You are on page 1of 25

This article was downloaded by: [Harokopio University]

On: 22 April 2012, At: 14:07

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Educational Psychologist
Publication details, including instructions for authors and subscription information:

Children's Metacognition About Reading: issues in

Definition, Measurement, and Instruction
Janis E. Jacobs & Scott G. Paris

Available online: 22 Jun 2011

To cite this article: Janis E. Jacobs & Scott G. Paris (1987): Children's Metacognition About Reading: issues in Definition,
Measurement, and Instruction, Educational Psychologist, 22:3-4, 255-278

To link to this article:


Full terms and conditions of use:

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to
anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should
be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,
proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in
connection with or arising out of the use of this material.
Copyright o 1987, Lawrence Erlbaum Associates, Inc.

Children's Metacognition About Reading:

Issues in Definition, Measurement, and
Janis E. Jacobs
University of Nebraska
Scott G. Paris
University of Michigan
Downloaded by [Harokopio University] at 14:07 22 April 2012

Metacognition has become a popular term in theories of cognitivedevelopment

and reading. 'What children know about the goals, tasks, and strategies of read-
ing can influence how well they plan and monitor their own reading. Despite the
appeal of metacognition and the emphasis on executive management of read-
ing, there have been relatively few empirical studies that measure or promote
children's metacognition about reading. This article briefly reviews some key
studies of children's knowledge about reading and identifies some difficulties in
defining and measuring metacognition. An Index of Reading Awareness (IRA)
is offered as an informal assessment of metacognition derived from both theory
and empirical data that can be used to measure children's understanding of
reading comprehension processes. Data are presented illustrating that the in-
strument is slensitive to developmental and instructional difference~in chil-
dren's metacognition about reading. It is argued that the usefulness of the term
metacognition depends on the development of both formal and informal tests
of children's knowledge about reading and effective instruction that promotes

Metacognition refers to thinking about thinking. It is a global construct that

has not been defined precisely nor measured frequently. Yet it is an appealing
term for both itlheoretical and practical reasons. Metacognition focuses on
self-regulated thinking- what people know and how they apply that knowl-
edge to particular tasks. The emphasis on executive and cognitive manage-
ment of reading has bolstered theories of comprehension by infusing them

Requests for reprints should be sent to Janis E. Jacobs, 203 Burnett, University of Nebraska,
Lincoln, NE 68588-0308.
with strategic reading and higher order thinking. Metacognition has also pro-
vided some new avenues of instruction for teachers by focusing on the pro-
cesses rather than the products of reading.
Metacognition has been a popular term in research on reading in the past
10 years because it emphasizes how readers plan, monitor, and repair their
own comprehension. Many articles written for teacher educators and curric-
ulum supervisors have enthusiastically endorsed metacognition as a key in-
gredient for cognitive instruction that leads to independentlearning (Babbs &
Moe, 1983; Costa, 1986; Loper, 1980; Spring, 1985; Tei & Stewart, 1985).
The enthusiastic acceptance of the importance of metacognition can be
traced to several factors. First, metacognition emphasizes active participa-
tion by the reader in task analysis and strategic reading. Research during the
1970s indicated that poor readers seldom use effective strategies to aid com-
prehension (Baker & Brown, 1984; Ryan, 1981). Skilled readers use a variety
of tactics and often predict what happens next in a story, look forward and
backward in the passage, and check their own understanding as they read.
Downloaded by [Harokopio University] at 14:07 22 April 2012

The independent use of appropriate strategies discriminates skilled from less

skilled readers. Second, interviews with young children indicated how little
they understood about reading. Beginning readers are often confused about
the directionality of print and whether one reads the pictures or the words.
Older children are often confused about what to do when they confront an
unknown word, why skimming text can be useful, or how rereading might fa-
cilitate understanding. NaivetC and misconceptions about reading are preva-
lent among unskilled readers. The third reason that metacognition has be-
come popular is that it offers a tangible alternative to traditional instruction.
Educators who are unsatisfied with traditional basal readers because they
provide so little guidance on comprehension strategies (Durkin, 1981) have
used the term metacognition to create instruction that fosters thinking strate-
gies before, during, and after reading. Thus, theoretical and practical inter-
ests in reading strategies, knowledge, and instructional innovations have
converged on metacognition because the concept provides tools for thinking
and tools for teaching.
Enthusiasm for the term metacognition is not universal though. Some
reading educators believe that an emphasis on reflection, planning, and cog-
nitive strategies will lead to excessive drills on cognitive skills removed from
the context of reading. They worry that children will be taught about reading
rather than taught how to read with fluency and enjoyment. Other critics
point out that the value of metacognition has yet to be demonstrated; it is
only a theoretical promise at this point. They argue that metacognition has
been used in many ways to refer to a wide variety of reading behaviors and
has not been measured in any consistent fashion. Furthermore, some instruc-
tion, such as reciprocal teaching (Palincsar & Brown, 1984), has been labeled
as a "metacognitive curriculum" but has failed to provide direct measures of
students' metacognition before or after the intervention. Only a few studies
have shown that instructing students about metacognitive aspects of reading
facilitates their comprehension quickly or significantly.
We believe the enthusiasm and criticism generated by the construct of
metacognition are both well deserved. Theories of reading have neglected
cognitive aspects of strategies and self-regulated learning for too long, and
metacognition offers some promise as a remedy. But the term may have been
embraced and accepted prematurely. If metacognition is to be useful for re-
searchers and teachers alike, it must be studied more extensively and in
greater detail. In particular, more attention needs to be paid to the definition
of metacognition and the creation of appropriate measures of metacog-
nition. Of course, the ultimate test of the value of the construct is whether or
not it helps students read better. Therefore, experimental studies that manip-
ulate metacogniition directly are needed. In the following sections we briefly
discuss the pro~blemsof definition, measurement, andl instruction and then
describe a study that attempts to improve the construct of metacognition in
Downloaded by [Harokopio University] at 14:07 22 April 2012

all three areas.


Wellman (1985) defined metacognition as a "fuzzy concept" that cannot be

pinned down or demarcated easily. Like others before him, Wellman es-
chewed conceptual or operational definitions of metalcognition and instead
relied on behavioral examples. Flavell and Wellman (1977) characterized
children's metarmemory as knowledge about task variables, person variables,
and strategy variables that influence remembering. In a subsequent article,
Flavell(1979) expanded the concept of metacognition to include the individu-
al's sensitivity to the need to use metacognition, the individual's "metacog-
nitive experiences" or affective reactions, and the ways in which people use
metacognition to orchestrate their thinking. Thus, knowledge, motivation,
and affect are all involved in metacognition according to Flavell. Ann Brown
and her colleagues chose to emphasize executive strategies such as planning,
monitoring, revising, and repairing comprehension (see:Brown, Armbruster,
&Baker, 1984). Brown, Bransford, Ferrara, and Campione (1983) traced the
historical and theoretical interest in metacognition and distinguished meta-
cognition that is used to regulate one's own learning from metacognition that
is used in instructional settings to supervise someone else's thinking (i.e., self-
regulation vs. ather regulation). But Brown et al. (1983) described meta-
cognition as generally intentional and "cold" in contrast to Flavell who sug-
gested that metacognition can oftentimes be unconscious and laden with

