or to misinterpret source parameters, which some-
times results in gullibility errors and inappropriate information selection (Brem,
Russell, & Weems, 2001).
There is also some evidence that providing source information in menus or
content maps can influence students’ construction of document models from
electronic information resources. Britt et al. (1996) examined the effect of net-
working and content representations on university students’ comprehension of
a set of multiple documents in history. The documents were presented electron-
ically either in a linear (page turning) or in a hypertext format. Both versions
contained a table of contents that listed the documents according to source pa-
rameters (authors, date, short title). The hypertext format allowed direct access to
the documents from the contents page, and contained direct cross-reference links
within the content pages. Britt et al. manipulated the organization of documents
in the table of contents. In the structured condition, the documents were listed
in an order consistent with the subject matter and with rhetorical relationships
450 LE BIGOT AND ROUET

(e.g., co-reference) among them. In the scrambled condition, the order of presen-
tation conveyed no sense of structure. Looking at students’ navigation patterns,
they found that the table of contents was used more often in the hypertext and
in the scrambled conditions. In other words, students relied more heavily on the
top-level representation of the document set when the navigation scheme was
more sophisticated (i.e., hypertext vs. linear arrangement), and when the global
organization of the documents was less obvious. This suggests that when con-
fronted with multiple source environments, students indeed try to build up an
integrated representation of “who said what.” Other studies have found evidence
that learning from multi-source environments may have positive impacts on col-
lege students’ writing about complex issues (Jacobson & Spiro, 1995; Wiley &
Voss, 1999).
In summary, students’ responses to both task and presentation format condi-
tion the effectiveness of Web-based document learning. In the case of multiple
document learning, a structured table of contents focusing on document sources,
as well as the use of separate frames for links and contents seem to be most
beneficial to multiple document comprehension. It is unclear, however, if label-
ing the documents may also play a role, as none of the studies mentioned above
examined this issue. In particular, it is of interest to find out whether signaling
the source of a piece of information (e.g., by mentioning the identity of the
document author) would foster the construction of document models.

PRESENT STUDY

The aim of the study was to find out whether a simple content representation
focusing on sources rather than contents would facilitate students’ deep com-
prehension of multisource hypertext introducing specific expository contents.
Based on prior studies of multiple document comprehension (Britt et al., 1996;
Perfetti et al., 1999; Wiley, 2001), we hypothesized that a source-based content
representation would facilitate students’ understanding of document model, that
is, the semantic contents of the hypertext nodes and their organization into a
consistent (though potentially conflicting) whole. Conversely, a topic-based con-
tent representation would encourage a more piecemeal, detail-oriented document
model. These influences would result in differences at a comprehension post-test,
as well as in an essay-writing task. A source-based menu would promote the
reference to documents in students’ essays, especially in the argument-writing
task.
Because previous studies of hypertext learning have found strong effects of
study assignments on students’ comprehension as well as interactions between
assignment and presentation format variables (e.g., Dee-Lucas & Larkin, 1995;
Wiley & Voss, 1999), we also were interested in examining the impact of a
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 451

summary versus an argument-writing task on students’ reading strategies and


on their writing. Instructions to write an argument based on the materials would
result in essays containing more transformed, better interconnected information,
compared to a summarization task (Scardamalia & Bereiter, 1987; Wiley & Voss,
1999).
We expected that patterns of document reading and the quality of students’
essays would vary according to the participants’ level of prior knowledge, specif-
ically that low prior knowledge students would need more time, on average,
to comprehend each hypertext node, and that they may reach lower levels of
content comprehension. We also conjectured that the type of content repre-
sentation would interact with both the readers’ level of prior knowledge and
the type of study assignment. Based on previous studies of content representa-
tion in hypertext (Britt et al., 1996; Potelle & Rouet, 2003), we hypothesized
that a source-based content representation would be more beneficial to students
with a higher level of prior knowledge, working under more demanding task
directions.

Method
We invited volunteer university students to participate in a one-session study,
during which they were asked to study a small-scale hypertext on the topic of
social influence. The students were asked to study the materials either with the
perspective of summarizing the information or with the perspective of writing an
argument on the topic of social influence. Half of the participants in each group
received a version of the hypertext in which the documents were listed according
to the topic covered (e.g., “Conformity”), while the other half received a version
of the hypertexts in which the documents were listed according to source param-
eters in addition to topic (e.g., “Asch (1951): Conformity”). Students’ selection
and study time were recorded throughout the study period. After studying the
materials the students were asked to write a one-page essay either summarizing
or developing an argument on the topic. Essays were analyzed for topics and
the use of connectives. In addition the participants answered a multiple choice
comprehension questionnaire.

Participants
Sixty-five students from a university in France took part in this study. Their
average age was 26.5 years (range D 19–53, S:D: D 6:24). The initial sample
consisted of 30 men and 35 women. Forty-two participants were undergraduate
students, and 23 were graduate students. Twenty-seven had studied psychology
for at least two years. All participants were native French speakers and were
452 LE BIGOT AND ROUET

randomly assigned to four initial experimental conditions. They were subse-


quently divided into two contrasted prior knowledge subgroups (see procedure
below). The final sample included only the participants whose scores were above
or below the contrasting criterion (N D 52).

