See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/232440754

The Effects of Readers' Misconceptions on

Comprehension of Scientific Text.

ARTICLE in JOURNAL OF EDUCATIONAL PSYCHOLOGY APRIL 2005

Impact Factor: 3.08 DOI: 10.1037/0022-0663.97.2.235

CITATIONS DOWNLOADS VIEWS

63 18 180

2 AUTHORS, INCLUDING:

Paul van den Broek

Leiden University
104 PUBLICATIONS 2,681 CITATIONS

SEE PROFILE

Available from: Paul van den Broek

Retrieved on: 17 June 2015
Journal of Educational Psychology Copyright 2005 by the American Psychological Association
2005, Vol. 97, No. 2, 235245 0022-0663/05/$12.00 DOI: 10.1037/0022-0663.97.2.235

The Effects of Readers Misconceptions on Comprehension

of Scientific Text
Panayiota Kendeou and Paul van den Broek
University of Minnesota

The aim of the present study was to investigate the effects of readers misconceptions on text compre-
hension. College students with misconceptions in science were asked to read and recall a text that
contradicted their misconceptions. Students with no misconceptions served as the control group. Both
online (think-aloud, reading times) and offline (recall) measures were obtained. The results suggest that
readers misconceptions often do not affect the online processes themselves but do influence the content
of those processes and, consequently, the offline memory representation for the text after reading is
completed.

Keywords: misconceptions, comprehension, scientific text

Much of the learning that takes place in and out of schools is
based on successful comprehension of texts. Readers actively
construct a memory representation of the text that critically de-
pends on their interpretation in light of prior knowledge. The
success of the comprehension process depends on the integration
of readers prior knowledge with textual information (Goldman &
Bisanz, 2002; Kintsch, 1988, 1998; van den Broek, Virtue, Ever-
son, Tzeng, & Sung, 2002). The powerful effects of readers prior
knowledge in text comprehension have been documented early on
(Bartlett, 1932), resulting in a large body of literature indicating
that prior knowledge increases memory of texts for both young and
adult readers (e.g., Chiesi, Spilich, & Voss, 1979; Dochy, Segers,
& Buehl, 1999; Means & Voss, 1985; Recht & Leslie, 1988). The
role of inaccurate prior knowledge, however, has received far less
attention despite the fact that readers with inaccurate knowledge
are the default case rather than the exception (Driver, Squires,
Rushworth, & Wood-Robinson, 1994). Our aim in this article is to
explore the effects of inaccurate prior knowledge on text compre-
hension, focusing both on the final product of reading a text and on
the actual processes that take place during reading and lead to this
product.
To assess the effects of prior knowledge on text comprehension,
one must consider both offline products and online processes (Just
& Carpenter, 1980; Kintsch & van Dijk, 1978; Magliano &
Graesser, 1991; Pressley, 2000). On the one hand, offline investi-
gation provides information about the final mental representation
of the text. Thus, it allows researchers to gauge readers memory
Panayiota Kendeou and Paul van den Broek, Department of Educational
Psychology, University of Minnesota.
This research was supported by the Fulbright Scholarship Program and
the Center for Cognitive Sciences at the University of Minnesota through
National Institute of Child Health and Human Development Grant HD-
07151. We thank Douglas Huffman for his assistance in the development
of the science questionnaire.
Correspondence concerning this article should be addressed to Panay-
iota Kendeou, Department of Educational Psychology, University of Min-
nesota, 206 Burton Hall, 178 Pillsbury Drive SE, Minneapolis, MN 55455.
E-mail: kend0040@umn.edu
and understanding of the text in its entirety, the final product of
reading. On the other hand, online investigation provides informa-
tion about the different types of processes in which readers engage
during actual reading that lead or fail to leadto a coherent and
accurate memory representation, including inferring, elaborating,
explaining, summarizing, paraphrasing, integrating information,
and so on. Insofar as background knowledgeincluding miscon-
ceptionsinfluences readers comprehension, a primary way in
which it does so is by influencing the processes during reading.
Thus, simultaneous consideration of offline and online aspects of
comprehension is important because both aspects are causally
related: online processes that take place during reading lead to the
offline product of reading (e.g., Goldman & Varma, 1995; Kintsch,
1988; Trabasso & Suh, 1993; van den Broek, Young, Tzeng, &
Linderholm, 1999; Zwaan & Singer, 2003).
To understand the online and offline effects of prior knowledge
on text comprehension, one must consider both quantity and qual-
ity of a persons prior knowledge (Dole, 2000; Kendeou, Rapp, &
van den Broek, 2004; Lipson, 1982; Maria & MacGinitie, 1987;
Shapiro, 2004). Quantity refers to the amount of knowledge one
has, whereas quality refers to the accuracy of that knowledge.
Quantity of prior knowledge has been the main focus of research
in which experts and novices in a domain are compared. Experts in
a domain engage in more inference generation and integration of
those inferences with textual information than do novices (Leon &
Perez, 2001; Noordman & Vonk, 1992). Experts also recall more
and are more accurate in their recalls than are novices, provided
that the topic of the text is related to their expertise (Alexander,
Kulikowich, & Schulze, 1994; Royer, Carlo, Dufresne, & Mestre,
1996; Schneider, Koerkel, & Weinert, 1989; Voss & Bisanz,
1985).
Quality of prior knowledge has been investigated specifically
with respect to students inaccurate ideas in science. Inaccurate
ideas, also called misconceptions, have been found to interfere in
the process of acquiring new knowledge (Alvermann et al., 1985;
Diakidoy & Kendeou, 2001; Maria & MacGinitie, 1987). Individ-
uals construct different mental models by which they explain
concepts and processes in the physical world; indeed, they do so
already as young children and before receiving formal instruction