Basic characteristics of metacognition have defied consensus. Some re-

searchers argue that metacognition refers to conscious knowledge and delib-
erate actions, whereas others suggest that metacognition can be tacit and au-
tomatic. Some argue that metacognition involves emotion and motivation,
whereas others suggest that it is better conceptualized as knowledge without
affect. There is little agreement about appropriate measures for meta-
cognition, although most agree that verbal reports and interviews can lead to
distorted and unreliable estimates of what people know about their own
thinking. Researchers have generally circumvented the problem of defining
metacognition by referring to two broad classes of metacognition: (a) knowl-
edge that one has about a cognitive domain (e.g., reading, memory, or learn-
ing) and (b) executive strategies that regulate thinking (e.g., planning and
monitoring; see Lawson, 1985). It is surprising that more detailed definitions
of metacognition have not appeared in the literature.
Our own definitions of metacognition have emerged in the course of our
research on children's reading and learning and have been described in sev-
Downloaded by [Harokopio University] at 14:07 22 April 2012

eral sources (Paris & Jacobs, 1984; Paris, Jacobs, & Cross, 1987; Paris,
Lipson, & Wixson, 1983). These views are summarized briefly because they
guided the creation of our measures of metacognition as well as our instruc-
tional interventions. First, we define metacognition as any knowledge about
cognitive states or processes that can be shared between individuals. That is,
knowledge about cognition can be demonstrated, communicated, examined,
and discussed. Often, metacognition is exchanged verbally or used privately
but, for us, the essential defining feature is that metacognition can be made
public. Thus, it is reportable, conscious awareness about cognitive aspects of
thinking. For example, understanding how the characteristics of a passage
impede or facilitate comprehension is metacognition about reading. If an in-
dividual pauses when a new word is encountered in text and rereads sur-
rounding words to figure it out but is unaware of the process and cannot re-
port the activity, it is superfluous to attribute metacognition to the behavior.
Automatic skills, no matter how sophisticated, do not necessarily imply
metacognition on the part of the learner. Perhaps metacognition was a com-
ponent of the skill during learning or development, but if awareness is no
longer available or required, then inferences about a tacit executive or ho-
munculus are unwarranted inferences about a deus ex machina. By re-
stricting the definition to shared knowledge we constrain the term considera-
bly but avoid false inferences about the occurrence of metacognition when it
cannot be observed or measured.
Like many others, we divide metacognition into two broad categories:
(a) self-appraisal of cognition and (b) self-management of thinking. Self-
appraisal refers to the static assessment of what an individual knows about a
given domain or task. The appraisal can be of one's abilities or knowledge, or
it might involve an evaluation of the task or consideration of strategies to be
used. In any domain, these appraisals of thinking appear to fall into three
broad subcategories that we refer to as declarative, procedural, and condi-
tional knowledge. Declarative knowledge refers to what is known in a propo-
sitional manner. For example, a student might know that topic familiarity
and prior knowledge influence reading speed and comprehension or that
rereading facilitates memory. Procedural knowledge refers to an awareness
of processes of thinking. For example, a student could know how to skim,
how to use context, how to underline, how to summarize, or how to find the
main idea while: reading. Awareness of cognitive processes involved in think-
ing is a fundamental aspect of metacognition. Conditictnal knowledge refers
to an awareness of the conditions that influence learning such as why strate-
gies are effective, when they should be applied and when they are appropri-
ate. For example, students can become aware of the value of periodic
paraphrasing as a means for monitoring comprehension, yet they also need
to realize that paraphrasing is a strategy used selectively for some purposes
with some types of text. Any kind of self-appraisal of cognition can be classi-
Downloaded by [Harokopio University] at 14:07 22 April 2012

fied as either declarative, procedural, or conditional knowledge.

Self-management refers to the dynamic aspects of translating knowledge
into action. Three types of executive processes can encompass the activities
of self-regulated thinking. The first is plcnning, which refers to the selective
coordination of a cognitive means to a cognitive goal. For example, profi-
cient readers adjust their rate of reading and standards of comprehension ac-
cording to the purposes of the tasks and constraints that are imposed (Baker,
1985). The second dimension of self-management is evaluation. For exam-
ple, readers can evaluate their own understanding as they pause, paraphrase,
answer questions, or summarize information in text. Evaluating thinking is
an ongoing process in any domain. The third aspect is regulation. Self-
managed thinking requires an individual to monitor progress and then revise
or modify plans and strategies depending on how well they are working. Self-
regulation allows the learner to adjust to changing task demands as well as to
successes and failures.
In summary, our restrictive definition of metacognition emphasizes public
knowledge about cognitive activities. We also compartmentalize metacog-
nition into a simple taxonomy of self-appraisal and self-management with
three subcategories in each. The definition and taxonomy facilitate measure-
ment and instruction, as we illustrate in the following sections. We hasten,
however, to add this comment. Metacognition is not, in our way of thinking,
the goal of development or education. Becoming aware of the contents and
procedures of one's mind can be helpful for solving problems and completing
tasks. It is this functional, means-ends relation that we emphasize in our
view of metacognition and not reflection for its own sake. Being mindful and
being metacognil.ive are goal-oriented aspects of thinking that are amenable
to instruction by others as well as to self-regulation (Paris & Oka, 1986).

Most studies of children's knowledge about reading have used interviews to

measure metacognition. Reid (1966) conducted one of the first studies to ask
children what they thought about reading. She found that 4- and 5-year-olds
did not know the goals of reading or the function of letters, words, or punc-
tuation. In related research, Clay (1972) found that beginning readers were
often confused about whether they should read the pictures or the print.
Johns (1980) observed a similar lack of knowledge among beginning readers.
Wixson, Bosky, Yochum, and Alvermann (1984) found that many young
children believe that the purpose of reading is to pronounce all the words
without mistakes and that good reading includes verbatim recall of the text.
The first study of metacognition with older readers was conducted by
Myers and Paris (1978) who used Flavell and Wellman's (1977) categories of
person, task, and strategy variables to construct an interview. Myers and
Paris used children's free responses to a scripted interview to examine the
Downloaded by [Harokopio University] at 14:07 22 April 2012