Materials
Hypertext. Seven short texts about different aspects of social influence were
composed from several social psychology textbooks used for teaching social in-
fluence in university courses. We decided to rewrite the texts to control for
length, writing style, and readability. The topic was chosen because it was rel-
evant to psychology students’ purposes. One text was about social influence in
general, three texts discussed majority influence, and three discussed minority
influence. Each text was organized into four paragraphs with (a) a title, (b) an
introduction to the topic, (c) a brief presentation of an experiment, and (d) a
conclusion.
The texts were presented in the form of hypertext (developed using Microsoft
Visual Basic), with a menu page leading to each individual text. The menu
was either an alphabetical list of topics or the corresponding list of sources
(see Table 1). In the topic list condition, the title focused on the contents and
the text was written in a descriptive form (e.g., “research found that : : : ”). In
the source list condition, the title featured the author and publication date of
a relevant research article, and the text was attributed to the source through
personal involvement markers (e.g., “we found that : : : ”). Apart from those
manipulations, the texts were identical in both presentation conditions. Each
text totaled 145 words (more or less 10 words). The texts did not refer to each
other (see Table 2 for an example). A button at the bottom of each text page
allowed the reader to return to the menu page. After all the texts had been read
at least once, an “answer questions” button appeared on the menu page.

TABLE 1
Presentation Formats (Menus) for Topics or Sources

List of Topics List of Sources

Conformity Asch (1951): Conformity


Majority influence Milgram (1974): Majority influence
Minority influence Moscovici et al. (1969): Minority influence
Social influence Sherif (1936): Social influence
Innovation Schacter (1951): Innovation
Normalization Montmollin (1966): Normalization
Norm and consensus Paicheler (1974): Norm and consensus
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 453

TABLE 2
Example of the Text on the Topic of Conformity in the Third Level
of the Hierarchical Structure (Below the Majority Influence Level)
in the Two Presentation Formats (Topic, Source)

Topic Format Source Format Paragraph

Conformity Asch (1951): Conformity Title or Source

Conformity is a change in behavior Conformity is a change in behavior Introduction


by which a person responds to the by which a person responds to the
pressure of a group in a particular pressure of a group in a particular
situation. It is characterized by situation. It is characterized by
changes in the way the individual changes in the way the individual
behaves and the adoption of norms behaves and the adoption of norms
that are either introduced to or that are either introduced to or
imposed on him or her. imposed on him or her.
Asch (1951) illustrates the conformity We illustrated the conformity process Presentation of
process with a nonambiguous with a non-ambiguous perceptual one study
perceptual task assessing the task assessing the “length of
“length of segments.” The naïve segments.” The naïve participants
participants were placed with were placed with several
several accomplices. The material accomplices. The material
comprised one standard segment comprised one standard segment
and three other segments of various and three other segments of various
lengths. The participants had to lengths. The participants had to
estimate the length of 3 segments estimate the length of 3 segments
compared with the standard compared with the standard
segment. The results showed that a segment. The results showed that a
large proportion of the naïve large proportion of the naïve
participants’ answers complied participants’ answers complied
with those of the accomplices, even with those of the accomplices, even
if these were incorrect. if these were incorrect.
This study shows how participants This study shows how participants Conclusion
comply with norms in reaction to comply with norms in reaction to
group pressure. group pressure.

Comprehension task. Comprehension was assessed by means of a 16-item


multiple-choice questionnaire for pre- and post-tests. Four categories of ques-
tions were created from the combination of two dimensions of comprehen-
sion (Hofman & van Oostendorp, 1999; Kintsch, 1998): the structural level
and the representational level. The two structural levels were microstructure
(i.e., questions about local details) and macrostructure (i.e., question about main
ideas); the two representational levels were text-based (i.e., questions based
on literal information) and situation model (i.e., questions based on the pro-
duction of inferences). The four question categories were Microstructure—
Text-based (i.e., questions about literal details, i.e., a paraphrase; N D 4),
454 LE BIGOT AND ROUET

Microstructure—Situation model (i.e., questions requiring local inferences;


N D 4), Macrostructure—Text-based (i.e., questions about explicit main ideas;
N D 4), and Macrostructure—Situation model questions (i.e., questions requir-
ing global inferences; N D 4). Appendix A presents an example of each category
of question.
Four possible answers were provided with each question, with only one cor-
rect answer. In addition, a “don’t know” answer was supplied with each question
to avoid random selections. After reading a question, the participants ticked an
answer among the five options. In order to make sure that the answer could
be found in the source texts, we conducted a pilot study with 36 undergradu-
ate students with no prior training in social psychology. The participants were
asked to answer the questions with the texts available. The accuracy rate was
well above chance, with a minimum of 40% for each question, which suggests
that inexpert participants could use the information in the source texts to answer
the questions.

Procedure and Design


The data collection took place during a single 90-minute session. Students par-
ticipated in small groups (maximum 5 participants). The session consisted of
four distinct stages: (a) pretest, (b) practice, (c) hypertext reading, and (d) writing
and post-test.

Pretest. The first stage involved an assessment of the participants’ prior


knowledge in social psychology. The participants directly answered the compre-
hension questionnaire using the cursor as a pointing device. Furthermore, they
were strongly encouraged to answer, “Don’t know” if they did not know the
correct answer.