235
236 KENDEOU AND VAN DEN BROEK
on those concepts (Carey, 1985). These mental models usually
contain misconceptions that are far from the scientifically correct
ideas (Diakidoy, Vosniadou, & Hawks, 1997; Nussbaum & No-
vak, 1976; Vosniadou & Brewer, 1992, 1994). Such misconcep-
tions have been found to be very resistant to change and to hinder
learning from texts (Alvermann et al., 1985; Lipson, 1982; Peeck,
van den Bosch, & Kreupeling, 1982; Smith, Readence, & Alver-
mann, 1984).
Although quantity and quality of background knowledge often
go hand in handa person with more knowledge about an area
may also have more accurate knowledge than another personthis
need not always be the case. Even experts in a domain may hold
inaccurate ideas about concepts (Borges, Horizonte, & Gilbert,
1999; Stocklmayer & Treagust, 1996). Although inaccurate ideas
by experts may be quite different from those by novices, the
underlying problems posed by the misunderstandings are similar
(Perkins & Simmons, 1988). Given the fact that inaccuracies in
knowledge are pervasive especially among studentsit is im-
portant to investigate the effects of such inaccuracies on the
process and product of reading.
The effects of readers lack of sufficient and accurate prior
knowledge are moderated by the structure of the text, with some
formats resulting in better comprehension and learning than others.
As a result, various methods for revising texts have been proposed,
ranging from simple single-sentence revisions to complete restruc-
turing of texts by content experts (Beck, McKeown, Sinatra, &
Loxterman, 1991; Britton, Van Dusen, Gulgoz, & Glynn, 1989).
Tests of these methods have shown that students comprehend
better when texts are more coherent and when implicit connections
between text elements are made more explicit (Britton & Gulgoz,
1991; Linderholm et al., 2000). Other methods have focused on
including refutation statements in the text that explicitly contrast
misconceptions and correct notions and explain how potential
misconceptions are less plausible than the correct notions (Alver-
mann & Hague, 1989; Hynd & Guzzetti, 1998). These refutation
texts have been found to enhance students learning in cases in
which students hold misconceptions related to the topic of the texts
(Alvermann & Hynd, 1989; Diakidoy, Kendeou, & Ioannides,
2003; Dole, 2000; Guzzetti, Snyder, Glass, & Gamas, 1993).
The focus on altering texts to improve comprehension and
learning reflects an often implicit shared assumption that improved
comprehension depends on improving readers active, online pro-
cessing of texts. Altered texts would make it more likely that these
online processes succeed, and this would improve comprehension.
Improved processing could pertain to a number of things including
generating explanations (Chi, Bassok, Lewis, Reimann, & Glaser,
1989), generating and answering questions (King, 1994;
Woloshyn, Paivo, & Pressley, 1994), monitoring comprehension
(Baker, 1979; Baker & Anderson, 1982), adopting appropriate
standards of coherence (van den Broek et al., 2002), or generating
knowledge-based inferences (Cote, Goldman, & Saul, 1998;
Graesser & Bertus, 1998). Improved online processing presumably
would result in a more complete and more coherent mental repre-
sentation of the text that efficiently integrates textual information
with prior knowledge (Kintsch, 1988).
In the current studies, we directly investigate the processing of
scientific texts as a function of the quality of readers prior knowl-
edge related to the text. Focusing on this issue is important because
although readers with misconceptions are the default case rather
than the exception (Driver et al., 1994) and although misconcep-
tions have a negative impact on what students remember from the
text (Alvermann et al., 1985; Lipson, 1982), little is known about
exactly how and under what circumstances these misconceptions
affect the comprehension processes as they occur online. Further-
more, although we have strong evidence from research in reading
and science education about the effects of misconceptions on the
offline representation of the text (see Guzzetti et al., 1993, for a
review), we know much less about the processes by which such
representations are constructed during reading. In the present
study, we address this issue by investigating the online and offline
effects of misconceptions on text comprehension. Our investiga-
tion follows the three-pronged method suggested by Magliano and
Graesser (1991) that coordinates predictions based on theories of
discourse processing, evidence from verbal protocols, and evi-
dence from behavioral measures. In Experiment 1, we use a
think-aloud methodology to investigate the cognitive processes
that take place when readers with and without misconceptions read
scientifically correct information presented in a simple science
text. Readers memory for the text is also assessed using free
recall. In Experiment 2, we use a reading-time methodology to
provide converging evidence for and expand on the results ob-
tained in Experiment 1. In both experiments we used a text on
electrical current because students often have misconceptions
about the nature of electrical current and its properties (Borges et
al., 1999; Gentner & Gentner, 1983; Heller & Finley, 1992;
Osborne & Freyberg, 1985; Stocklmayer & Treagust, 1996).
Experiment 1
The purpose of this experiment was to investigate the effects of
misconceptions on the cognitive processes of science text compre-
hension and to determine whether misconceptions are represented
in readers memory for the text. We used a think-aloud method-
ology to investigate the different cognitive processes online be-
cause it allows us to consider a variety of readers responses
(Ericsson & Simon, 1993; Pressley & Afflerbach, 1995; Trabasso
& Suh, 1993). Furthermore, this methodology has received exten-
sive validation as a tool to reveal comprehension processes in
reading (Afflerbach, 2002; Cote & Goldman, 1999; Magliano &
Graesser, 1991; Magliano & Millis, 2003; Zwaan & Brown, 1996).
With respect to the online effects of misconceptions, there are
several possible scenarios for the anticipated results. One scenario
is that readers processing is not affected by their having miscon-
ceptions. This scenario would follow, for example, if readers do
not actually use their background knowledge as they process the
text. We argue that this scenario is an unlikely one, given the
extensive research findings cited earlier in this articlethat dif-
ferences in the amount and accuracy of background knowledge do
affect the product of reading. Moreover, psychological research in
text comprehension has provided considerable evidence that the
readers knowledge has profound effects on the readers online
text processing itself (e.g., inference generation; see Leon & Perez,
2001; Noordman & Vonk, 1992). Thus, it is likely that readers
misconceptions also affect online processing. Such effects could
involve the actual processes or the contents of those processes (or
both). Thus, a second possible scenario is that readers with mis-
conceptions engage in different types of processes than do readers
without misconceptions. A third possible scenario is that readers
misconceptions do not affect the online comprehension processes
 EFFECTS OF READERS MISCONCEPTIONS
per se but the content of these processes when they involve prior
knowledge.
With respect to the offline effects, there are also several possible
scenarios. One scenario is that readers memory is not affected by
their having misconceptions. Again, we argue that this scenario is
an unlikely one, given prior findings that differences in back-
ground knowledge do affect the product of reading (e.g., recall; see
Chiesi et al., 1979; Royer et al., 1996). Also, students with mis-
conceptions do not perform as well as students without miscon-
ceptions on postreading tasks such as application problems,
multiple-choice questions, and so forth (e.g., Maria & MacGinitie,
1987). On the basis of these findings it is likely that readers
misconceptions affect readers offline text representation. Such
effects could involve the quantity of memory of the text (i.e., how
much readers remember), the content of this memory (i.e., how
accurate their memory is), or both. Thus, a second possible sce-
nario is that readers with misconceptions remember less informa-
tion from the text than do readers with no misconceptions. A third
scenario is that misconceptions not only affect the quantity of
memory but also the content of memory. In this case, we would
anticipate readers with misconceptions to remember less informa-
tion from the text and to be less accurate in what they do recall
than readers with no misconceptions.
Given that the online processes form the basis for the construc-
tion of a memory representation (Goldman & Varma, 1995;
Kintsch, 1988; Langston & Trabasso, 1999; Myers & OBrien,
1998; van den Broek et al., 1999; Zwaan, 1999), an important
additional expectation is that the observed patterns of online pro-
cesses will be reflected in the patterns of results for the offline
memory task. Put differently, conclusions about the processes that
take place during reading should be consistent with the observa-
tions concerning the end product of reading and vice versa.
Method
Participants
Sixty-three undergraduate college students enrolled in introductory psy-
chology courses participated in this study. The data of 6 students were
eliminated from the study because the prior-knowledge inclusion criterion
was not satisfied. Of the remaining participants, 34 were women and 23
were men, with an age range of 18 28 years. Students were recruited from
the University of Minnesota and received extra course credit for partici-
pating. The misconception group (as identified using the procedures de-
scribed below) comprised 29 participants. The nonmisconception group
comprised the remaining 28 participants.
Materials
Prior knowledge tests. A science questionnaire and an interview were
used to identify students prior knowledge. A pilot study was conducted to
develop and validate the science questionnaire and the interview questions.
In this study, 24 undergraduate psychology students were administered an
open-ended science questionnaire on electricity. The pilot questionnaire
consisted of 10 questions designed to elicit participants incorrect knowl-
edge related to electricity. Questions were adapted from a longer instru-
ment used previously in applied research (Huffman & Michlin, 2001).
Participants responses to each question were scored as correct or incorrect.
All incorrect responses were analyzed further to identify common incorrect
ideas about electricity. This analysis showed that a large number of
participants had inaccurate knowledge related to the nature of electrical
current and simple electrical circuits. Particularly, 58% of the participants
provided answers that indicated that electrical current is a substance that is
237
stored in a battery. Also, 38% of the participants provided a number of
incorrect explanations about how the bulb in a circuit lights up. In addition,
67% of the participants provided answers that indicated they believed that
electrical current is consumed either in the bulb or in the circuit. These
identified incorrect ideas were also consistent with research that identified
common misconceptions in electricity (Borges et al., 1999; Osborne &
Freyberg, 1985).
The science questionnaire used in the pilot study was revised to include
16 multiple-choice items that were related to the nature of electrical
current, simple electrical circuits, and the fact that electrical current is not
consumed in the bulb or the circuit (see Appendix A for sample questions).
Each question included one correct answer and a number of incorrect
alternatives that corresponded to common student misconceptions (identi-
fied by the pilot study). The reliability of the instrument fell within
moderate range (Cronbachs .61).
The interview consisted of a hands-on activity and three accompanying
questions. The hands-on activity required students to attempt lighting up a
bulb by using a battery, a bulb, and a few wires. The three questions
targeted three common readers misconceptions about electrical current
identified in the pilot study: the nature of electrical current, the processes
by which electrical current lights up a bulb, and the fact that electrical
current is not consumed in the bulb.
Reading comprehension test. The Passage Comprehension and Read-
ing Vocabulary subtests of the Woodcock Language Proficiency Battery
Revised (Woodcock, 1991) were used to assess students reading compre-
hension ability. The average score on the two tests was the students
reading comprehension score.
Texts. Three expository texts were used. Two texts, five sentences long
each and entitled Tornado and TV, respectively, were used for practice.
Both texts were adapted from Millis, Morgan, and Graesser (1990). The
third text was 19 sentences long and entitled How Flashlights Work (see
Appendix B). The text was adapted from Bloomfield (2001). The text
about flashlights contained three target sentences that stated the correct
knowledge about electrical current and simple circuits. Each target sen-
tence related to one of the three common misconceptions about electrical
current. For example, the target sentence Electrical current is not con-
sumed in the bulb related to students misconception that electrical current
is consumed in the bulb.
Procedure
Participants were tested individually in an hour-long session. Partici-
pants first completed the interview and the science questionnaire, followed
by the Woodcock Passage Comprehension test. Participants were then
asked to read the two practice texts. Texts appeared one sentence at a time
on a computer screen. Participants could move from one sentence to the
next with a mouse click. Texts were interrupted at prespecified points with
a blank screen. When the blank screen appeared, students were instructed
to think aloud about the preceding sentence. The experimenter provided
feedback after each text. Participants were instructed to read the third text
on their own and to think aloud whenever the blank screen appeared. The
presentation of the third text was interrupted every three or four sentences,
for a total of five times. Two blank screens were used for control purposes,
whereas the remaining three blank screens followed the three target sen-
tences. Participants responses were recorded. After completing reading,
participants were administered the Woodcock Reading Vocabulary test.
Finally, participants were asked to write down everything they could
remember from the text about flashlights.
Scoring
Prior knowledge tests. The interview questions and the science ques-
tionnaire were scored dichotomously. Each correct response received a
score of 1, whereas each incorrect response received a score of 0. The
observed total scores for the interview questions ranged from 0 to 3
(possible range 0 to 3), whereas the observed total scores for the science
238 KENDEOU AND VAN DEN BROEK