knowledge about reading reported by 8- and 12-year-olds. They found that

older children were more aware than younger children of the effects of many
variables on reading and the utility of strategies of comprehension. In a sub-
sequent study, Paris and Myers (1981) compared good and poor fourth-
grade readers who were matched for age, sex, and arithmetic achievement. It
was observed that good readers knew more about reading strategies, detected
errors more often while reading, and had better memory for text than did
poor readers. Canney and Winograd (1979) studied children's beliefs about
reading using an interview technique with an experimental manipulation.
Children who were 8, 10, 12, and 14 years old were presented with passages
that were either coherent or disrupted. When children were asked if each
passage could be read and why, it was found that young and poor readers at-
tended more to decoding, whereas proficient readers knew that making sense
was the goal of reading. Kobasigawa, Ransom, and Holland (1980) observed
that 10-year-olds had only a rudimentary understanding of skimming,
whereas 14-year-olds could describe and use skimming as a strategy to aid
Two large studies of children's metacognition require more detailed com-
ments. Paris and Jacobs (1984) reported a study in which third- and fifth-
grade children were given metacognitive interviews before and after a meta-
cognitive intervention designed to increase their understanding and use of
comprehension strategies. The interview was a modification of the Myers
and Paris (1978) instrument. The original interview included 33 Likert-scale
items and 19open-ended questions. Fifteen of the open-ended questions were
consolidated into an interview score, with 5 items pertaining to each sub-
category of self-managed reading (i.e., planning, evaluation, and regula-
tion). Each of the 15 open-ended questions was scored 0, 1, or 2 points de-
pending on the level of strategic awareness indicated in the child's response.
The composite scores on the 30-point scales ranged from 7 to 26. The reading
awareness scores were significantly higher at the posttest for 10-year-olds
than for 8-year-olds and significantly higher for the children in the experi-
mental classrooms.
In addition, Paris and Jacobs (1984) asked students if they used strategies
such as rereading and predicting. The following four questions were asked:
1. Do you ever go back and read things over and over again?
2. Do you ever stop while you are reading and try to guess what happens
3. Do you ever take notes, underline words, or try to imagine yourself in a
story while you read?
4. Do you try to think about what the sentences mean and how they go
Analysis of covariance indicated significantly greater reported use of the
Downloaded by [Harokopio University] at 14:07 22 April 2012

strategies at the posttest by children in the experimentall groups.

When participants in the study were divided into groups of high, middle,
and low awareness about reading based on their metacognitive scores, it was
evident that children with high awareness consistently scored higher than
other children on standardized reading tests, cloze tasks, and error detection
tests. The correlations between metacognition and these various measures of
comprehension were modest (.24 to .40), but significant. Thus, Paris and Ja-
cobs (1984) demonstrated that interview measures can yield information
about metacognition that distinguisheschildren on the basis of age and com-
prehension skillls and is also sensitive to instructional treatments within
Another large study of children's metacognition reached a different con-
clusion, however. Forrest-Pressley and Waller (1984) reported an extensive
correlational study of children's awareness about reading (decoding, com-
prehension, and strategy efficiency) and cognition (language, memory, and
attention). They correlated measures of verbalization (i.e., reported aware-
ness) and performance for 144 children in Grades 3 and 6 on 12 experimental
tasks, Gates-Ma.cGinitie reading achievement scores, and nonverbal IQ
scores. Their data reveal some interesting and some puzzling correlations.
For example, composite scores of verbalization about comprehension were
derived from the following questions:
1. How do yolu know when you are ready to write a test?
2. Would you know how well you had done on the test before you got it
back? How?
3. How would you know when you knew enough about a game to be able
to teach it to someone else?
Although Forrest-Pressley and Waller (1984) did not discuss the intercorre-
lations, data in their Appendix U reveal that the composite scores correlated r
= .83 with their measure of comprehension performance and r = 3 2 with
Gates-MacGinitie scores at Grade 3. The comparable correlations at Grade 6
were .80 and .57. These correlations are remarkably high and suggest strong
relations between reported awareness and reading performance.
Other correlations between verbalization and performance in that study
were much lower, however. Awareness about decoding was assessed by com-
posite scores derived from four questions:

1. What do you do when you come to a word that you do not know?
2. Is there a difference between knowing what a word "says" and knowing
what a word "means"?
3. Is it better to sound out a word that you do not know or ask someone
what it says?
4. What do you do when you come to a whole sentence that you do not
Downloaded by [Harokopio University] at 14:07 22 April 2012


The composite score correlated only r = .12 with decoding performance and
r = .23 with Gates-MacGinitie scores at Grade 3 and rs = .12 and .20 at
Grade 6 for comparable correlations. Although reported knowledge about
decoding improved with age and reading ability, verbalizations about de-
coding and reading performance were barely related.
The final aspect of reading awareness that can be derived from Forrest-
Pressley and Waller's (1984) data concerns strategies. The verbalization com-
posite score was derived from children's responses to the following questions:

1. What do you do when you read in preparation for a test?

2. Is there anything that you can do to make what you are reading easier to
3. How would you find the name of a place in a story?
4. How would you remember a story so that you could tell it to a friend
5. How much of the story would you remember?
6 . How would you think of a title for a story?

At Grade 3, the composite awareness score correlated r = .35 with the au-
thors' measure of strategy efficiency, r = .30 with the comprehension per-
formance measure, and r = .30 with Gates-MacGinitie scores. The compar-
able correlations at Grade 6 were rs = .44, .36, and .44.
Forrest-Pressley and Waller (1984) correlated awareness measures with
traditional comprehension scores derived from paragraph reading followed
by questions. The relatively low correlations that they observed between
metacognition and comprehension may be due to their choice of measures -
Gates-MacGinitie reading achievement scores, sentence completion in a
multiple-choice format (i.e., recognition of sentences from test), and an unu-
sual "strategy efficiency" score. This last measure was computed by dividing
the passage comprehension score from the sentence-completiontask by the
logarithm of the time that the child took to read each passage. Evidently the
time to read the passages had little impact on the measure because strategy ef-
ficiency and comprehension scores correlated rs = .9T and .96 at Grades 3
and 6 . In fact, the "strategy efficiency" scores also correlated rs = .71 and .79
with the Gates-MacGinitie achievement scores, which suggests that all three
measures assessed similar characteristics of speeded paragraph reading and
question answering. The actual comprehension strategies and their effi-
ciency, though, were not assessed.
The data off~era range of relations between reading awareness and per-
formance from strong to weak that Forrest-Pressley and Waller (1984) did
not reconcile. It is difficult to compare their findings to other studies because
Downloaded by [Harokopio University] at 14:07 22 April 2012

they did not compute a total verbalization score an~dcorrelate reported

awareness on the 13interview items to reading performance, nor did they dis-
cuss the intercorrelations reported here. Instead, they used repeated stepwise
regression analyses to predict Gates-MacGinitie scores with each analysis
employing specific performance and verbalization measures for the six tasks
(and nonverbal IQ scores). In every case, the performance measures pre-
dicted Gates-MiacGinitie scores first with little variance accounted for by ver-
balization measures. In an overall stepwise regressiorr analysis predicting
Gates-MacGinitie scores with the three verbalization and three perforrnance
measures, the performance measures were the first three predictors at each
grade level accounting for more than 67% of the variance at Grade 3 and
77% of the variance at Grade 6. From these analyses, Forrest-Pressley and
Waller concluded that verbalization or metacognition is not related strongly
to children's reading.
Their conclusions, though, may be pessimistic because of their failure to
interpret the intercorrelations, compute a composite awareness score, and
conduct hierarchical regression analyses with verbalization scores entered
first. The importance of metacognition may be underestimated if the analy-
ses partial out the relations among various performance measures first and
only allow metacognition to account for the residual variance. The over-
reliance on standardized reading achievement scores is a serious weakness as
well because they are so highly correlated with intelligence and test-taking
skills that they may not be sensitive to specific knowledge about reading nor
specific reading strategies (Paris, Cross, & Lipson, 1984). Basically, the
Gates-MacGiniti~ereading test and the sentence-completiontask created by
Forrest-Pressley and Waller (1984) to assess comprehension afford minimal
opportunities for rich strategies because of the premium placed on rapid

decoding and question answering. For example, in Forrest-Pressley and

Waller's data, the Gates-MacGinitie scores correlated r = .45 with memory
performance and r = .40 with nonverbal IQ at Grade 3, and rs = .51 and .57
for comparable scores at Grade 6. Their measures of reading comprehension
were confounded, and it is likely that their correlations between reading
awareness and performance measures seriously underestimate the relation.
All these interview studies used children's reported knowledge about read-
ing as the primary data. Collectively, the studies have revealed progressive
awareness about reading with increasing age and reading ability. However,
the use of verbal reports has been criticized by these authors and others. Gar-
ner (1987) reported eight criticisms of verbal report data on children's read-
ing, which are paraphrased in the following axioms:

1. Some cognitive knowledge and processes are tacit and inaccessible, yet
reflection on concurrent cognitive processing can often be accurate.
2. Verbal report data may include participants' rationalizations, elicited
Downloaded by [Harokopio University] at 14:07 22 April 2012

mimicry, and fabrications because of the demand characteristics of the

3. Children may lack the language and verbal facility to discuss mental
4. The behavior and characteristics of the interviewer can elicit answers
perceived to be socially desirable.
5. Reliability of interviews is rarely assessed.
6. Forgetting may interfere with introspective reports.
7. Children have difficulty answering questions about hypothetical situ-
8. Asking questions during or after cognitive processing can disrupt
Because of these difficulties in measuring children's metacognition, many
researchers have chosen to infer knowledge and executive control from dem-
onstrations of children's planful or reflective use of cognitive skills. Al-
though these inferences may be warranted on occasion, they run the risk of
assuming that children understand more than they actually do about the vari-
ables that influence thinking. When these inferences are coupled with impre-
cise definitions and measurements of metacognition, the construct can take
on an undeserved ubiquity and executive power in explanations of cognitive
processing. More precise definitions and better measurement techniques can
reduce some of these problems.


Most metacognitive training studies have examined memory strategies in

laboratory tasks. For example, it has been shown that the provision of feed-
back about the utility of memory strategies can significantly increase chil-
dren's continue~duse of the strategies (Fabricius & Hagen, 1984; Kennedy &
Miller, 1976; Paris, Newman, & McYey, 1982). Numerous studies have also
shown that children can be taught to monitor their own performance and to
use executive strategies (Borkowski & Kurtz, 1987; Lodlico, Ghatala, Levin,
Pressley, & Bell, 1983; Pressley, Borkowski, & O'Sullivan, 1985). Thus, in-
structional studbes that have increased the accuracy of children's self-ap-
praisal or the effectiveness of self-management have improved memory
Reading researchers have hypothesized that metacognition plays a similar
role in reading (Baker & Brown, 1984). Yet, there is considerably less evi-
dence to support this claim. If reading awareness is related to compre-
hension, then teachers should be able to facilitate comprehension in their
students by including information about cognitive self-appraisal and self-
management in their reading lessons. Unfortunately, teachers in inter-
mediate grades rarely provide explicit instruction to enhance comprehension
Downloaded by [Harokopio University] at 14:07 22 April 2012

(Durkin, 1979) and teachers' manuals provide little information about com-
prehension strategies (Durkin, 1981). There are at least three notable excep-
tions during the past few years that have illustrated the benefits of meta-
cognitive instruction for children's reading comprehension. With our col-
leagues, we have demonstrated that elementary-school children can be taught
comprehension strategies explicitly. In a program called "Informed Strate-
gies for Learning," 8- to 12-year-oldswere taught what comprehension strat-
egies are, how they operate, when they should be used, and why they are ef-
fective. Explaining thinking skills that can be used before, during, and after
reading proved beneficial to children of all ages and reading abilities. Chil-
dren in the experimental classes showed significant gains from pretest to
posttest in reading awareness and strategic reading (Paris, Cross, & Lilpson,
1984; Paris & Jacobs, 1984; Paris & Oka, 1986).
In a second demonstration of the importance of metacognition, Palincsar
and Brown (1984) instituted a procedure of peer tutoring that they called "re-
ciprocal teaching" because students alternated roles of tutor and learner. In-
struction focused on the independent use of four important strategies: self-
questioning, summarizing, paraphrasing, and predicting information in text.
After 20 consecutive days of instruction, children who participated in recip-
rocal teaching showed significant gains in reading comprehension and mem-
ory. No measures of awareness or strategic reading were taken in these stud-
ies, though. Cleaxly, additional research is needed to demonstrate the effec-
tiveness of enhancing students' metacognition about reading.
Gerald Duffy and his colleagues examined teachers' explanations of read-
ing lessons and comprehension strategies in a variety of studies. They found
that teachers who provide explicit descriptions of strategiesto be learned and
used during reading lessons promote students' understanding of lesson con-
tent (Duffy, Roehler, & Rackliffe, 1986). In another study, the researchers

showed that teachers could be trained to model metacognitive approaches to

reading so that students received explicit verbal explanations about strategies
to be learned (Duffy, Roehler, Meloth et al., 1986). Twenty-two fifth-grade
teachers were trained to incorporate explicit explanations into their routine
basal-reading skill instruction. "Specifically, teachers were taught to empha-
size the mental processing one does when using the skills prescribed in the ba-
sal textbook" (p. 244). A five-step lesson format was taught: introduction,
modeling, guided interaction, practice, and application. Results of the study
revealed that teachers learned quickly to be more explicit in their explana-
tions of the lessons. Furthermore, students in treatment classrooms showed
significant increases in metacognition as measured by their responses to the
following questions:

1. What were you learning in the lesson I just saw?

2. When would you use what was taught in the lesson?
3. How do you do what you were taught to do?
Downloaded by [Harokopio University] at 14:07 22 April 2012

Despite these changes in teachers' and students' metacognition, there was no

change in students' Gates-MacGinitie reading scores. Unfortunately, no
other measures of reading performance were obtained and thus it is unclear if
the treatment was ineffective or if the outcome measure was insensitive.