Practice task. In the second stage, participants practiced doing the written
assignment and using the computer interface. They were introduced to the inter-
face via a hypertext system dealing with an unrelated topic (crude oil market).
They were told to read a set (4) of very short texts displayed on the computer,
during a 7–8 minute period.
Depending on the experimental condition, the participants were verbally in-
structed to read the hypertext with a view to writing either a summary of
the texts or an argument based on them, following a procedure used by Wi-
ley and Voss (1999). The instructions for the summarization condition were as
follows:
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 455

You will have to read and understand texts about the crude oil market. You will
then have to write a one-page (about 5–10 lines) summary based on this set of
texts. The summary will have to present the main ideas expressed in these texts on
the subject of the crude oil market. Please pay attention, as you will not be able
to take notes while you read. You will therefore have to read the texts carefully
so that you remember them when you come to write your summary. You have
approximately 7–8 minutes to read the texts. (Underlining is added here for the
purpose of clarity.)

For the participants in the argumentation group, the word “summary” was
replaced by “argument” and the term “main ideas” by “your opinion based on
ideas.” Thus the two sets of directions emphasized different study perspectives
(i.e., reading for main ideas vs. reading to develop one’s own informed opinion),
while remaining very similar in length and style.
The texts were displayed on a computer screen, with a menu leading to each
individual text. The participants were advised to read and review the texts, in
whichever order they wished. After seven minutes, text presentation automat-
ically ceased. The participants could stop reading earlier by clicking on the
“answer questions” button. They then wrote their summary or argument for
5 minutes or until they had written roughly 10 sentences. Finally, they answered
a multiple-choice comprehension questionnaire with four items.

Main task. In the third stage, the participants went through the main task
according to a procedure similar to the practice phase (“crude oil market” was
replaced by “social influence,” “about 5–10 lines” by “about 15–20 lines,” and
“7–8 minutes” by “15 minutes”). However, an additional manipulation of menu
content was introduced. For approximately half the participants, the menu was
made up of a list of topics, whereas for the remaining participants, the menu
was made up of a list of sources. Again, approximately half the participants
were given a summarization assignment, and the remaining an argumentation
assignment. The written assignment given in the third phase was consistent with
the instructions in the practice phase. After 15 minutes, text presentation was
automatically halted. However, the participants could stop reading earlier by
clicking on the “answer questions” button.

Writing and post-test. After reading the texts, the Lefavrais (1968) read-
ing test was used both as a control of the participants’ reading ability and as an
intervening task (i.e., to avoid a recency effect). Then, depending on the con-
dition, the participants were asked to write a summary or an argument, within
a maximum delay of 20 minutes. Finally, they were given the comprehension
456 LE BIGOT AND ROUET

TABLE 3
Number of Participants Per Group

Written Assignment
Presentation Prior
Format Knowledge Argument Summary Total

Topic Low 7 9 16
High 3 5 8
Total 10 14
Source Low 9 6 15
High 5 8 13
Total 14 14

questionnaire. The 16 items were presented on the computer screen in a random


order. The participants were given unlimited time to respond.

Design and contrasting groups. The study was based on a three-factor


between-participants experimental design. The factors were (a) written assign-
ment (summarization vs. argumentation), (b) presentation format (list of topics
vs. list of sources), and (c) and level of prior knowledge (high vs. low). Level
of prior knowledge was based on participants’ scores on the pretest. Participants
with a score of 0 or 1 correct answer and who had never studied social psy-
chology were categorized as possessing low prior knowledge (n D 31, average
age D 27.2 years, S.D. D 6.72), whereas those with a pre-test score of 3 or more
correct answers and who had been studying social psychology for at least two
years were categorized as possessing high prior knowledge (n D 21, average
age D 25.9 years, S.D. D 4.25). Participants who did not fit into either cate-
gory (e.g., with a score of only 2) were excluded from the analysis (n D 13).
Thus, the final sample comprised 52 participants (see Table 3 for distribution
per group).

Measurements
Text and menu reading time. To gauge the amount of time devoted to
knowledge acquisition and construction, the overall reading time was measured
in seconds. This included the time spent reading the texts and menu.

Comprehension questionnaire scores. The proportion of correct an-


swers (out of a maximum possible of four) was computed for each category
of question.
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 457

Essays. The short essays written by participants were used to assess the
level of comprehension achieved, through an analysis of length, connectives,
transformed information, and references to documents.

Words (token). The total number of words was used to assess the partic-
ipants’ verbosity and to calculate the type-token ratios (TTR; Holmes, 1994).
TTR is a statistical measurement of the wealth of vocabulary used in discourse.
It corresponds to the number of different lexical units (i.e., the number of dif-
ferent forms: Type) in the vocabulary contained in the sample divided by the
number of units (i.e., total number of words: Token) forming the sample.

Connectives. Connectives are words or short expressions linking up idea


units within a sentence or between two sentences. The use of connectives in
writing is characteristic of the construction of a situation model (Mac Lure &
Geva, 1983). Moreover, some types of connectives are typical of certain forms
of writing. For example, the use of temporal connectives (e.g., “first”) reflects
a narrative or expository style, whereas the use of adversative connectives (e.g.,
“however”) reflects an argumentative style. All the connectives used in the par-
ticipants’ writings were recorded and assigned to one of five categories: (a) com-
parative/alternative (e.g., “like,” “also”), (b) causal/consequence (e.g., “because,”
“indeed”), (c) adversative/concessive (e.g., “but,” “however”), (d) temporal (e.g.,
“for instance,” “first”), or (e) circumstantial (e.g., “as for,” “concerning”). The
identification and categorization of connectives was performed by two indepen-
dent scorers according to a double blind scoring procedure. The inter-scorer
reliability was assessed using Cohen’s Kappa on a sample of 38 essays, and on
the connectors identified by both scorers. We obtained a Kappa value of .658 (ac-
ceptable level of agreement). All differences were resolved through discussion.
The TTR for each connective category was calculated for each essay.

