EDUCATIONAL PSYCHOLOGIST, 40(4), 245–255

Copyright © 2005, Lawrence Erlbaum Associates, Inc.

LIN, SCHWARTZ,
TEACHERS’ METACOGNITION
HATANO

Toward Teachers’ Adaptive Metacognition

Xiaodong Lin
Teachers College
Columbia University

Daniel L. Schwartz
School of Education
Stanford University

Giyoo Hatano
Department of XXXXXXX
University of Air, Japan

In this article, we compare conventional uses of metacognition with the kinds of metacognition
required by the teaching profession. We discover that many of problems and tasks used in suc-
cessful metacognitive interventions tend to be reasonably well-defined problems of limited du-
ration, with known solutions. Teaching has unique qualities that differentiate it from many of
the tasks and environments that metacognitive interventions have supported. Teachers often
confront highly variable situations. This led us to believe that successful teaching can benefit
from what we call adaptive metacognition, which involves change to oneself and to one’s envi-
ronment, in response to a wide range of classroom social and instructional variability. We pres-
ent several examples to illustrate the nature of metacognition required by teachers and the chal-
lenges of helping teachers recognize situations that require adaptive metacognition. We
conclude the article by describing an approach, critical event-based instruction, which we have
recently developed to help teachers appreciate the need for metacognitive adaptation by seeing
the novelty in everyday recurrent classroom events.

Conventionally, metacognitive instruction is often used to ing the environment (and students), and this presents unique
help learners monitor and control the effectiveness and accu- challenges. As a result, we have been led to believe that suc-
racy of their own understanding and problem-solving behav- cessful teaching can benefit from what we call reflective ad-
iors in a particular subject matter (e.g., reading, writing, aptation (Lin, 2001a) or adaptive metacognition.
mathematics). In this article, we argue that conventional ap- Adaptive metacognition involves both the adaptation of
plications of metacognition fall short when it comes to the one’s self and one’s environment in response to a wide range
challenges teachers often face. Teaching has some unique of classroom variability. Many teachers tell us that each class
qualities that differentiate it from many of the tasks that is quite different, and each presents its own challenges and
metacognitive interventions have supported. For instance, charms. This situation is different from the stable environ-
the problems encountered in teaching can require days if not ments for which most metacognitive interventions have been
months to resolve. Moreover, what counts as a good solution designed. For repetitive tasks, like solving word problems, it
depends on clarifying and reconciling competing values, for is possible to provide students with metacognitive strategies
example, those of the teachers, the school district, and the that can eventually be routinized into problem solving
students. Finally, metacognitive research has largely been schemas. Teachers, on the other hand, confront highly vari-
about one’s individual thoughts, but teaching involves adapt- able situations from student to student and class to class. One
solution does not fill all, and teachers need metacognitive ap-
Correspondence should be addressed to Xiaodong Lin, Teachers Col- proaches that support adaptation and not just improved effi-
lege, Columbia University, 525 W. 120th St., Box 8, New York, NY 10027. ciency for completing recurrent cognitive tasks (Schwartz,
E-mail: [email protected] Bransford, & Sears, 2005).
246 LIN, SCHWARTZ, HATANO