The study that is reported in the remainder of this article describes an in-
formal inventory of reading awareness that can be used in classroom re-
search. We describe briefly a study that confirms the effectiveness of meta-
cognitive instruction. During the past 7 years, our research team at the Uni-
versity of Michigan has been developing and testing an instructional program
that we call "Informed Strategiesfor Learning" (ISL; Paris, Cross, Jacobs, et
al., 1984; Paris, Cross, & Lipson, 1984). ISL is a research project designed to
test the relation between children's awareness about reading strategies and
their reading skills, but it is also a practical approach to classroom instruc-
tion. There are three basic objectives in ISL. The first is to increase young
children's understanding of reading tasks, goals, and strategies by describing
what, how, and why various strategies influence reading. The second objec-
tive is to provide an experimental test of the relation between metacognition
and performance: Can children's reading skills and comprehension levels be
promoted by teaching them information about reading strategies? The third
objective is to develop an instructional method for informing children about
reading that is interesting, easy for teachers to use, and suitable for young

readers. ISL involves direct instruction by classroom teachers to the entire

class and includes discussion and teacher-student interaction so that students
can increase their understanding about reading as well as their skills and mo-
tivation. Further details about the lessons and materials can be found in Paris

Evaluating Students' Comprehension Strategies

Students. Forty-six teachers from 18 schools (2 districts) in southeast-

ern Michigan volunteered to participate in the experimental reading curricu-
lum; 25 teachers in the same two districts provided control classrooms. There
were 783 third graders and 801 fifth graders from intact classrooms parti-
cipating in the study. Approximately 20% of the children were from black,
Middle Eastern, or Asian families with the remainder from white families.
There were similar numbers of boys and girls.
Downloaded by [Harokopio University] at 14:07 22 April 2012

Index of Rteading Awareness (IRA). A multiple-choice Index of

Reading Awareness (IRA) was constructed to provide data about children's
reading awareness and to serve as an index of the treatment's effectiveness.
Data from the IliA were also analyzed in relation to children's performance
on a standardized comprehension test. A multiple-choice test of metacog-
nition avoids some of the pitfalls of verbal reports noted by Garner (1987)
and others. Although the IRA does not solve all the problems mentioned, it
was designed to meet four criteria.
First, a multiple-choice test is more objective than inAerviews that may in-
volve interpretations of open-ended responses, experimenter bias, or fabri-
cated responses. Although the IRA allows guessing, it does not place shy or
inarticulate children at a disadvantage. Second, we wanted to construct a
measure that was based on empirical research of children's responses to
metacognitive questions, so that it accurately reflected children's knowledge
about reading strategies rather than our beliefs about what they know. Ease
of administration and scoring was the third criterion. We wanted a measure
that could be administered to groups in a limited amount of time and that
could be scored quickly by teachers. This is particularly important if the in-
strument is to be used by classroom teachers as part of reading skill invento-
ries. Finally, we wanted to create a measure that would be sensitive to individ-
ual and age-related differences in awareness about reading. The IRA was
designed for children in the third and fifth grades with grade-equivalentread-
ing abilities from second to seventh grade. Both the age and skill ranges are
critical in most theoretical accounts of metacognition (e.g., Baker & Brown,
l984), and the ir~!strument'susefulness depends on the degree to which it can
discriminate changes in awareness with age and reading ability. We also
wanted an instrument sensitive enough to distinguish between increases in
awareness occurring during the course of the normal school year and in-
creases due to instruction.
The IRA was designed to measure three aspects of metacognition: evalua-
tion, planning, and regulation (Paris & Lindauer, 1982). Fifteen of the IRA
items were the questions used by Paris and Jacobs (1984). The response fre-
quencies from the earlier interview were used to develop representative
multiple-choice responses as the three alternatives for each question. In addi-
tion, we devised a fourth category of questions that measured children's un-
derstanding about how specific strategies met various reading goals. These
questions were designed to measure conditionalknowledge, a term coined by
Paris et al. (1983) to denote children's understanding of when and why partic-
ular strategies should be applied. The IRA questions assessed children's
knowledge about reading and their abilities to evaluate tasks, goals, and per-
sonal skills; to plan ahead for specific purposes; to monitor progress while
reading; and to recruit fix-up strategies as needed. The IRA questions and
alternatives are shown in Table 1.
Downloaded by [Harokopio University] at 14:07 22 April 2012

The entire IRA included 20 questions, each with three alternatives repre-
senting an inappropriate response (0 points), a partially adequate answer (1
point), and a strategic response (2 points). The order of the choices was
randomized. The rationale for assigning point values to each alternative was
as follows. Responses in the zero category were inappropriate or denied the
problem. Responses in the 1-point category were adequate responses based
on decoding, external features of the text, or vague references to affective or
cognitive reactions, but with no mention of a specific strategy. Choices re-
ceiving 2 points were strategic responses that were evaluative or planful and
exhibited awareness of reading goals and strategies. The following example
illustrates the scoring:
If you are reading for science or social studies, what would you do to re-
member the information?

a. Ask yourself questions about the important ideas. (2 points)

b. Skip the parts you don't understand. (0 points)
c. Concentrate and try hard to remember. (1 point)
In this example, choice b received zero points because it is an inappropriate
response to the goal of remembering science or social studies information. In
some situations it may be a reasonable strategy for getting through the mate-
rial, but not in this case. The answer c received 1 point because it describes a
general cognitive act indicating an understanding that some extra effort and
special thinking will be required to remember the material. However, no spe-
cific strategy for the goal is mentioned. In this example, choice a received 2
points because it specifies an active self-questioning strategy that allows stu-
dents to monitor retention of the material. Scores for the 20 questions on the
multiple-choice IRA were combined to produce a total score ranging from 0
to 40 points.
Index o f Reading Awareness Items

Score Items


1. What is the hardest part about reading for you?

1 a. Sounding out the hard words.
2 b. When you don't understand the story.
0 c. Nothing is hard about reading for you.
2. What would help you become a better reader?
1 a. If more people would help you when you read.
0 b. Reading easier books with shorter words.
2 c. Checking to make sure you understand what you read.
3. What is special about the first sentence or two in a story?
1 a. Thjey always begin with "Once upon a time . . ."
0 b. Thie first sentences are the most interesting.
2 c. They often tell what the story is about.
Downloaded by [Harokopio University] at 14:07 22 April 2012

4. How are the last sentences of a story special?

1 a. Thqy are the exciting, action sentences.
2 b. They tell you what happened.
0 c. They are harder to read.
5. How can you tell which sentences are the most important ones in a story?
2 a. They're the ones that tell the most about the characters and what happens.
1 b. They're the most interesting ones.
0 c. All of them are important.


1. If you could only read some of the sentences in the story because you were in a
hurry, which ones would you read?
a. Read the sentences in the middle of the story.
b. Read the sentences that tell you the most about the story.
c. Read the interesting, exciting sentences.
2. When you tell other people about what you read, what do you tell them?
a. What happened in the story.
b. The number of pages in the book.
c. Who the characters are.
3. If the teacher told you to read a story to remember the general meaning, what would
you do?
a. Skim through the story to find the main parts.
b. Read all of the story and try to remember everything.
c. Read the story and remember all of the words.
4. Before you start to read, what kind of plans do you make to help you read better?
a. Yon don't make any plans. You just start reading.
b. You choose a comfortable place.
c. You think about why you are reading.
5. If you hatd to read very fast and could only read some words, which ones would you
try to read?
a. Read the new vocabulary words because they are important.
b. Read the words that you could pronounce.
c. Read the words that tell the most about the story.
TABLE 1 (Continued)

Score Items


1. What things do you read faster than others?

1 a. Books that are easy to read.
2 b. When you've read the story before.
0 c. Books that have a lot of pictures.
2. Why do you go back and read things over again?
1 a. Because it is good practice.
2 b. Because you didn't understand it.
0 c. Because you forgot some words.
3. What do you do if you come to a word and you don't know what it means?
2 a. Use the words around it to figure it out.
1 b. Ask someone else.
0 c. Go on to the next word.
4. What do you do if you don't know what a whole sentence means?
1 a. Read it again.
Downloaded by [Harokopio University] at 14:07 22 April 2012

0 b. Sound out all of the words.