Idea units: Borrowed, transformed, and added information. In tasks


that require students to write from sources, the amount of transformed informa-
tion is considered a good indicator of in-depth comprehension (Scardamalia &
Bereiter, 1987). The participants’ essays were segmented into idea units, that
is, sentences containing one or more related items of information. Each idea
unit was then categorized as being borrowed, transformed, or added, in line
with the definitions established by Scardamalia and Bereiter. The procedure was
carried out in accordance with Wiley and Voss’s (1996) specification of these
definitions:

Borrowed information was taken directly or paraphrased from the presented mate-
rial. Sentences were marked as transformed when they contained some presented
458 LE BIGOT AND ROUET

information in combination with a new claim or fact or included two bits of pre-
sented information not previously connected. A sentence was coded as added when
it contained only novel information (pp. 68–69).

References to documents. Four types of explicit references to documents


in the essays were considered. Direct references were either general expressions
referring to documents (e.g., “According to the texts : : : ”), or specific author
names (e.g., “Paicheler found that : : : ”). Indirect references were also consid-
ered, that is, verbal expressions that referred to either the studies in general
(e.g., “the studies,” “the experiments”), or to a specific study (e.g., “the blue-
green paradigm”). The number of references of each type was scored in each
essay by two independent judges with an inter-rater agreement of 94%. The
disagreements were resolved through discussion.

Predictions
In accordance with the text comprehension literature, we expected low prior
knowledge participants to spend more time reading the texts and to achieve a
lesser level of comprehension, especially at the macrostructural and inferential
levels. High-knowledge participants were also expected to write essays with a
higher density of connectives and more transformed information, regardless of
task and presentation format.
Consistent with prior studies of multiple document comprehension, we ex-
pected an interaction between presentation format and task on students’ achieve-
ment of text comprehension, and on the type of information included in students’
essays. More specifically, we expected the combination of argument task and
source presentation to improve students’ construction of an integrated situation
model, resulting in more accurate answers to the macrostructure and situation
model questions. The argument task was also expected to enhance students’
transformation of information and addition of extraneous information in their
essays. Arguments would also contain more causal, comparative, and adversative
connectives than summaries, whereas summaries might contain more temporal
and circumstantial connectives. Finally, we expected the source list presenta-
tion to increase the number of explicit references to source texts in students’
essays, reflecting the benefits of highlighting sources on students’ construction
of a documents model (Perfetti et al., 1999; Wiley & Voss, 1999).

RESULTS

Unless otherwise specified, the data were analyzed using an analysis of variance
(GLM procedure). The written assignment (summarization vs. argumentation),
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 459

presentation format (list of topics vs. list of sources), and level of prior knowl-
edge (high vs. low) were treated as between-participants factors. However, the
contrasting procedure made it impossible to balance the groups, and for one
experimental condition there were fewer than five data points. Consequently,
interaction effects involving more than two between-participants factors were
excluded from the analysis. Regarding the statistical analysis of the comprehen-
sion measure, the structural (micro- vs. macrostructural) and representational
(text-base vs. situation model) levels of the questionnaire were added as within-
participants factors. Regarding the writing assignment, connective categories and
idea units were added as within-participants factors.

Text and Menu Reading Time


Because of the global limit on study time (15 minutes), text and menu reading
times were subjected to separate ANOVAs. In order to stabilize variance and
normalize distribution, statistical analyses were performed on logarithmic trans-
formed reading time data. To facilitate interpretation, however, the untransformed
data are presented in Table 4. As expected, text reading times were significantly
longer for low prior knowledge than for high prior knowledge participants, re-
gardless of task or presentation conditions (F Œ1; 44 D 4:81, MSe D :011,
p < :05). Reading times were also higher in the summary than in the argument
task, and in the source than in the topic presentation format, but the difference
did not reach significance. There was also no significant interaction.
An exploratory analysis of menu reading times showed that they were longer
when the participants were required to produce a summary rather than an
argument (83.76 seconds vs. 65.70 seconds, respectively, F Œ1; 44 D 6:26,
MSe D :073, p < :05), suggesting that the summary task may have led the
participants to consider the texts as a whole.

TABLE 4
Text Reading Times (in Seconds) According to Presentation Format,
Written Assignment, and Prior Knowledge
(Standard Deviations in Brackets)

Written Assignment
Presentation Prior
Format Knowledge Argument Summary Total

Topic Low 763 (279) 823 (151) 802 (211)


High 635 (203) 714 (96) 684 (137)
Source Low 843 (134) 881 (45) 858 (107)
High 685 (188) 811 (113) 762 (152)
460 LE BIGOT AND ROUET