questionnaire ranged from 4 to 16 (possible range 0 to 16). The criterion for
inclusion in the nonmisconception group required participants to score a 3
on the interview questions and an 11 or higher in the science questionnaire.
These participants had none of the three misconceptions related to the
concept of electrical current. The criterion for inclusion in the misconcep-
tion group required participants to score a 0 on the interview questions and
10 or lower on the science questionnaire. These participants had all three
specific misconceptions related to the concept of electrical current. The
data from participants who scored a 1 or a 2 in the interview were
eliminated from the study.
Think-aloud protocols. Students responses during the think-aloud
procedure were transcribed. Each protocol was parsed into clauses. Each
clause was categorized on the basis of two coding schemes. The first
coding scheme was adapted from Pritchard (1990). This coding scheme
consisted of four categories and was used to identify the different processes
readers engaged in during reading. Categories included understanding
(readers expressed their understanding, e.g., I knew that. It makes sense),
uncertainty confusion (readers expressed their confusion, e.g., I am not
sure about that. I am just guessing here), explanations (readers provided
explanations that were based on their prior knowledge, textual information,
or both prior knowledge and textual information, e.g., The electrons
collide with the atoms of the filament. That is actually resistance), and
paraphrases (readers made comments that captured the gist meaning of
text sentences, e.g., The current is not used in the bulb). Responses that
did not fall into any of these categories were coded as other. This last
general category included responses such as repeating the sentence, eval-
uations of the text, or associations that were not relevant to the text.
The second coding scheme was developed to further analyze the content
of explanations. First, each explanation was coded with respect to the role
of readers prior knowledge. Responses that comprised information beyond
the information presented in the text were categorized as knowledge-based
explanations, whereas responses that comprised information only pre-
sented in the text were categorized as text-based explanations. Second, to
investigate possible content effects of prior knowledge, we further coded
knowledge-based explanations as valid or invalid on the basis of their
scientific accuracy.
Two independent coders completed parsing and coding the think-aloud
protocols on the basis of the two coding schemes blind to students group
assignment. Twenty-five percent of the protocols were coded in common.
Agreement was 85%. Disagreements between coders were resolved
through discussion.
Recall protocols. Students written recall protocols were parsed into
clauses. Each clause was matched to the text sentences according to a gist
criterion. Two additional categories were developed for inferences (see van
den Broek, 1994, for a review), statements that were not explicitly men-
tioned in the text but could be inferred on the basis of the text and prior
knowledge: Inferences that were scientifically correct were coded as valid,
whereas inferences that were scientifically incorrect were coded as invalid.
Two independent coders completed parsing and coding the protocols blind
to students group assignment. Twenty-five percent of the protocols were
coded in common. Agreement was 91%. Disagreements between coders
were resolved through discussion.
Types of Comprehension Processes
Participants responses during the think-aloud procedure were
classified with respect to the five selected categories. A multivar-
iate analysis of variance (MANOVA) was conducted, with group
(misconception, nonmisconception) as the independent variable
and frequency for each of the five response categories across
participants as the dependent variables. The means are provided in
Table 1. There were no significant differences between the mis-
conception and the nonmisconception groups in any of the five
categories, F(5, 51) 1.29, p .05 (for the follow-up univariate
tests, all Fs 3, all ps .05).
A second MANOVA was performed, with each of the two
subcategories of explanations, knowledge-based and text-based, as
the dependent variables. The analysis included group (misconcep-
tion, nonmisconception) as the independent variable and the num-
ber of responses per category as the dependent variables. The
means are provided in Table 1. There were no significant differ-
ences between the misconception and the nonmisconception
groups with respect to the frequency of knowledge-based or text-
based explanations, F(2, 54) 0.11, p .05 (for the follow-up
univariate tests, all Fs 1, all ps .05).
Knowledge-based explanations for all participants were further
classified as valid or invalid. A MANOVA was performed, with
group (misconception, nonmisconception) as the independent vari-
able and valid or invalid knowledge-based explanations as the
dependent variables. This analysis revealed a significant effect of
group, F(2, 54) 3.68, p .05, 2 .12. As can be seen in Table
1, readers in the misconception group included fewer valid
knowledge-based explanations than did readers in the nonmiscon-
ception group, F(1, 55) 5.49, p .05, 2 .10. Readers in the
misconception group included more invalid, knowledge-based
explanations than did readers in the nonmisconception group, but
this difference was not statistically significant, F(1, 55) 1.25,
p .05.
Recall of the Text
The responses of the participants in the recall procedure were
coded with respect to three categories (recall score, valid infer-
Table 1
Means of the Types of Processes by Misconception and
Nonmisconception Groups
Misconception Nonmisconception
(N 29) (N 28)
Response category M SD M SD

Uncertaintyconfusion 2.07 1.57 1.36 1.47

Results Explanations 3.31 2.37 3.54 2.96
Text-based 1.24 1.24 1.21 1.25
Preliminary analyses showed that there were no significant Knowledge-based 2.07 1.81 2.32 2.00
Valid 0.58* 0.68 1.32* 1.54
differences in Woodcock reading comprehension score between Invalid 1.48 1.59 1.00 1.65
the misconception (M 532.53, SD 9.50) and nonmisconcep- Understanding 2.38 1.37 2.89 1.66
tion groups (M 537.53, SD 9.55), F(1, 55) 3.36, p .05. Paraphrase 3.62 2.63 3.32 2.01
As expected, the misconception group had a significantly lower Other 1.86 1.50 1.54 1.03
mean science score (M 7.89, SD 1.67) than did the nonmis- Note. Numbers indicate the average frequency for each response cate-
conception group (M 12.00, SD 1.94), F(1, 55) 73.01, p gory.
.01, 2 .57. * p .05.
 EFFECTS OF READERS MISCONCEPTIONS 239