2 c. Think about the other sentences in the paragraph.
5. What parts of the story do you skip as you read?
1 a. The hard words and parts you don't understand.
2 b. The unimportant parts that don't mean anything for the story.
0 c. You never skip anything.

Conditional Knowledge

1. If you are reading a story for fun, what would you do?
a. Look at the pictures to get the meaning.
b. Read the story as fast as you can.
c. Imagine the story like a movie in your mind.
2. If you are reading for science or social studies, what would you do to remember the
a. Ask yourself questions about the important ideas.
b. Skip the parts you don't understand.
c. Concentrate and try hard to remember it.
3. If you are reading for a test, which would help the most?
a. Read the story as many times as possible.
b. Talk about it with somebody to make sure you understand it.
c. Say the sentences over and over.
4. If you are reading a library book to write a book report, which would help you the
a. Sound out words you don't know.
b. Write it down in your own words.
c. Skip the parts you don't understand.
5. Which of these is the best way to remember a story?
a. Say every word over and over.
b. Think about remembering it.
c. Write it down in your own words.
The control group provided a comparison of average changes in awareness
to be expected during the school year for children who did not receive ISL in-
struction. To check this, pretest and posttest scores for the control group (n
= 544) were correlated. After an 8-month interval, the Pearson prod-
uct-moment correlation was r = 3 , p < .001. To insure the appropriate-
ness of the measure for children of different ages, all items were tested sepa-
rately for third and fifth graders. No ceiling or floor effects were found for
either grade and no items were unusually skewed. The standard deviation for
items ranged between .49 and 239.

Standardized comprehension test. The comprehension subtest of

the Gates-MacGinitie Reading Test (MacGinitie, 1978) was used to measure
reading level because it is a group-administered, normatively referenced test
that provides raw, percentile, and extended-scalescores. The comprehension
subtest measures children's abilities to answer questions about text informa-
tion they have read. Equivalent forms of the test were used for each grade
Downloaded by [Harokopio University] at 14:07 22 April 2012

(Level C for third graders and Level D for fifth graders). Form 1was used for
the pretest and Form 2 was used for the posttest at each level. The pretest
extended-scale sicores were used to divide children in each grade into three
reading levels. Low, medium, and high reader groups were formed by parti-
tioning our sample at the 33rd and 67th percentiles of the national norms.
This resulted in a similar distribution of reader groups at both grades.

Teaching Comprehension Strategies

Testing. The tasks were administered to intact classes in conjunction

with other tests in the larger reading project. The pretests were administered
from September to October by trained experimenters during two 1-hr ses-
sions. Posttests on the same tasks were administered in April and May of the
same school year. Experimenters instructed children to circle the "best an-
swer for you" on the IRA. Experimenters read each question and each choice
aloud to insure that all children were able to understand the questions and to
ensure a uniform pace. Experimenters paused until all children had selected
answers before going on to the next question. Questions and choices were re-
peated when necessary. The Gates-MacGinitie Reading Test is a timed test.
Experimenters read specific instructions at the beginnin,gof the test and chil-
dren worked on their own for the allotted time period.

Treatment. ISL was designed to teach children about the existence and
appropriate use of various comprehension strategies. The methods incorpo-
rated aspects of both informed and self-control training (Brown et al., 1984)
so that instruction provided explicit information, practice, and discussions
about cognitive strategies. Students were not only told about strategies, they
were also persuaded by teachers to use the skills when reading both narrative
and expository texts.
Group discussions were structured by the materials and lesson plans pro-
vided to teachers. Each strategy was represented by a metaphor. For exarn-
ple, to describe comprehension monitoring, we drew analogies between traf-
fic signs (e.g., stop, yield, caution, dead end) and similar cues in text that
might signal occasions to paraphrase, to look up new words in the dictionary,
to slow down the reading rate, or to reread. Each metaphor was portrayed on
a colorful and large bulletin board display that teachers used as a reference
point in each lesson. The metaphors were also repeated on work sheets that
accompanied each lesson (e.g., traffic signs) and allowed students immediate
practice with strategies. Each successive lesson and work sheet required stu-
dents to recall and select appropriate strategies with fewer explicit instruc-
tions so that students gradually assumed more responsibiIity for guiding their
own reading. Group discussions before the activities focused on how, when,
Downloaded by [Harokopio University] at 14:07 22 April 2012

and why to use the strategies; discussions following the seatwork provided
shared feedback about the ease, benefits, or difficulties of using each strat-
egy. Thus, each lesson included explicit information about comprehension
strategies as well as discussions and practice that were intended to convince
students about the value of using strategies independently.
The entire package of instructional materials included 20 modules ar-
ranged in groups of 5 that were sequentially aimed at (a) planning and
preparing to read, (b) identifying meaning, (c) reasoning about text content,
and (d) comprehension monitoring. The first 4 modules in each set intro-
duced a distinct strategy; the fifth module was a review that integrated and
repeated the previous four strategies. Although there are many strategies that
are worthwhile candidates for instruction, we chose strategies identified by
research as critical for comprehension and within the grasp of 8- to 10-year-
olds. Strategies that involved construction of text meaning (e.g., elaboration,
inference, integration, activation of prior knowledge, and summarization)
and comprehension monitoring (e.g ., rereading, self-questioning, checking
consistencies, and paraphrasing) were especially emphasized throughout the
Each module included three lessons, each of which was designed to be pre-
sented to the entire class in approximately 30 min. The series of lessons was
designed to "fade" the teachers' support and place more responsibility on stu-
dents to recruit and apply strategies. The first two lessons introduced the met-
aphor, described the strategy, and explained how to use the strategy while
reading. These lessons were followed by group practice and feedback, then
individual practice with texts and worksheets. The third lesson was referred
to as a "bridging" lesson because the reading material was to be taken from
other parts of the teachers' curriculum such as social studies or science so that

students could Itearn directly that these strategies should be transferred to

content-area reading. The materials were described and explained to teachers
in four separate workshops and periodic in-service meetings throughout the
school year.