Comprehension Questionnaire
Prior knowledge significantly affected students’ scores on the comprehension
post-test. Students with high prior knowledge obtained significantly better scores
than low prior knowledge students (F Œ1; 44 D 4:00, MSe D 1:582, p D
:05). Univariate analyses showed that this effect was most noticeable in the
macrostructure-text-based questions (F Œ1; 44 D 7:45, MSe D 946, p < :01), on
which the high-expertise participants performed best. Moreover, the text-based
questions were easier than the situation model questions (F Œ1; 44 D 43:18,
MSe D :912, p < :0001). The other comparisons were not statistically signifi-
cant (see Table 5).
We did not find the expected effects of task and presentation format on com-
prehension. There was, however, a three-way interaction between task, presenta-
tion format, and the structural level of questions (F Œ1; 44 D 8:50, MSe D :743,
p < :01). Students in the summary task condition answered microstructural
questions more accurately than the macrostructural ones when they experienced
a topic-based presentation, the reverse being true for the source-based presenta-
tion (F Œ1; 44 D 10:77, MSe D :743, p < :01). None of the other comparisons
was statistically significant. In particular, no significant effects were found for
the argumentation assignment (see Figure 1).
Thus, participants with a high level of prior knowledge achieved a higher level
of comprehension at the macrostructural level. Moreover, the summarization
assignment combined with a source-based presentation enhanced macrostructural
comprehension to the detriment of the microstructural level. Conversely, the
topic-based presentation favored microstructural comprehension. The source-
based presentation may have fostered the identification of global connections
across documents as students were constructing an integrated representation of

TABLE 5
Mean Comprehension Scores According to Participants’ Prior Knowledge and
Question Types. The Groups Were Collapsed Across Writing Tasks and
Presentation Format Conditions (Standard Deviations in Brackets)

Structural Level

Microstructure Macrostructure

Representational Level Textbase Situation Textbase Situation

Prior knowledge
Low 2.84 (1.07) 1.42 (1.18) 1.77 (1.02) 1.68 (0.87)
High 2.95 (0.92) 1.50 (0.81) 2.62 (0.87) 2.05 (1.02)
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 461

FIGURE 1 Mean comprehension scores according to writing task, presentation format,


and structural level of questions.

the hypertext (i.e., the intertext model). Neither task nor presentation format had
any effect on inferential versus literal questions.

Analysis of Essays
Writing length. A control analysis of the length of students’ essays found
that they were rather homogeneous across levels of expertise and conditions,
with an average of 119.2 words and a standard deviation of 32.2. The nature
of the writing task, however, did affect students’ writing, with the arguments
being slightly but significantly longer than the summaries (133.5 vs. 106.9,
F Œ1; 44 D 8:02, MSe D 811, p < :01). No other comparison was statistically
significant.

Connectives. The Greenhouse-Geisser epsilon (") correction method was


used to determine F ratios for within-participant effects where appropriate (spher-
icity assumption). In this case, the degrees of freedom were corrected to the
nearest whole values. There was an average of 3.48 (S:D: D 2:11) connec-
tives per text (see Table 6). The analysis revealed a connective category ef-
fect (" D :831, F Œ3; 146 D 3:094, MSe < :001, p < :05). This effect was
462 LE BIGOT AND ROUET

TABLE 6
Mean Number of Connectives According to Participants’ Writing Task
(Standard Deviations in Brackets). The Groups Were Collapsed
Across Prior Knowledge and Presentation Format Conditions

Connectives Causal/Cons. Temporal Compar. Advers. Circums.

Writing task
Argument .0112 .0033 .0089 .0065 .0021
(.0092) (.0066) (.0111) (.0069) (.0050)
Summary .0048 .0076 .0048 .0056 .0024
(.0073) (.0094) (.0065) (.0075) (.0057)

modulated by the two-way connective category  written assignment interac-


tion (" D :831, F Œ3; 146 D 2:848, MSe < :001, p < :05). More specifi-
cally, causal/consequence connectives were used more often in the argumen-
tations than in the summaries (F Œ1; 44 D 6:190, MSe < :001, p < :05)
whereas temporal ones were used marginally more often in the summaries
(F Œ1; 44 D 3:454, MSe < :001, p < :07). None of the other comparisons
was statistically significant.

Type of idea units. There was an average of 5.46 (S:D: D 1:80) idea
units per text (see Figure 2). The analysis revealed an idea unit category effect
(" D :753, F Œ2; 66 D 21:62, MSe D 2:199, p < :0001). This effect was mod-
ulated by the predicted two-way idea unit category  writing task interaction

FIGURE 2 Mean number of idea units according to writing task and idea unit category.
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 463

(" D :753, F Œ2; 66 D 45:79, MSe D 2:199, p < :0001). More specifically, ar-
guments contained more transformed and added information (F Œ1; 44 D 34:15,
MSe D 2:053, p < :0001 and F Œ1; 44 D 22:53, MSe D :643, p < :0001,
respectively), whereas there was more borrowed information in the summaries
(F Œ1; 44 D 40:96, MSe D 2:866, p < :0001). None of the other comparisons
was statistically significant. In particular, we did not find any impact of presen-
tation format or prior knowledge on the type of idea unit included in students’
essays.
In summary, the argument writing task prompted the use of causal/consequence
connectives and the transformation of ideas contained in the texts. The summa-
rization assignment, on the other hand, promoted the use of temporal connectives
and the inclusion of a large number of borrowed ideas in the writings.