We argue that one of the great challenges of metacognition egy training and creating social environments to support re-
for teachers is to help them recognize that apparently routine flective discourse (Lin, 2001b). Strategy-training studies
situations often have a number of hidden features that may originally used direct instruction, usually in a one-to-one sit-
make it quite different from what they believe, and therefore uation (Brown et al., 1983). More recently, there has been an
they require adaptation. This is because teaching is a deeply increase in the use of modeling and prompting. For example,
social act involving peers, students, and parents. Conse- Lin and Lehman (1999) used prompts to help preservice
quently, there are usually many hidden values that need to be teachers learn about strategies for controlling variables in a
made explicit so teachers can determine what type of adapta- computer-simulated science experiment. Some students re-
tion is acceptable (e.g., Dweck, 1999; Lin & Schwartz, 2003). ceived periodic questions about how well they were achiev-
Moreover, in contrast to the problems and tasks confronted in ing their goals. Other students received alternate prompts
typical metacognitive interventions, teachers must find a way (e.g., how are you feeling about yourself right now) or no
to effectively communicate and interact with people of differ- prompts at all. The students prompted to reflect on their goals
ent values to make their adaptations successful. Therefore, outperformed the other students on a subsequent transfer
teaching is always cross-cultural to some extent, because problem that did not include any prompting. Similar results
teachers and students rarely share the exactly same values and have been obtained in other studies with children and adults,
experiences. Problem solving that involves making changes to using tasks like the Tower-of-Hanoi, the Katona card prob-
one’s social environments and other people has its own set of lems, chess problems, human circulatory system problems,
metacognitive requirements. In this article, we report several mathematics problems, and computer-simulation problems
of our metacognitive studies that are designed to help teachers that included search (Safari Search), patterns (The Pond),
see through the surface of the classroom and begin the process and logic (Rocky’s Boots; e.g., Berardi-Coletta, Buyer,
of adapting to different social norms and situations that many Dominowski, & Rellinger, 1995; Chi, Deleeuw, Chiu, &
classes and students bring. LaVancher, 1994; King, 1992; Schoenfeld, 1987).
Our article consists of three sections. In the first section of Computers have been successful in modeling
the article, we describe some of the successes in the metacgonitive strategies. For example, they can make often
metacognitive literature and extract contextual characteris- tacit thinking processes overt so they become externalized
tics that seem responsible for their success including stable and accessible as objects of close reflection and evaluation
problems, environments, and values. In the second section, (Lin, Hmelo, Kinzer, & Secules, 1999). Graesser and
we present two examples to clarify how teaching can lack McNamara (2005) used computers as tutors to help students
these characteristics, and what that entails for adaptive generate questions (e.g., why, what-if, how) and develop
metacognition. In the last section, we describe some of the self-explanation strategies while reading text. Other re-
techniques we have developed to help teachers appreciate the searchers use technology to help students develop self-cor-
need for metacognitive adaptation, particularly in situations rection skills for problem solving (Mathan & Koedinger,
that appear routine on the surface level. One of our primary 2005), as well as use effective monitoring strategies in their
goals is to make sure that teachers do not apply assumptions online scientific inquiries (e.g., change their topic if they fail
of “routine” metacognition to teaching situations where to quickly find related information, Quintana, Zhang, &
adaptive metacognition is appropriate. For example, we pro- Krajcik, 2005). Computers have also been used to facilitate
vide them with recurrent problems, but we help them learn to collaborative problem solving and reflection. These include
avoid the tendency to schematize these recurrent problems so the uses of computers as advisors for science inquiry (White
they do not seek out the particular, but often hidden, details & Frederiksen, 2005) and creating an online forum that al-
that require adaptation. lows learners to share and critique each other’s understand-
ing (Computer-Supported Intentional Learning Environment
[CSILE], by Scardamalia & Bereiter, 1991; and Learning
Through Collaborative Visualization [Co Vis], by Pea,
CHARACTERISTICS OF TYPICAL
1994). In this work, the emphasis is on the creation of social
METACOGNITIVE INTERVENTIONS WITH
support for the adoption of effective metacognitive strate-
AND WITHOUT COMPUTERS
gies. For example, in the CSILE program, after being ex-
posed to reflection discourses in the classrooms, students
Cognitive Characteristics of Interventions
tended to develop the habits of using explanation strategies to
Metacognition, or the awareness and regulation of the pro- engage in metacognitive social conversations about the tasks
cess of one’s thinking, has been recognized as a critical ingre- and problems at hands. In a more automated version of social
dient to successful learning (Brown, Bransford, Ferrara, & support, Biswas, Schwartz, Leelawong, Vye, and TAG-V
Campione, 1983; Flavell, 1987; Hacker, Dunlosky, & (2005) had fifth-grade students teach a computer agent to
Graesser, 1998; Pressley, Etten, Yokoi, Freebern, & Meter, solve problems about the oxygen cycle in pond ecology. The
1998). Over the past decades, there have been two main ap- “teachable agent” initiated observations about how well it
proaches for teaching metacognitive monitoring skills: Strat- understood; for example, it would refuse to take a test, stating
 TEACHERS’ METACOGNITION 247