Effects of Teaching Comprehension Strategies

A comparison of pretest and posttest scores for the experimental and con-
trol groups permits us to contrast increases in reading awareness occurring
during the course of an average school year with increases due to the effec-
tiveness of the instruction. The first section included analyses that show the
effect of the intervention on children's awareness scores and its relation to sex
and grade level. The second section describes the impact on children of dif-
ferent reading abilities. The analyses presented answer the questions: (a) Did
classroom instruction improve reading awareness? and (b) Do children with
different abilities in reading benefit equally from instruction about reading
Downloaded by [Harokopio University] at 14:07 22 April 2012


Effects of cl,assroom instruction. The 20-item IRA provided a 40-

point scale of reading awareness that is the main index of children's knowl-
edge about reading in all analyses reported here. Pretest reading awareness
scores ranged from 12 to 39, with a mean of 23.0 for third graders and 26.9
for fifth graders,,The impact of the instructional intervention on children's
reading awareness was assessed by conducting a 2 (Treatment) x 2 (Grade)
x 2 (Sex) analysis of covariance (ANCOVA), using the pretest awareness
scores as covariates. This analysis revealed that children in the experimental
group received significantly higher scores than their control counterparts
after a year of metacognitive instruction, F(l, 1574) = 90.8 1,p < .001. Fifth
graders also scored significantly higher than third graders on the IRA, F(l,
1573) = 19.13, p c .001; and girls scored significantlyhigher than boys, F(1,
11574) = 37.88, p < .001. However, no interactions between treatment,
grade, and sex arere found.
Although all children in Grade 3 increased their reading awareness during
the school year, tlhe gains made by the experimentalgroulp were clearly larger.
the two groups showed nearly equivalent reading awareness at the beginning
of the year, starting within 1 point of each other, but by the end of the in-
structional period, the gap had widened to nearly 3 points (which is approxi-
mately equal to one standard deviation). A similar pattern of gains was found
in the fifth grade. Although the slopes were slightly less dramatic, the fifth
graders participating in the instructional program gained significantly more
than the children in the control classes. Overall, fifth graders scored about an
average of 3 points higher than third graders in both groups.
A closer look at the performance sf girls and boys on the IRA indicates
that girls scored significantly higher than boys in both grades. Although girls
and boys in the third grade started the year with scores within 1 point of each
other, girls in both treatment groups performed better by the end of the year.
In the fifth grade, boys started more than 1 point lower than girls and the dif-
ference was maintained throughout the year for both treatment groups.

Relation between reading ability and reading awareness. Children

were divided into low, medium, and high reader groups by using the pretest
scores on the Gates-MacGinitie comprehension subtest. To test the relation
between awareness, initial reading ability and instruction, a 2 (Grade) x 2
(Treatment) x 2 (Sex) x 3 (Reader) ANCOVA was performed using the pre-
test awareness scores as covariates. This analysis yielded a highly significant
main effect for reader, F(2, 1567) = 44.27, p < .001, indicating that good
readers are more aware of important reading variables and poor readers are
less aware. This finding also indicates that the IRA is sensitive to reader dif-
ferences. A significant interaction between reading level and treatment in this
Downloaded by [Harokopio University] at 14:07 22 April 2012

analysis would support the hypothesis that one group of readers received
greater benefits from the instruction than another group. However, no inter-
action was found using this procedure, which means that readers at all levels
benefited from the ISL program. Again we found large effects, similar to
those reported earlier, for grade, F(1, 1436) = 28.81, p < .001; sex, F(1,
1536) = 4 5 . 9 9 , ~< .001; and treatment, F(6, 1536) = 84.45, p < .001. No
significant interactions were found.


One objective of this study was to investigate the effects of classroom in-
struction on reading awareness. The data presented here provide convincing
evidence that a classroom-based program of metacognitive instruction can
improve children's awareness and understanding of reading strategies. We
also found that fifth graders exhibited greater reading awareness than third
graders and that girls are more aware of reading strategies than are boys.
Most striking, however, was the finding that girls and boys at all reading lev-
els and in both grades benefited from the instructional program. The effect
was not limited to a particular subgroup of children.
A second objective was to develop a measure of metacognition about read-
ing to study the effects of instruction, and we described the construction of
the measure in detail. The IRA is an objective test, based on empirical data,
that is easy to administer in classrooms. The data also indicate that the IRA is
sensitive to changes in awareness due to individual differences in age, sex,
and reading ability. This informal assessment can help to differentiate be-
tween children who know a lot about evaluating, planning, and regulating

their own readiing and those children who do not. However, certain caveats
should be mentioned. First, the IRA reflects our conceptual framework of
reading awareness and measures knowledge about some concepts that we be-
lieve are related to reading comprehension. These concepts were also the ba-
sis for the weelcly lessons. Although it is important to connect definitions,
measures, and interventions conceptually and operationally, there are other
strategies and knowledge that may be just as important for effective reading.
Second, our data indicate that awareness increases without special instruc-
tion. An instructional program such as ISL promotes reading awareness and
increases knowledge about cognitive skills and strategies more than occurs
without such instruction; however, children's understanding usually im-
proves during the course of a school year.
The success of ISL indicates that metacognition is an important part of
proficient reading and that it can be taught in the classroom. Indeed, many
teachers already emphasize children's conceptual understanding about read-
ing and may use techniques similar to ISL, direct explanation, or reciprocal
Downloaded by [Harokopio University] at 14:07 22 April 2012

teaching. Metacognitive instruction can be added to the classroom curricu-

lum, particularly in reading, writing, and language arts, But metacognition is
not a panacea for children's reading problems and should not be the curricu-
lum objective. Understanding the processes of thinking while reading is a
tool to read moire effectively and not an end in itself.
Finally, this study highlights the importance of developing measures of
metacognition in conjunction with instructional programs. Measuring chil-
dren's metacognition and thinking skills directly helps teachers to diagnose
specific misconceptionsand nonstrategic reading. In this manner, cyclical as-
sessment and instruction can become intertwined to encourage students'
learning. Recip~rocalteaching programs, for example, include informal, re-
peated assessments to monitor student progress and to identify targets for
remediation (Palincsar, Brown, & Martin, this issue).
It is clear that children's awareness about reading and their use of effective
strategies can be promoted by instruction. There are many methods and ma-
terials that can be developed to teach and evaluate metacognition; the initial
success of ISL should encourage future efforts. Now that the first glow of
metacognition as a "new approach" to reading has faded, the challenge is to
continue to tackle the tough issues of defining, measuring, and fostering stu-
dents' metacoginitive approaches to reading.


This research was supported by National Institute of Education Grant

NIE-G-82-0019. We appreciate the help of David Cross, Evelyn Oka, Ann
Marie DeBritto, and David Saarnio who collaborated with us on the project.
Special thanks also go to the teachers and students of Waterford and Farm-
ington school districts for their cooperation. We thank Karen Wixson for her
helpful comments on an earlier version of this article.

Babbs, P., & Moe, A. (1983). Metacognition: Key for independent learning from text. Reading
Teacher, 36, 422-426.
Baker, L. (1985). How do we know when we don't understand? Standards for evaluating text
comprehension. In D. L. Forrest-Pressley, G. E. MacKinnon, & T. G. Waller (Eds.),
Metacognition, cognition, and human performance (Vol. 1, pp. 155-205). New York:
Baker, L., &brown, A. L. (1984). Metacognitive skills of reading. In D. Pearson, R. B a r , M.
Kamil, & P. Mosenthal (Eds.), Handbook of reading research (pp. 353-394). New York:
Borkowski, J. G., & Kurtz, B. E. (1987). Metacognition and executive control. In J. Borkowski
& J. Day (Eds.), Intelligence and cognition in special children: Perspectives on retardation,
Downloaded by [Harokopio University] at 14:07 22 April 2012

learning disabilities, and giftedness. Norwood, NJ: Ablex.