References to the documents. Overall students’ essays contained quite a


few explicit references to the documents. There was a total of 105 occurrences,
with 12 direct general references (or 0.2 per essay), 34 direct specific (or 0.7
per essay), 29 indirect general (0.6), and 30 indirect specific (0.6). Twelve es-
says (23%) contained 0 references; 14 (27%) essays contained 1 reference; and
26 (50%) essays contained 2 references or more, with a maximum number of
6 references (1 participant). Preliminary analyses showed that the level of ex-
pertise and the writing instructions had no effect on the number of references.
Therefore, we decided to perform a non-parametric analysis of the influence of
the presentation format on the distribution of total references.
Figure 3 clearly shows an impact of presentation format on the distribution of
references. Forty-two percent of students in the topic list condition never referred
to documents in their essays, whereas 61% of those in the source list condition
included two references or more. The contrast was significant (2 Œ2 D 8:68,
p D :013), indicating that the source list condition increased students’ references
to documents.

DISCUSSION

When studying complex online materials, such as a set of Web sites retrieved
using a search engine, readers need to select, evaluate, and integrate heteroge-
neous, potentially inconsistent sources of information. Thus, students need to
acquire advanced literacy skills if they are to benefit from exposure to mul-
tiple sources (Brem et al., 2001; Rouet et al., 1997; Spivey & King, 1989).
Our purpose was to find out if an emphasis on sources as opposed to content,
combined with an argument-oriented study task (Wiley & Voss, 1999) could
enhance students’ comprehension of multiple documents presented online. We
464 LE BIGOT AND ROUET

FIGURE 3 Number of references to documents per essay as a function of presentation


format (e.g., 42% of participants in the Topic list condition made 0 reference to documents
in their essay).

hypothesized that content representation and tasks would interact with students’
level of background knowledge in the content area.
The study showed a strong effect of the type of writing assignment on several
dependent measures, and a more limited but significant impact of presentation
format. The participants who were assigned an argument task focused on the
contents of the hypertext pages and wrote essays that contained more trans-
formed and added information, and a greater density of causal connectives, as
opposed to more borrowed information with temporal connectives in the sum-
mary condition. Thus, like prior studies of multiple text comprehension (e.g.,
Wiley & Voss, 1999) the argument writing task promoted students’ appropriation
and integration of the materials. There was no direct benefit of the argument
writing task, however, on students’ performance on the comprehension question-
naire. In particular, although the amount of information transformed in a written
essay is considered a sign of deep comprehension (Scardamalia & Bereiter,
1987), this was not reflected in performances on the comprehension question-
naire. One possible explanation is that the questionnaire lacked the sensitivity
required to assess such qualitative shifts in students’ processing of complex text
information.
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 465

The presentation format manipulation affected neither reading times nor


comprehension measures. Under the summary writing condition, however, the
source-based presentation improved macrostructural comprehension, whereas
the topic-based presentation promoted microstructural comprehension. Thus, the
source-based presentation tended to facilitate the identification of macrostruc-
tural links between the documents for all readers, regardless of their level of
prior knowledge. The source-based presentation also affected the structure of
students’ essays, by promoting the inclusion of explicit references (direct or
indirect, general or specific) to documents. The source-based presentation also
helped students establish connections at the intertextual level, a key step toward
the construction of an integrated document model (Perfetti et al., 1999; Spivey
& King, 1989). Whereas prior research on hypertext has focused on the shape
of content representations and navigation tools (e.g., menus vs. concept maps;
Potelle & Rouet, 2003), the present study shows that labeling hypertext links may
influence students’ comprehension of multiple documents. Labels that promote
the identification and comparison of information sources may foster students’
construction of a structured document model (Britt et al., 1996; Perfetti et al.,
1999).
Consistent with prior studies of text comprehension, we also found a sta-
tistically significant main effect of prior knowledge on navigation and compre-
hension measures. Although high prior knowledge participants spent less time
reading the hypertext nodes, they scored better on the comprehension test than
those with low prior knowledge. The analysis showed that these differences
were mainly due to better responses to the macrostructural/text-based questions.
Consistent with prior studies of hypertext comprehension, students’ prior knowl-
edge enabled them to figure out the global semantic/rhetorical structure of the
hypertext as they went through the contents (Hofman & van Oostendorp, 1999;
McDonald & Stevenson, 1998; Rouet, 2003).
The study has some significant implications for the use of Web-based re-
sources in educational contexts. When studying a given content area, students
or users often have to read, assess, and integrate several sources of information
(Britt & Angliskas, 2002). For instance, they have to be capable of assessing
the usefulness of each source (Rouet et al., 1997). They also have to integrate
information from different documents, remembering which document contained
which item of information. These processes may draw on specific cognitive
resources that require a minimum level of skill or domain expertise.
The findings suggest that learning tasks that require students to go through
complex sets of online documents will be better achieved if the students are
provided with specific study objectives. Specific directions might help students
build up consistent navigation strategies (Britt et al., 1996; Dee-Lucas & Larkin,
1995), which are critical when reading hypertext. Unlike short, single texts,
466 LE BIGOT AND ROUET