ences, invalid inferences). A MANOVA was performed, with
group (misconception, nonmisconception) as the independent vari-
able and recall score, valid inferences, and invalid inferences as the
dependent variables. This analysis revealed a significant effect of
group, F(3, 53) 14.08, p .01, 2 .44. As can be seen in
Table 2, there were significant differences between the miscon-
ception and the nonmisconception groups with respect to all three
variables. Readers with misconceptions recalled less textual infor-
mation after reading than did readers with no misconceptions, F(1,
55) 24.70, p .01, 2 .31. Readers with misconceptions
generated more invalid inferences, F(1, 55) 23.31, p .01,
2 .30, and fewer valid inferences, F(1, 55) 7.38, p .01,
2 .12, than did readers with no misconceptions.
To rule out the possibility that differences in recall and inference
generation were due to readers individual differences in reading
comprehension ability, we performed a multivariate analysis of
covariance (MANCOVA) with the Woodcock reading comprehen-
sion score as the covariate, group (misconception, nonmisconcep-
tion) as the independent variable, and recall score, number of
invalid inferences, and number of valid inferences as the depen-
dent variables. The data showed that group was a significant
predictor for readers recall, F(1, 54) 20.33, p .01, 2 .27,
invalid inferences, F(1, 54) 18.50, p .01, 2 .25, and valid
inferences, F(1, 54) 4.89, p .05, 2 .08, over and above
readers comprehension ability.
Discussion
The purpose of this experiment was to investigate the effects of
readers misconceptions on the online processes and the offline
end result of reading in science texts. The online results showed
that readers with misconceptions and readers without misconcep-
tions engaged in the same types of processes during reading.
Readers expressed their understanding, uncertainty, or confusion
and explained and paraphrased equally often regardless of whether
they had misconceptions. The content of these processes differed,
however, at points in the text where prior knowledge played a
direct role. At these points, readers in the misconception group
generated significantly fewer valid, knowledge-based explanations
than did readers in the nonmisconception group. Thus, readers
misconceptions affected their attempts to explain information in
the text during reading.
The offline results showed that misconceptions intruded in
readers memory for the text. Readers with misconceptions re-
called less textual information and generated more invalid infer-
ences and fewer valid inferences than did readers with no miscon-
Table 2
Mean Recall by Misconception and Nonmisconception Groups
Misconception Nonmisconception
(N 29) (N 28)
Variable M SD M SD
Recall score 7.72*** 1.84 10.92*** 2.91
Invalid inferences 1.48*** 1.05 0.39*** 0.56
Valid inferences 0.83** 1.10 1.64** 1.16
Note. Numbers indicate the average frequency for each variable.
** p .01. *** p .001.
ceptions. These results provide strong evidence that
misconceptions have an intrusive effect on both the quantity and
the quality of students memory representation of text.
Online and offline results are consistent in that they show
similar effects of readers misconceptions. In the case of
knowledge-based inferences, readers misconceptions affected the
quality of knowledge-based explanations online and the quality of
inferences offline. Indeed, the number of readers invalid,
knowledge-based explanations generated during think-aloud was
significantly related to the number of invalid inferences included
in recall (r .35, p .01). Thus, the quality of individuals
processing of the text was related to the quality of their represen-
tation of the text.
In this experiment, we used a think-aloud methodology because
we wanted to investigate online processing and to obtain a variety
of students responses. Think-aloud procedures provide a rich
source of data, but they also have several limitations. One limita-
tion is that this methodology reveals only a subset of cognitive
processes during reading, namely those of which readers are
aware. However, readers also engage in automatic, subconscious
processes. For instance, when readers encounter information in the
text that is inconsistent with information in the preceding text, they
slow down without necessarily being aware of doing so (Albrecht
& OBrien, 1993; Myers, OBrien, Albrecht, & Mason, 1994). A
second limitation of the think-aloud methodology is that it may
have altered readers natural processing of the text, resulting in
either better or worse processing. Because of these limitations, it is
important to obtain converging evidence by using a less intrusive
and more automatic measure (Magliano & Graesser, 1991). To do
so, in Experiment 2 we obtained reading-time measures for the
target sentences. Reading times at word, sentence, or text levels
have been used extensively in psychological research as an indi-
cator of processing (Lorch & van den Broek, 1997; OBrien, 1995;
Rapp, Gerrig, & Prentice, 2001; Zwaan & Singer, 2003).
Experiment 2
In Experiment 2, we used a reading-time measure to investigate
the effects of misconceptions on the cognitive processes during
comprehension of science texts. Reading-time measures are unob-
trusive and reflect both conscious and subconscious processes. As
in Experiment 1, we also obtained offline measures.
With respect to the online effects of misconceptions, there are
two possible scenarios for the anticipated results. One scenario is
that readers actual processing of the text is affected by their
having misconceptions. This scenario would be indicated if readers
with misconceptions read one or more of the three target sentences
or the entire text at a significantly different speed than do readers
with no misconceptions. A second scenario, in line with the results
obtained from the first experiment, is that readers misconceptions
do not affect the actual processes during reading, although they
may affect the content of those processes. This scenario would be
indicated if we observe no significant differences in reading times
between the two groups of readers.
With respect to the offline effects, on the basis of the findings
obtained in Experiment 1, we expected that readers misconcep-
tions affect readers offline text representation. As in Experiment
1, such effects would involve the quantity of memory of the text
(i.e., how much readers remember) and the content of this memory
(i.e., how accurate is the memory). This expectation assumes that
240 KENDEOU AND VAN DEN BROEK
the think-aloud methodology in the first experiment did not alter
readers natural online processing in a way that affected the offline
text representation. If we observe substantially different results for
the offline task in Experiment 2 than in Experiment 1, this would
indicate that the think-aloud task did alter the reading process and
product.
Method
Participants
Forty-three undergraduate college students enrolled in introductory psy-
chology courses participated in this study. The data of 6 students were
eliminated from the study because the prior knowledge inclusion criterion
was not satisfied. Of the remaining participants, 20 were women and 17
were men, with an age range of 18 26 years old. Students were recruited
from the University of Minnesota and received extra course credit for
participating. The misconception group, as identified using the procedures
described in Experiment 1, comprised 18 participants. The nonmisconcep-
tion group comprised the remaining 19 participants.
Materials
The materials were the same as used in Experiment 1.
Procedure
The procedure was the same as in Experiment 1, except that participants
were instructed to read silently one sentence at a time and the text was not
interrupted. Reading times for the target sentences and the entire text were
collected.
Scoring
Prior knowledge tests. The interview and the science questionnaire
were scored as in Experiment 1. The reliability of the instrument fell within
moderate range (Cronbachs .65).
Reading times. The reading times for the three target sentences were
recorded. Reading times that were more than two standard deviations from
the mean were replaced with two standard deviations. This method re-
placed approximately 4.5% of the data.
Recall protocols. The recall protocols were scored as in Experiment 1.
Two independent coders coded 25% of the protocols in common. Agree-
ment was 93%. Disagreements between coders were resolved through
discussion.
Results
Preliminary analyses showed that there were no significant
differences in Woodcock reading comprehension scores between
the misconception (M 531.36, SD 7.98) and the nonmiscon-
ception groups (M 535.79, SD 7.88), F(1, 35) 2.88, p Table 3
.05. As expected, the misconception group had a significantly
lower mean science score (M 8.11, SD 2.02) than did the
nonmisconception group (M 12.53, SD 1.47), F(1, 35)
58.12, p .01, 2 .62.
Analysis of Reading Times
The reading times for the three target sentences and the total
reading time were entered into a MANOVA. The analysis included
group (misconception, nonmisconception) as the independent vari-
able and response times for the three sentences and the total
reading time as the dependent variables. The data are shown in
Table 3. There were no significant differences between the mis-
conception and the nonmisconception groups with respect to any
of the three sentences, F(3, 33) 0.33, p .05 (for the follow-up
univariate tests, all Fs 1, all ps .05), or the total reading time,
F(1, 35) 0.56, p .05. Thus, there was no evidence that readers
with misconceptions spent more time reading the target sentences
or the entire text than did readers in the nonmisconception group.
Analysis of Recall Protocols
The responses of the participants during the recall procedure
were coded with respect to three categories (recall score, valid
inferences, invalid inferences). A MANOVA was performed, with
group (misconception, nonmisconception) as the independent vari-
able and recall score, valid inferences, and invalid inferences as the
dependent variables. This analysis was significant, F(3, 33)
12.24, p .01, 2 .53. The data in Table 4 show significant
differences between the misconception and the nonmisconception
groups with respect to all three variables. Specifically, readers with
misconceptions recalled less textual information after reading
compared with readers with no misconceptions, F(1, 35) 7.20,
p .01, 2 .17. Readers with misconceptions generated more