Brown, A. L., Armbruster, B. B., & Baker, L. (1984). The role of metacognition in reading and
studying. In J. Orasanu (Ed.), A decade of reading research: Implications for practice
(pp. 49-76). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Brown, A. L., Bransford, J. D., Ferrara, R. A., & Campione, J. C. (1983). Learning, remem-
bering, and understanding. In J. H. Flavell& E. M. Markman (Eds.), Handbook of child
psychology: Vol. 3. Cognitive development (pp. 77-166). New York: Wiley.
Canney, G., & Wiograd, P. (1979). Schemata for reading and reading comprehension per-
formance (Tech. Rep. No. 120). Urbana-Champaign: University of Illinois, Center for the
Study of Reading.
Clay, M. M. (1972). Reading: The patterning of complex behavior. Auckland, New Zealand:
Costa, A. L. (1986). Mediating the metacognitive. Educational Leadership, 42, 57-62.
Duffy, G., Roehler, L., Meloth, M., Vavrus, L., Book, C., Putnam, J., & Wesselman, R.
(1986). The relationship between explicit verbal explanation during reading skill instruction
and student awareness and achievement: A study of reading teacher effects. Reading Research
Quarterly, 21, 237-252.
Duffy, G., Roehler, L., & Rackliffe, G. (1986). How teachers' instructional talk influences
students' understanding of lesson content. Elementary School Journal, 87, 3-16.
Durkin, D. (1979). What classroom observations reveal about reading comprehension instruc-
tion. Reading Research Quarterly, 14, 481-533.
Durkin, D. (1981). Reading comprehension instruction in five basal reading series. Reading
Research Quarterly, 16, 515-544.
Fabricius, W. V., & Hagen, J. W. (1984). Use of causal attributions about recall performance to
assess metarnemory and predict strategicmemory behavior in young children. Developmental
Psychology, 20, 975-987.
Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive-
developmental inquiry. American Psychologist, 34, 907-91 1.
Flavell, J. H., & Wellman, H. M. (1977). Metamemory. In R. V. Kail & J. W. Hagen (Eds.),
Perspectives on the development of memory and cognition (pp. 3-34). Hillsdale, NJ:
Lawrence Erlbaum Associates, Inc.
Forrest-Pressley, D. L., & Waller, T. G. (1984). Metacognition, cognition, and reading. New
York: Springer-Verlag.
Garner, R. (1987). Metacognition and reading comprehension.Norwood, NJ: Ablex.
Johns, J. L. (1980). First graders' concepts about print. Reading Research Quarterly, 15,
Kennedy, B. A., & Miller, D. J. (1976). Persistent use of verbal rehearsal as a function of infor-
mation about its value. Child Development, 47,566-569.
Kobasigawa, A., liansom, C. C., & Holland, C. J. (1980). Children's knowledge about skim-
ming. Alberta Journal of Educational Research, 26, 169-182.
Lawson, M. (1985). Metacognition. In J. Biggs & J. Kirby (Eds.), Cognitive strategies. New
York: Academic
Lodico, M. G., Ghatala, E. S., Levin, J. R., Pressley, M., & Bell, J. A. (1983). The effects of
strategy-monitoringon children's selection of effective memory strategies. Journal of Experi-
mental Child Psychology, 35, 263-277.
Loper, A. B. (19800. Metacognitive development: Implications for cognitive training. Excep-
tional Educatioir Quarterly, 1, 1-18.
MacGinitie, W. H. (1978). Gates-MacGinitie Reading Tests (2nd ed.). Boston: Houghton
Myers, M., & Paris, S. G. (1978). Children's metacognitive knowledge about reading. Journalof
Educational Psychology, 70, 680-690.
Palincsar, A. S., & Brown, A. L. (1984). Reciprocal teaching of comprehension fostering and
Downloaded by [Harokopio University] at 14:07 22 April 2012

monitoring activfties. Cognition and Instruction, 1, 117- 175.

Paris, S. G. (1986). Teaching children to guide their reading and learning. In T. Raphael & R.
Reynolds (Eds.), Contexts of school-based literacy (pp. 115-130). New York: Random
Paris, S. G., Cross, D. R., Jacobs, J. E., Oka, E. R., DeBritto, A. M., & Saarnio, D. (1984,
April). Improving children's metacognition and reading comprehension with classroom in-
struction. Paper presented at the annual meeting of the American Educational Research Asso-
ciation, New Orleans.
Paris, S. G., Cross, D. R., & Lipson, M. Y. (1984). Informed strategies for learning: A program
to improve chikcken's reading awareness and comprehension. Journal of Education Psychol-
ogy, 76, 1239-1252.
Paris, S. G., & Jacobs, J. E. (1984). The benefits of informed instruction for children's reading
awareness and comprehension skills. Child Development, 55,2083-2093.
Paris, S . G., Jacobs, J. E., & Cross, D. R. (1987). Toward an individualistic psychology of ex-
ceptional children. In J. Borkowski & J. Day (Eds.), Intelligence and cognition in special
children: Comparative approaches to retardation, learning disabilities, and gvtedness.
Norwood, NJ: Ablex.
Paris, S. G., & Lindauer, B. K. (1982). The development of cognitive skills during childhood. In
B. Wolman (Ed.), Handbook of developmentalpsychology (pp. 333-349). Englewood Cliffs,
NJ: Prentice-Hd.
Paris, S. G., Lipson, M., & Wkson, K. K. (1983). Becoming a strat~egic reader. Contemporary
Educational Psychology, 8,293-316.
Paris, S. G., L Myers, M. (1981). Comprehensionmonitoring, memory and study strategies of
good and poor readers. Journal of Reading Behavior, 8, 5-22.
Paris, S. G., Newman, R. S., & McVey, K. A. (1982). Leaming the functional significance of
mnemonic actions: A microgenetic study of strategy acquisition. Journal of Experimental
Child Psychology, 34, 490-509.
Paris, S. G., & Oka, E. R. (1986). Children's reading strategies, metacognition, and motivation.
Developmental Review, 6, 25-56.
Pressley, M., Borkowski, J. G., & O'Sullivan, J. T. (1985). Children's metarnemory and the
teaching of memory strategies. In D. L. Forrest-Pressley,G. E. MacKinnon, & T. G. Waller
Gds.), Metacognition, cognition, and human performance. New York: Academic.
Reid, J. (1966). Learning to think about reading. Educational Research, 9, 56-62.

Ryan, E. B. (1981). Identifying and remediating failures in reading comprehension: Toward an

instructional approach for poor comprehenders. In T. G. Waller & G. E. MacKinnon (Eds.),
Advances in reading research (Vol. 3, pp. 223-261). New York: Academic.
Spring, H. T. (1985). Teacher decision making: A metacognitive approach. Reading Teacher,
39, 290-295.
Tei, E., & Stewart, 0. (1985). Effective studying from text: Applying metacognitive strategies.
Forum for Reading, 16,46-55.
Wellman, H. M. (1985). The origins of metacognition. In D. L. Forrest-Pressley, G. E.
MacKinnon, & T. G. Waller (Eds.), Metacognition, cognition, and human performance. New
York: Academic.
Wison, K. K., Bosky, A. B., Yochum, M. N., &Alvermann, D. E. (1984). An interview for as-
sessing students' perceptions of classroom reading tasks. Reading Teacher, 37, 346-352.
Downloaded by [Harokopio University] at 14:07 22 April 2012