which are normally read linearly from top to bottom, multiple online documents
require deliberate, purposeful strategies on the part of the reader. Because of
the amount of information available, students need to set up criteria in order
to prioritize their selections and to organize into hierarchy content information.
Search questions, instructional objectives, or writing assignments are useful cues
for students to set up their own study goals (Rouet & Vidal-Abarca, 2002).
In conclusion, our study suggests that the provision of source information in
hypertext content representation has some value in the context of instructional
activities. Highlighting the sources would provide a means of explicitly identi-
fying the origin of a hypertext node, its content, author, and so forth. Source
information would provide a useful cue when building up a global understanding
of multiple document information, or document model (Perfetti et al., 1999).
Several limits of the present study must be pointed out. First, we used a
single set of documents on a single topic. The generality of the result across do-
mains and student populations needs to be further assessed. Second, because we
used a between-participant design, the sample size was relatively small, which
increases the risk of sampling biases and precludes any large-scale generaliza-
tion. Third, the results were not totally consistent across dependent measures.
In particular, some of the benefits observed in the writing task did not extend to
the comprehension measure. This may indicate a lack of sensitivity of multiple
choice comprehension questions at the level of global, strategic comprehension
processes.
Finally, many issues relating to source and content representation in hypertext
remain open for further investigation. How much information should designers
provide about the source and contents of multi-source hypertext nodes? Are the
author name and date of publication sufficient or should other information be
added, such as details about the publishing institution, the intended audience,
and so forth? Would a record of the source features of visited Web sites in
bookmarks be of any benefit (for instance, in a personalized information space)?
And exactly when and how do learners assess a particular source (who produced
this information and why)? To what extent do learners actually encode those
parameters in their long-term memory of the hypertext contents? These complex
and important questions certainly require more research, so that researchers can
develop comprehensive theories of naturalistic text and discourse processing,
and so that educators can make effective use of new information resources in
the classroom.

REFERENCES

Brem, S. K., Russell, J., & Weems, L. (2001). Science on the Web: Students’ evaluation of scientific
arguments. Discourse Processes, 32, 191–213.
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 467

Britt, M. A., & Angliskas, C. (2002). Improving students’ ability to identify and use source infor-
mation. Cognition and Instruction, 20, 485–522.
Britt, M. A., Rouet, J.-F., & Perfetti, C. A. (1996). Using hypertext to study and reason about
historical evidence. In J.-F. Rouet, J. J. Levonen, A. P. Dillon, & R. J. Spiro (Eds.), Hypertext
and cognition (pp. 43–72). Mahwah, NJ: Erlbaum.
Charney, D. (1994). The impact of hypertext on processes of reading and writing. In S. J. Hilligoss
& C. L. Selfe (Eds.), Literacy and computers: The complications of teaching and learning with
technology (pp. 238–263). New York: Modern Language Association.
Chen, S. Y., Fan, J.-P., & Macredie, R. D. (2006). Navigation in hypermedia learning systems:
Experts vs. novices. Computers in Human Behavior, 22, 251–266.
Coiro, J. (2003, February). Reading comprehension on the Internet: Expanding our understanding
of reading comprehension to encompass new literacies [Exploring Literacy on the Internet depart-
ment]. The Reading Teacher, 56(6). Retrieved January 2, 2007, from http://www.readingonline.org/
Dee-Lucas, D., & Larkin, J. H. (1995). Learning from electronic texts: Effects of interactive
overviews for information access. Cognition and Instruction, 13, 431–468.
Dillon, A. (1994). Designing usable electronic text: Ergonomics aspects of human information usage.
London: Taylor and Francis.
Foltz, P. W. (1996). Comprehension, coherence and strategies in hypertext and linear text. In J.-F.
Rouet, R. J. Levonen, A. P. Dillon, & R. J. Spiro (Eds.), Hypertext and cognition (pp. 109–136).
Mahwah, NJ: Erlbaum.
Hofman, R., & van Oostendorp, H. (1999). Cognitive effects of a structural overview in a hypertext.
British Journal of Educational Technology, 30, 129–140.
Holmes, D. I. (1994). Authorship attribution. Computers and the Humanities, 28, 87–106.
Jacobson, M. J., & Spiro, R. J. (1995). Hypertext learning environments, cognitive flexibility, and the
transfer of complex knowledge: An empirical investigation. Journal of Educational Computing
Research, 12, 301–333.
Kintsch, W. (1998). Comprehension: A paradigm for cognition. Cambridge, MA: Cambridge Uni-
versity Press.
Kintsch, W., & Yarbrough, J. C. (1982). Role of rhetorical structure in text comprehension. Journal
of Educational Psychology, 74, 828–834.
Lefavrais, P. (1968). La pipe et le rat [The pipe and the rat]. Paris: Éditions et Applications
Psychologiques.
Leu, D. J., Jr., Kinzer, C. K., Coiro, J., & Cammack, D. W. (2004). Toward a theory of new
literacies emerging from the Internet and other information and communication technologies. In
R. B. Ruddell & N. Unrau (Eds.), Theoretical models and processes of reading (pp. 1570–1613).
Newark, DE: International Reading Association. Retrieved January 2, 2007, from: http://www.
readingonline.org/
Mac Lure, E., & Geva, E. (1983). The development of the cohesive use of adversative conjunctions
in discourse. Discourse Processes, 6, 411–432.
McDonald, S., & Stevenson, J. (1998). Navigation in hyperspace: An evaluation of the effects of
navigational tools and subject matter expertise on browsing and information retrieval in hypertext.
Interacting With Computers, 10, 129–142.
McNamara, D. S., Kintsch, E., Songer, N. S., & Kintsch, W. (1996). Are good texts always better?
Interactions of text coherence, background knowledge, and levels of understanding in learning
from text. Cognition and Instruction, 14, 1–43.
Nilsson, R. M., & Mayer, R. E. (2002). The effects of graphical organizers giving cues to the
structure of a hypertext document on users’ navigation strategies and performance. International
Journal of Human-Computer Studies, 57, 1–26.
468 LE BIGOT AND ROUET