invalid inferences, F(1, 35) 14.84, p .01, 2 .30, and fewer
valid inferences, F(1, 35) 18.75, p .01, 2 .35, compared
with readers with no misconceptions.
To rule out the possibility that differences in recall and inference
generation were due to readers individual differences in reading
comprehension ability, we performed a MANCOVA with Wood-
cock reading comprehension score as the covariate, group (mis-
conception, nonmisconception) as the independent variable, and
recall score, number of invalid inferences, and number of valid
inferences as the dependent variables. The data showed that group
was a significant predictor for readers invalid inferences, F(1,
34) 15.08, p .01, 2 .31, and valid inferences, F(1, 34)
19.89, p .01, 2 .37, over and above readers comprehension
ability. Group was only a marginally significant predictor for
readers recall, F(1, 34) 3.99, p .05, 2 .11.
Discussion
The purpose of this experiment was to investigate readers
online processes during the reading of scientific texts, including
processes of which readers are not aware, and to obtain converging
evidence for the recall results in Experiment 1.
The online results demonstrate that readers with misconceptions
did not spend more time reading any of the sentences that explic-
Mean Reading Times of the Target Sentences and Text by
Misconception and Nonmisconception Groups
Misconception Nonmisconception
(N 18) (N 19)
Variable M SD M SD
Sentence 11 4,471 1,627 4,552 1,648
Sentence 14 3,432 981 3,189 1,281
Sentence 18 3,342 892 3,176 1,016
Total reading time 77,958 19,043 74,321 21,115
Note. Numbers indicate the average time in milliseconds.
 EFFECTS OF READERS MISCONCEPTIONS
Table 4
Mean Recall by Misconception and Nonmisconception Groups
Misconception Nonmisconception
(N 18) (N 19)
Variable M SD M SD
Recall score 8.17** 2.54 10.58** 2.89
Invalid inferences 1.83*** 1.20 0.53*** 0.84
Valid inferences 0.17*** 0.38 1.39*** 1.11
Note. Numbers indicate the average frequency for each variable.
** p .01. *** p .001.
itly conflicted with their knowledge or reading the entire text than
did readers with no misconceptions. These results support the
conclusion from Experiment 1 that readers misconceptions do not
result in different types of processes during reading. The offline
results replicated the results from Experiment 1 that readers
misconceptions intruded in their memory of the text. The fact that
the offline results were consistent across the two experiments
indicates that the think-aloud methodology used in Experiment 1
did not alter readers processing.
General Discussion
The aims of the present study were to investigate the online and
offline effects of readers misconceptions on the comprehension of
scientific texts. With regards to the online processes that take place
as the reader proceeds through a text, we demonstrated that readers
with misconceptions engage in the same types of processes during
reading as do readers with no misconceptions but that the content
of those processes differs. During the think-aloud procedure in
Experiment 1, readers expressed their understanding, uncertainty,
and confusion and explained and paraphrased to the same degree
regardless of whether they had misconceptions. At those points in
the text where prior knowledge plays an important role, as is the
case for knowledge-based explanations, both groups generated
equal numbers of inferences, but it is important to note that readers
with misconceptions generated fewer valid explanations than did
readers with no misconceptions. In Experiment 2, we obtained
reading times to verify the conclusion that misconceptions do not
alter the processes themselves. This methodology is less intrusive
than a think-aloud task and, unlike think-aloud procedures, is
sensitive to automatic as well as strategic processes. Here, readers
with misconceptions did not differ from their counterparts without
misconceptions in their reading speed for the text as a whole nor
for the sentences that contained information that directly contra-
dicted their misconceptions. Thus, there is no evidence that readers
altered their processing of the text when their prior knowledge
conflicted with the information in the text.
The online processes in which readers engage as they read a text
form the basis for the offline outcome of their comprehension, so
one would expect that differences in online processes have impli-
cations for the eventual memory representation of the text. Indeed,
the offline results in both experiments indicate that readers mis-
conceptions affect both the quantity and the quality of their mem-
ory for the text. Readers with misconceptions remembered less
information from the text and included more inaccurate and fewer
accurate inferences in their recalls than did readers with no mis-
241
conceptions. Given that the two groups of readers did not differ in
their overall reading comprehension ability, as measured by the
Woodcock reading comprehension test, or in the total number of
inferences they included during recall, these differences appear to
be due directly to the readers prior knowledge.
These results have important implications for our understanding
of the way in which misconceptions influence the comprehension
of scientific texts. There is ample evidence that having miscon-
ceptions vis-a`-vis the content of a text leads to poor understanding
and memory for the text (see Guzzetti et al., 1993, for a review).
Our results allow us to go beyond these findings and consider how
the actual reading processes themselves, as they unfold during
reading of the text, are influenced by a reader having misconcep-
tions and lead to an impoverished mental representation of the text
after reading is completed. Readers with misconceptions proceed
through a text at the same speed and with the same types of
processes as do readers without misconceptions. When the text
calls for it, they activate and integrate their background knowledge
with the textual information as do readers without misconceptions
(see also Hannon & Daneman, 2001; Kintsch, 1988). But the
content of these knowledge-based inferences reflects the miscon-
ceptions, resulting in fewer valid explanations.
Specifically, misconceptions exerted their influence on the ex-
planatory inferences that readers generated during reading. Such
inferences have been found to be crucial in understanding expos-
itory texts (Ohlsson, 2002) and can determine students success in
other tasks such as problem-solving in science (Chi, de Leeuw,
Chiu, & LaVancher, 1994). Indeed, they are deemed central to
successful reading comprehension in current theoretical models.
They provide the glue that provides coherence to a readers mental
representation of a text (e.g., Graesser, Singer, & Trabasso, 1994;
Kintsch, 1988; van den Broek et al., 1999) and lead to a feeling of
satisfaction in the reader that he or she indeed understands a
certain phenomenon or situation (Brewer, Chinn, & Samarapun-
gavan, 2000). The importance of explanatory inferences is also
reflected in the fact that many interventions are aimed at fostering
the skills necessary for and increasing the likelihood that readers
will construct such inferences, for example, using either externally
or internally generated questions (Pressley, 2000; van den Broek,
Tzeng, Risden, Trabasso, & Basche, 2001) or through self-
explanation reading training (McNamara, 2004). Given the impor-
tance of such inferences, it is not surprising that we found that
misconceptions in the end led to deficient mental representations
by the readers. The invalid inferences generated during reading not
only found their way into this mental representation but also led to
generally less consistent and internally coherent representations.
Thus, readers misconceptions appear to have an interference
effect on memory for the text, resulting in both quantitative and
qualitative problems. From the results of our study, we can con-
clude that an important part of this interference originates during
reading, particularly during the generation of knowledge-based
inferences, such as knowledge-based explanations.
Research in reading and science education has shown consis-
tently that readers with misconceptions are relatively unaware of
conflicts between their knowledge and the information in tradi-
tional expository text (Guzzetti et al., 1993). This finding stemmed
primarily from results obtained offline, in which students learning
was ineffective as measured after reading using free recall, appli-
cation questions, or problem solving. We provided additional
evidence that during online reading of traditional expository text
242 KENDEOU AND VAN DEN BROEK
readers do not become aware of their misconceptions. This is quite
startling: Here we have a situation where a reader realizes that a
piece of information in the text needs explanation, recruits his or
her relevant background knowledge, and generates an explanation,
without realizing that the explanation actually contradicts the
to-be-explained text. Had readers detected the inconsistency be-
tween their knowledge and the text, they could have engaged in
deeper processing in an attempt to establish coherence (Glenberg,
Wilkinson, & Epstein, 1983; Graesser et al., 1994; McNamara &
Kintsch, 1996) or reconcile the inconsistent information (e.g.,
Hakala & OBrien, 1995; Linderholm, Virtue, Tzeng, & van den
Broek, 2004). Failure to do so suggests that when average readers
generate explanatory inferences they have a tendency to do so in a
relatively shallow fashion (cf. Elshout-Mohr & van Daalen-
Kapteijns, 2002; Otero, 2002; Rouet & Vidal-Abarca, 2002).
Our findings are consistent with research on knowledge incor-
poration versus compartmentalization. Incorporation is the inte-
gration of new information from the text with the readers existing
body of knowledge, whereas compartmentalization is the isolation
of new information from the text from prior knowledge (Potts &
Peterson, 1985; Potts, Peterson, St. John, & Kirson, 1990; Potts,
St. John, & Kirson, 1989). The evidence from our experiments
suggests that our readers engaged in knowledge compartmental-
ization. Although they activated and used their prior knowledge
when needed during reading (e.g., knowledge-based explanations),
they failed to incorporate textual information into their prior
knowledge.
Such compartmentalization presents an important educational
problem. How do educators foster incorporation rather than com-
partmentalization? An important prerequisite would be that readers