Perfetti, C. A., Rouet, J.-F., & Britt, M. A. (1999). Toward a theory of documents representation. In
H. van Oostendorp & S. R. Goldman (Eds.), The construction of mental representations during
reading (pp. 71–98). Mahwah, NJ: Erlbaum.
Potelle, H., & Rouet, J.-F. (2003). Effects of content representation and readers’ prior knowledge
on the comprehension of hypertext. International Journal of Human-Computer Studies, 58, 327–
345.
Reinking, D. (1998). Synthesizing technological transformations of literacy in a post-typographic
world. In D. Reinking, M. C. McKenna, L. D. Labbo, & R. D. Kieffer (Eds.), Handbook of
literacy and technology: Transformations in a post-typographic world (pp. xi–xxx). Mahwah,
NJ: Erlbaum.
Rieh, S. Y. (2002). Judgments of information quality and cognitive authority in the Web. Journal of
the American Society for Information Science and Technology, 53, 145–161.
Rouet, J.-F. (2003). What was I looking for? The influence of task specificity and prior knowledge
on students’ search strategies in hypertext. Interacting With Computers, 15, 409–428.
Rouet, J.-F. (2006). The skills of document use. From text comprehension to Web-based learning.
Mahwah, NJ: Erlbaum.
Rouet, J.-F., Favart, M., Britt, M. A., & Perfetti, C. A. (1997). Studying and using multiple documents
in history: Effects of discipline expertise. Cognition and Instruction, 15, 85–106.
Rouet, J.-F., & Vidal-Abarca, E. (2002). “Mining for meaning”: A cognitive examination of inserted
questions in learning from scientific text. In J. Otero, J. A. Leon, & A. C. Graesser (Eds.), The
psychology of science text comprehension (pp. 417–436). Mahwah, NJ: Erlbaum.
Scardamalia, M., & Bereiter, C. (1987). Knowledge telling and knowledge transforming in writ-
ten composition. In S. Rosenberg (Ed.), Advances in applied psycholinguistics: Vol.2, Reading,
writing, and language learning (pp. 142–175). Cambridge, England: Cambridge University Press.
Segev-Miller, R. (2004). Writing from sources: The effect of explicit instruction on college students’
processes and products. L1 Educational Studies in Language and Literature, 4, 5–33.
Spivey, N. N., & King, J. R. (1989). Readers as writers composing from sources. Reading Research
Quarterly, 24, 7–26.
Strømsø, H. I., & Bråten, I. (2002). Norwegian students’ use of multiple sources while reading
expository texts. Reading Research Quarterly, 37, 208–227.
van Dijk, T. A., & Kintsch, W. (1983). Strategies of discourse comprehension. Hillsdale, NJ:
Erlbaum.
Wiley, J. (2001). Supporting understanding through task and browser design. In proceedings of the
twenty-third annual conference of the Cognitive Science Society (pp. 1136–1143). Mahwah, NJ:
Erlbaum.
Wiley, J., & Voss, J. F. (1996). The effects of “playing historian” on learning in history. Applied
Cognitive Psychology, 10, S63–S72.
Wiley, J., & Voss, J. F. (1999). Constructing arguments from multiple sources: Tasks that pro-
mote understanding and not just memory for text. Journal of Educational Psychology, 91, 301–
331.
Wineburg, S. (1991). On the reading of historical texts: Notes on the breach between school and
academy. American Educational Research Journal, 28, 495–519.
Winne, P. H., & Hadwin, A. F. (1998). Studying as self-regulated learning. In D. Hacker, J. Dun-
lowsky, & A. Graesser (Eds.), Metacognition in educational theory and practice (pp. 277–304).
Mahwah, NJ: Erlbaum.
Zwaan, R. A., & Singer, M. (2003). Text comprehension. In A. C. Graesser, M. A. Gernsbacher, &
S. R. Goldman (Eds.), Handbook of discourse processes (pp. 83–122). Mahwah, NJ: Erlbaum.
COMPREHENSION OF MULTIPLE ONLINE DOCUMENTS 469

Ludovic Le Bigot is Assistant Professor of Ergonomics and Cognitive Psychology


at the University of Poitiers (France).

Jean-François Rouet is a senior researcher at the Centre National de la Recherche


Scientifique and director of the Language, Memory and Cognitive Development
Laboratory at the University of Poitiers (France).

APPENDIX A: SAMPLE ITEMS FROM THE


COMPREHENSION QUESTIONNAIRE*

Example of a Microstructure—Text-based question

What was the participants’ task in Asch’s (1951) study?

A. Evaluate the color of photographs


B. Evaluate the distance among dots
C. Evaluate the length of segments
D. Evaluate the proximity of individuals
E. Don’t know

Microstructure—Situation model question

What was the total number of segments used in the conformity study?

A. 3 segments
B. 4 segments
C. 5 segments
D. 6 segments
E. Don’t know

Macrostructure—Text-based question

What are the two main types of social influence?

A. Endogenous and exogenous influence


B. Minority and majority influence
C. Normative influence and conformity
D. Perceptual and procedural influence
E. Don’t know
470 LE BIGOT AND ROUET

Macrostructure—Situation model question


What is the temporal course of minority influence?
A. Permanent
B. Immediate
C. Intermittent
D. Latent
E. Don’t know


(Adapted from French. The complete questionnaire in French is available upon
request to the first author.)

View publication stats

You might also like