are aware that there are inconsistencies between their prior knowl-
edge and textual information. For example, one could construct
texts in such a way that the inaccuracies in ones background
knowledge and its incompatibility with the textual information are
explicitly mentioned. Indeed, one such type of text, refutation text
(in which readers misconceptions are explicitly mentioned and
refuted and the correct conceptions explained in close textual
proximity), has been found to be more effective in bringing about
conceptual change than is traditional expository text, as measured
by offline tasks such as recall and question answering (e.g., Dole,
2000; Guzzetti et al., 1993). In the context of the present theoret-
ical framework, we can speculate that the benefits of refutation
texts come about because they engage readers in different types of
processes than traditional expository texts do, resulting in greater
likelihood of incorporation instead of compartmentalization and,
hence, in more effective learning.
In this article, we have provided evidence that one source of
individual differences, readers misconceptions, has a profound
effect on science text comprehension. Other individual differ-
ences such as differences in working memory capacity or level
of engagement with the text may also play a role in science text
comprehension and indeed may interact with readers miscon-
ceptions. With respect to working memory capacity, extensive
evidence in the psychological literature has demonstrated that
in cases where processing is demandingfor example, as with
inconsistenciesperformance is much better for readers with
large working memory capacity than for readers with a small
capacity (e.g., Engle & Conway, 1998). These findings raise the
possibility that differences in working memory capacity may
interact with readers misconceptions. Likewise, level of en-
gagement with the text or interest may interact with readers
misconceptions given the relation between prior knowledge and
interest. Readers who have high knowledge in a domain also
tend to express high interest in the domain (Kintsch, 1980).
Other sources of individual differences, such as domain knowl-
edge, may be directly related to readers misconceptions. For
instance, it may be the case that readers with misconceptions
also tend to have a lower level of domain knowledge than do
readers without misconceptions, and conversely readers with a
low level of domain knowledge may tend to have more mis-
conceptions than do readers with a higher level of domain
knowledge. Exploration of such interactions and interdependen-
cies in the future is important because it will reveal how
comprehension processes vary as a function of reader charac-
teristics and text properties.
A thorough understanding of the online and offline effects of
misconceptions on text comprehension is important for the design
of effective texts and instructional techniques to promote students
learning from texts. Further, this understanding is necessary be-
cause readers with misconceptions will always be a reality in and
out of schools, and learning from text will always be a medium of
independent lifelong learning.
References
Afflerbach, P. (2002). Verbal reports and protocol analysis. In M. L.
Kamil, P. B. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Methods of
literacy research: The methodology chapters from the Handbook of
Reading Research, Volume 3 (pp. 87103). Mahwah, NJ: Erlbaum.
Albrecht, J. E., & OBrien, E. J. (1993). Updating a mental model:
Maintaining both local and global coherence. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 19, 10611070.
Alexander, P. A., Kulikowich, J. M., & Schulze, S. K. (1994). How
subject-matter knowledge affects recall and interest. American Educa-
tional Research Journal, 31, 313337.
Alvermann, D. E., & Hague, S. A. (1989). Comprehension of counterin-
tuitive science text: Effects of prior knowledge and text structure.
Journal of Educational Research, 82, 197202.
Alvermann, D. E., & Hynd, C. E. (1989). Effects of prior knowledge
activation modes and text structure on nonscience majors comprehen-
sion of physics. Journal of Educational Research, 83, 97102.
Alvermann, D. E., Smith, L. C., & Readence, J. E. (1985). Prior knowledge
activation and the comprehension of compatible and incompatible text.
Reading Research Quarterly, 20, 420 436.
Baker, L. (1979). Comprehension monitoring: Identifying and coping with
text confusions. Journal of Reading Behavior, 11, 365374.
Baker, L., & Anderson, R. I. (1982). Effects of inconsistent information on
text processing: Evidence from comprehension monitoring. Reading
Research Quarterly, 27, 281294.
Bartlett, F. C. (1932). Remembering: A study in experimental and social
psychology. Cambridge, England: Cambridge University Press.
Beck, I. L., McKeown, M. G., Sinatra, G. M., & Loxterman, J. A. (1991).
Revising social studies text from a text-processing perspective: Evidence
of improved comprehensibility. Reading Research Quarterly, 26, 251
276.
Bloomfield, L. A. (2001). How things work: The physics of everyday life.
New York: Wiley.
Borges, A. T., Horizonte, B., & Gilbert, J. K. (1999). Mental models of
electricity. International Journal of Science Education, 21, 95117.
Brewer, W. F., Chinn, C. A., & Samarapungavan, A. (2000). Explanations
in scientists and children. In F. C. Keil & R. A. Wilson (Eds.), Expla-
nation and cognition (pp. 279 298). Cambridge, MA: MIT Press.
Britton, B. K., & Gulgoz, S. (1991). Using Kintschs computational model
 EFFECTS OF READERS MISCONCEPTIONS
to improve instructional texts: Effects of repairing inference calls on
recall and cognitive structures. Journal of Educational Psychology, 83,
329 345.
Britton, B. K., Van Dusen, L., Gulgoz, S., & Glynn, S. M. (1989).
Instructional texts rewritten by five expert teams: Revisions and reten-
tion improvements. Journal of Educational Psychology, 81, 226 239.
Carey, S. (1985). Conceptual change in childhood. Cambridge, MA:
Bradford.
Chi, M. T. H., Bassok, M., Lewis, M. W., Reimann, P., & Glaser, R.
(1989). Self-explanations: How students study and use examples in
learning to solve problems. Cognitive Science, 13, 145182.
Chi, M. T. H., de Leeuw, N., Chiu, M. H., & LaVancher, C. (1994).
Eliciting self-explanations improves understanding. Cognitive Science,
18, 439 477.
Chiesi, H., Spilich, G., & Voss, J. (1979). Acquisition of domain-related
information in relation to high and low domain information. Journal of
Verbal Learning and Verbal Behavior, 18, 257273.
Cote, N., & Goldman, S. R. (1999). Building representations of informa-
tional text: Evidence from childrens think-aloud protocols. In H. van
Oostendorp & S. R. Goldman (Eds.), The construction of mental repre-
sentations during reading (pp. 169 193). Mahwah, NJ: Erlbaum.
Cote, N., Goldman, S. R., & Saul, E. U. (1998). Students making sense of
informational text: Relations between processing and representation.
Discourse Processes, 25, 153.
Diakidoy, I. N., & Kendeou, P. (2001). Facilitating conceptual change in
astronomy: A comparison of the effectiveness of two instructional
approaches. Learning and Instruction, 11, 120.
Diakidoy, I. N., Kendeou, P., & Ioannides, C. (2003). Reading about
energy: The effects of text structure in science learning and conceptual
change. Contemporary Educational Psychology, 28, 335356.
Diakidoy, I. N., Vosniadou, S., & Hawks, J. D. (1997). Conceptual change
in astronomy: Models of the earth and the day/night cycle in American-
Indian children. European Journal of Psychology of Education, 12,
159 184.
Dochy, F., Segers, M., & Buehl, M. M. (1999). The relation between
assessment practices and outcomes of studies: The case of research on
prior knowledge. Review of Educational Research, 69, 145186.
Dole, J. A. (2000). Readers, texts and conceptual change learning. Reading
and Writing Quarterly, 16, 99 118.
Driver, R., Squires, A., Rushworth, P., & Wood-Robinson, V. (1994).
Making sense of secondary science. New York: Routledge.
Elshout-Mohr, M., & van Daalen-Kapteijns, M. (2002). Situated regulation
of scientific text processing. In J. Otero, J. A. Leon, & A. C. Graesser
(Eds.), The psychology of science text comprehension (pp. 223252).
Mahwah, NJ: Erlbaum.
Engle, R. W., & Conway, A. R. A. (1998). Working memory and com-
prehension. In R. H. Logie & K. J. Gilhooly (Eds.), Working memory
and thinking (pp. 6791). East Sussex, England: Psychology Press.
Ericsson, K. A., & Simon, H. A. (1993). Protocol analysis: Verbal reports
as data. Cambridge, MA: MIT Press.
Gentner, D., & Gentner, D. R. (1983). Flowing waters or teeming crowds:
Mental models of electricity. In D. Gentner & A. L. Steves (Eds.),
Mental models (pp. 99 129). Mahwah, NJ: Erlbaum.
Glenberg, A. M., Wilkinson, A. C., & Epstein, W. (1983). The illusion of
knowing: Failure in a self-assessment of comprehension. Memory &
Cognition, 10, 597 602.
Goldman, S. R., & Bisanz, G. L. (2002). Toward a functional analysis of
scientific genres: Implications for understanding and learning processes.
In J. Otero, J. A. Leon, & A. C. Graesser (Eds.), The psychology of
science text comprehension (pp. 19 50). Mahwah, NJ: Erlbaum.
Goldman, S. R., & Varma, S. (1995). CAPing the construction-integration
model of discourse comprehension. In C. A. Weaver, S. Mannes, &
C. R. Fletcher (Eds.), Discourse comprehension: Essays in honor of
Walter Kinstch (pp. 337358). Hillsdale, NJ: Erlbaum.
Graesser, A. C., & Bertus, E. L. (1998). The construction of causal
243
inferences while reading expository texts on science and technology.
Scientific Studies of Reading, 2, 247269.
Graesser, A. C., Singer, M., & Trabasso, T. (1994). Constructing infer-
ences during narrative comprehension. Psychological Review, 101, 371
395.
Guzzetti, B. J., Snyder, T. E., Glass, G. V., & Gamas, W. S. (1993).
Promoting conceptual change in science: A comparative meta-analysis
of instructional interventions from reading education and science edu-
cation. Reading Research Quarterly, 28, 117159.
Hakala, C. M., & OBrien, E. J. (1995). Strategies for resolving coherence
breaks in reading. Discourse Processes, 20, 167185.
Hannon, B., & Daneman, M. (2001). A new tool for measuring and
understanding individual differences in the component processes of
reading comprehension. Journal of Educational Psychology, 93, 103
128.
Heller, P. M., & Finley, F. N. (1992). Variable uses of alternative concep-
tions: A case study in current electricity. Journal of Research in Science
Teaching, 29, 259 275.
Huffman, D., & Michlin, M. (2001). The impact of CPU on high school
physics: Current electricity results. Minneapolis: University of Minne-
sota, Center for Applied Research and Educational Improvement.
Hynd, C., & Guzzetti, B. J. (1998). When knowledge contradicts intuition:
Conceptual change. In C. Hynd (Ed.), Learning from text across con-
ceptual domains (pp. 139 164). Mahwah, NJ: Erlbaum.
Just, M. A., & Carpenter, P. A. (1980). A theory of reading: From eye
fixations to comprehension. Psychological Review, 87, 329 354.
Kendeou, P., Rapp, D. N., & van den Broek, P. (2004). The influence of
readers prior knowledge on text comprehension and learning from text.
In R. Nata (Ed.), Progress in education (Vol. 13, pp. 189 209). Haup-
pauge, NY: Nova Science.
King, A. (1994). Guiding knowledge construction in the classroom: Effects
of teaching children how to question and how to explain. American
Educational Research Journal, 31, 338 368.
Kintsch, W. (1980). Learning from texts, levels of comprehension, or: Why
anyone would read a story anyway. Poetics, 9, 8798.
Kintsch, W. (1988). The role of knowledge in discourse comprehension: A
construction integration model. Psychological Review, 95, 163182.
Kintsch, W. (1998). Comprehension: A paradigm for cognition. Cam-
bridge, England: Cambridge University Press.
Kintsch, W., & van Dijk, T. A. (1978). Toward a model of text compre-
hension and production. Psychological Review, 85, 363394.
Langston, M. C., & Trabasso, T. (1999). Modeling causal integration and
availability of information during comprehension of narrative texts. In
H. van Oostendorp & S. R. Goldman (Eds.), The construction of mental
representations during reading (pp. 29 69). Mahwah, NJ: Erlbaum.
Leon, J., & Perez, O. (2001). The influence of prior knowledge on the time
course of clinical diagnosis inferences: A comparison of experts and
novices. Discourse Processes, 31, 187213.
Linderholm, T., Gaddy, M., van den Broek, P., Mischinski, M., Crittenden,
A., & Samuels, S. J. (2000). Effects of causal text revisions on more- and
less-skilled readers comprehension of easy and difficult texts. Cognition
and Instruction, 18, 525556.
Linderholm, T., Virtue, S., Tzeng, Y., & van den Broek, P. (2004).
Fluctuations in the availability of information during reading: Capturing
cognitive processes using the landscape model. Discourse Processes, 37,
165186.
Lipson, M. (1982). Learning new information from text: The role of prior
knowledge and reading ability. Journal of Reading Behavior, 14, 243
261.
Lorch, R. E. J., & van den Broek, P. (1997). Understanding reading
comprehension: Current and future contributions to cognitive science.
Contemporary Educational Psychology, 22, 213246.
Magliano, J. P., & Graesser, A. C. (1991). A three-pronged method for
studying inference generation in literary text. Poetics, 20, 193232.
Magliano, J. P., & Millis, K. K. (2003). Assessing reading skill with a
244 KENDEOU AND VAN DEN BROEK
think-aloud procedure and latent semantic analysis. Cognition and In-
struction, 21, 251283.
Maria, K., & MacGinitie, W. (1987). Learning from texts that refute the
readers prior knowledge. Reading Research and Instruction, 26, 222238.
McNamara, D. S. (2004). SERT: Self-Explanation Reading Training. Dis-
course Processes, 38, 130.
McNamara, D. S., & Kintsch, W. (1996). Learning from texts: Effects of
prior knowledge and text coherence. Discourse Processes, 22, 247288.
Means, M. L., & Voss, J. F. (1985). Star wars: A developmental study of
expert novice structures. Journal of Memory and Language, 24, 746 757.
Millis, K. K., Morgan, D., & Graesser, A. C. (1990). The influence of
knowledge-based inferences on the reading time of expository text. In
A. C. Graesser & G. H. Bower (Eds.), The psychology of learning and
motivation (Vol. 25, pp. 197212). New York: Academic Press.
Myers, J. L., & OBrien, E. J. (1998). Accessing the discourse represen-
tation during reading. Discourse Processes, 26, 131157.
Myers, J. L., OBrien, E. J., Albrecht, J. E., & Mason, R. A. (1994).
Maintaining global coherence during reading. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 20, 876 886.
Noordman, L. G., & Vonk, W. (1992). Readers knowledge and control of
inferences in reading. Language and Cognitive Processes, 7, 373391.
Nussbaum, J., & Novak, J. D. (1976). An assessment of childrens concepts
of the earth utilizing structured interviews. Science Education, 60, 535
550.
OBrien, E. J. (1995). Automatic components of discourse comprehension.
In E. P. Lorch & E. J. OBrien (Eds.), Sources of coherence in reading
(pp. 159 176). Hillsdale, NJ: Erlbaum.
Ohlsson, S. (2002). Generating and understanding qualitative explanations.
In J. Otero, J. A. Leon, & A. C. Graesser (Eds.), The psychology of
science text comprehension (pp. 91128). Mahwah, NJ: Erlbaum.
Osborne, R., & Freyberg, P. (1985). Learning in science: The implications
of childrens science. Portsmouth, NH: Heinemann.
Otero, J. (2002). Noticing and fixing difficulties while understanding
science texts. In J. Otero, J. A. Leon, & A. C. Graesser (Eds.), The
psychology of science text comprehension (pp. 281307). Mahwah, NJ:
Erlbaum.
Peeck, J., van den Bosch, A. B., & Kreupeling, W. J. (1982). Effect of
mobilizing prior knowledge on learning from text. Journal of Educa-
tional Psychology, 74, 771777.
Perkins, D. N., & Simmons, R. (1988). An integrative model of miscon-
ceptions. Review of Educational Research, 58, 303326.
Potts, G. R., & Peterson, S. B. (1985). Incorporation versus compartmen-
talization in memory for discourse. Journal of Memory and Language,
24, 107118.
Potts, G. R., Peterson, S. B., St. John, M. F., & Kirson, D. (1990).
Independent access to world knowledge and newly learned facts. In
A. C. Graesser & G. H. Bower (Eds.), The psychology of learning and
motivation (Vol. 25, pp. 213232). New York: Academic Press.
Potts, G. R., St. John, M. F., & Kirson, D. (1989). Incorporating new
information into existing world knowledge. Cognitive Psychology, 21,
303333.
Pressley, M. (2000). What should comprehension instruction be the in-
struction of? In M. L. Kamil, P. B. Mosenthal, P. D. Pearson, & R. Barr
(Eds.), Handbook of reading research (Vol. 3, pp. 545586). Mahwah,
NJ: Erlbaum.
Pressley, M., & Afflerbach, P. (1995). Verbal protocols of reading: The
nature of constructively responsive reading. Hillsdale, NJ: Erlbaum.
Pritchard, R. (1990). The effects of cultural schemata on reading process-
ing strategies. Reading Research Quarterly, 24, 273295.
Rapp, D. N., Gerrig, R. J., & Prentice, D. A. (2001). Readers trait-based
models of characters in narrative comprehension. Journal of Memory
and Language, 45, 737750.
Recht, D. R., & Leslie, L. (1988). Effect of prior knowledge on good and
poor readers memory of text. Journal of Educational Psychology, 80,
16 20.
Rouet, J. F., & Vidal-Abarca, E. (2002). Mining for meaning: Cognitive
effects 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.
Royer, J. M., Carlo, M. S., Dufresne, R., & Mestre, J. (1996). The
assessment of levels of domain expertise while reading. Cognition and
Instruction, 14, 373 408.
Schneider, W., Koerkel, J., & Weinert, F. (1989). Domain-specific knowl-
edge and memory performance: A comparison of high- and low-aptitude
children. Journal of Educational Psychology, 81, 306 312.
Shapiro, A. M. (2004). How including prior knowledge as a subject
variable may change outcomes of learning research. American Educa-
tional Research Journal, 41, 159 189.
Smith, L. C., Readence, J. E., & Alvermann, D. E. (1984). Effects of
activating prior knowledge on retention of expository text. In J. A. Niles
& L. A. Harris (Eds.), Thirty-third yearbook of the National Reading
Conference (pp. 188 192). Rochester, NY: National Reading
Conference.
Stocklmayer, S. M., & Treagust, D. F. (1996). Images of electricity: How
do novices and experts model electric current? International Journal of
Science Education, 18, 163178.
Trabasso, T., & Suh, S. (1993). Understanding text: Achieving explanatory
coherence through on-line inferences and mental operations in working
memory. Discourse Processes, 16, 334.
van den Broek, P. (1994). Comprehension and memory of narrative texts.
In M. A. Gernsbacher (Ed.), Handbook of psycholinguistics (pp. 539
588). San Diego, CA: Academic Press.
van den Broek, P., Tzeng, Y., Risden, K., Trabasso, T., & Basche, P.
(2001). Inferential questioning: Effects on comprehension of narrative
texts as a function of grade and timing. Journal of Educational Psychol-
ogy, 93, 521529.
van den Broek, P., Virtue, S., Everson, M., Tzeng, Y., & Sung, Y. (2002).
Comprehension and memory of science texts: Inferential processes and
the construction of a mental representation. In J. Otero, J. A. Leon, &
A. C. Graesser (Eds.), The psychology of science text comprehension
(pp. 131154). Mahwah, NJ: Erlbaum.
van den Broek, P., Young, M., Tzeng, Y., & Linderholm, T. (1999). The
landscape model of reading: Inferences and on-line construction of a
memory representation. In H. van Oosterndorp & S. R. Goldman (Eds.),
The construction of mental representations during reading (pp. 7198).
Mahwah, NJ: Erlbaum.
Vosniadou, S., & Brewer, W. F. (1992). Mental models of the earth: A
study of conceptual change in childhood. Cognitive Psychology, 24,
535585.
Vosniadou, S., & Brewer, W. F. (1994). Mental models of the day/night
cycle. Cognitive Science, 18, 123183.
Voss, J. F., & Bisanz, G. L. (1985). Knowledge and the processing of
narrative and expository texts. In B. K. Britton & J. B. Black (Eds.),
Understanding expository texts (pp. 173198). Hillsdale, NJ: Erlbaum.
Woloshyn, V. E., Paivio, A., & Pressley, M. (1994). Use of elaborative
interrogation to help students acquire information consistent with prior
knowledge and information inconsistent with prior knowledge. Journal
of Educational Psychology, 86, 79 89.
Woodcock, R. W. (1991). Woodcock Language Proficiency Battery
Revised. Chicago: Riverside.
Zwaan, R. A. (1999). Five dimensions of narrative comprehension: The
events indexing model. In S. R. Goldman, A. C. Graesser, & P. van den
Broek (Eds.), Narrative comprehension, causality, and coherence (pp.
93110). Mahwah, NJ: Erlbaum.
Zwaan, R. A., & Brown, C. M. (1996). The influence of language profi-
ciency and comprehension skill on situation-model construction. Dis-
course Processes, 21, 289 328.
Zwaan, R. A., & Singer, M. (2003). Text comprehension. In A. C.
Graesser, M. A. Gernsbacher, & S. R. Goldman (Eds.), Handbook of
discourse processes (pp. 83121). Hillsdale, NJ: Erlbaum.
 EFFECTS OF READERS MISCONCEPTIONS 245

Appendix A

Sample Questions of the Science Questionnaire

1. Electrical current is: a) The wire that enters the radio.

a) The flow of chemical energy b) The wire that leaves the radio.

b) A substance stored in the battery c) Both wires have the same amount of electrical current.
c) The flow of electrons
3. Is there electrical current in the automobiles battery when the vehicle
d) The path of electricity is moving?

2. A radio is powered by a battery. Electrical current enters the radio a) Yes

flowing through one wire and leaves the radio through another. Which
wire has more electrical current? b) No

Appendix B

Expository Text: How Flashlights Work

1. There isnt much to a typical flashlight: 15. When the switch of the flashlight is turned off, it breaks the
circuit,
2. You flip a switch and out comes a beam of light.
16. and the electrical current is prevented from flowing through the
3. To understand how flashlights works, circuit.

4. you must understand how an electrical circuit works. 17. Without electrical current flowing the flashlight becomes dark.

5. The circuit of the flashlight contains one or more batteries, a 18. The electrical current is not consumed in the bulb.B3
light bulb, a switch,
19. What is consumed in the bulb is the chemical energy of the
6. and several metal strips that connect the components in a battery.
continuous loop.

7. When the switch of the flashlight is turned on,

Note. From How Things Work: The Physics of Everyday Life (p. 411), by
8. the chemical energy stored in the battery is transformed into
electrical energy. L. A. Bloomfield, 2001, New York: Wiley. Copyright 2001 by John Wiley
& Sons. Adapted with permission.
9. The electrical energy makes all free electrons move

10. from the negative terminal of the battery to the positive

B1
terminal. The fact that electrical current is the directed flow of free electrons
relates to readers preconception that electrical current is a substance stored
11. That directed flow of electrons is the electrical current.B1 in the battery.
B2
The fact that the collision of the electrons and the atoms in the
12. As electrical current flows through the bulbs filament,
filament causes the bulb to light relates to readers preconception that the
13. the electrons collide with the atoms of the filament. light in the bulb is the product of a chemical reaction.
B3
The fact that electrical current is not consumed in the bulb relates to
14. Thats why the filament becomes very hot and emits light.B2 readers preconception that electrical current is consumed in the bulb.

Received March 11, 2004

Revision received August 5, 2004
Accepted October 21, 2004