that the student had not asked it enough questions to make participants wisely refused to complete a task as given, be-
sure it understood. In a transfer test a month later, students cause they decided it was intractable.
learned about the nitrogen cycle on their own. Students who
worked with the metacognitive teachable agent learned more Shared values and goals. A third assumption of
during the transfer test than students who learned the oxygen many metacognitive interventions is that the participants and
cycle without metacognitive support from the agent. the instructors share common learning goals and values. We
Broadly speaking, these studies have focused on the fun- have not encountered studies that address issues where most
damental question of whether instruction in effective of the participants believe that successful problem solving
metacognitive strategies can facilitate learning and problem and deep comprehension are not their goal for learning. The
solving. The findings repeatedly suggest that people can be assumption seems to be that any given task should ultimately
very successful at metacognitive monitoring and that com- induce the same set of optimal activities from everybody in-
prehension and problem solving are usually improved, espe- volved. Hence, within the metacognitive literature, there is
cially when the utility and function of these effective minimal activity dedicated to illuminating cultural values
metacognitive strategies are made apparent. that might influence what constitutes a meaningful problem
or an acceptable solution. This might be because most of the
problems used are reasonably value free—there is little de-
Contextual Characteristics of Tasks Commonly bate about whether a problem should be addressed or what
Targeted by Metacognitive Interventions constitutes an optimal solution. Additionally, in most inter-
As we reviewed the metacognitive literature, we noticed that ventions, people engage in individualistic metacognitive pro-
the target tasks often involved underlying contextual features cesses in which the focus is on how they are thinking and pro-
that may have helped to make the metacognitive interven- gressing. Even when students are invited to monitor and
tions successful. These contextual features, however, may evaluate other people’s argumentations, as in many of the re-
not characterize many teaching situations; therefore, it is im- cent computer-supported collaborative learning programs,
portant to consider whether previously studied metacognitive the focus is often on other people’s thinking about the content
interventions are ideally designed for the contexts that teach- knowledge and tasks, rather than monitoring other people’s
ers confront. Here, we briefly review the contextual features values, goals, and the cultural contexts that contribute to the
of the tasks addressed by prior metacognitive training, in- formation of such values and goals.
cluding (a) well-defined problems with known optimal solu- In sum, successful metacognitive strategy interventions
tions; (b) stability of the learning environment; and (c) shared have chosen tasks that are reasonably well defined and value
values and goals. free; the environments for which they are preparing people
are fairly stable; the trainees and the instructors share com-
mon learning goals and values; and learners are responsive to
Well-defined problems. Many of the problems used in the need for methods of metacognition. As we describe next,
successful metacognitive interventions are well-defined these features are often missing in the reality of teaching
problems of limited duration. In most of the studies, partici- practice, which suggests that additional forms of
pants handle one problem or step at a time, and the problems metacognitive training might be useful.
have an optimal solution procedure. For instance, in our own
study with variable control problems, we used metacognitive
strategies to help people monitor their application of experi- TWO CASES: TEACHER METACOGNITION
mental designs that yield confound-free findings about insect AND THE CHALLENGES OF TRIGGERING
habitats. Because the problems have an optimal solution path ADAPTIVE METACOGNITION
and a natural segmentation, the metacognitive emphasis is on
monitoring and regulation during each step of problem solv- Teachers confront many problems that share characteristics
ing, rather than on the reflection that might occur in between with tasks found in the metacognitive literature. For example,
problem solving episodes. grading papers efficiently is a highly routine activity that
presents a challenging well-defined problem in a stable envi-
Stable learning environments. The learning environ- ronment of clear values, and teachers always have an eager-
ments for most metacognitive interventions are relatively sta- ness to learn more effective methods and how to avoid skim-
ble. The problems are usually identified for the students ming. At the same time, teaching introduces new
rather than arising naturally from their own practices, so metacognitive challenges, assuming that teachers do not treat
there is limited emphasis on problem finding. Plus, the set- teaching like following a recipe. Lampert (2001), for exam-
ting of problem solving and learning is often fixed, so that ple, pointed out how teaching problems are often ill-defined
there is little opportunity for adapting the environment or and involve a variety of values that may not be in complete
one’s range of solutions. For instance, we have yet to read a harmony. She explained that when she is teaching fifth-grade
positive metacognitive report about a situation in which the mathematics in an open classroom of discussion, she simul-
248 LIN, SCHWARTZ, HATANO

taneously has to teach the mathematics content, the need to an assignment of homework. She approached Jasper in the
complete assignments successfully, and the value of being same way, for example, by beginning with a lecture on how
civil and equitable to classmates. Teaching involves handling to proceed solving Jasper. The lecture and her highly sched-
many interacting factors simultaneously that often cannot be uled class structure did not work very well, because the Jas-
neatly decomposed and treated one at a time (see Dar- per video is more complex than bite-sized math problems.
ling-Hammond & Bransford, 2005). There is usually not a Jasper usually takes children several days to solve working in
single algorithm for achieving an optimal solution. Hence, teams, and there are different possible solutions. Students
teachers cannot metacognitively monitor whether they are disagreed with the teacher and with each other about what the
deviating from that algorithm. Instead, teachers often need to problems were and what were the best approaches to solve
reflect on their values and the consistency between their own the problems. The students also wanted to work in groups,
values and those of the other members in the community but they had not done group work on open-ended problems
(parents, students, principals) to guide them towards an ac- like this. Over time, many situations developed that were un-
ceptable solution. familiar to the teacher: Students competed over whose ideas
In this section, we present two cases. The first cases shows were correct within the group and who deserved credit; the
how the common contextual features that are found in many teacher had difficulty assessing student understanding with
typical metacognitive interventions disappear, particularly brief quizzes at the end of each class; and the teacher had dif-
when a teacher is introducing new curriculum to the students. ficulty leading class discussions when students offered many
This suggests that it may be worthwhile to entertain alterna- different possible solutions.
tive approaches to metacognition for teachers. The examples This situation denied the teacher three contextual features
in the second case illuminate some of the relatively unique that we identified as common to most metacognitive inter-
metacognitive challenges that teachers face, including the ventions. The problem of implementing the new curriculum
discovery and negotiation of conflicting values and practices. did not have a single, optimal solution. The classroom re-
These clarify the challenge of triggering adaptive sisted the usual stability of the teacher’s well-designed daily
metacognition in teachers’ everyday classroom practices. routine. The values of the students differed from the teacher,
and with each other. The demands of adaptation to the artifact
and its effects led the teacher to reflect on what she was doing
Case One: Teacher Adapting New Curriculum
and why. Much of her reflection had qualities often associ-
and How the Common Contextual Features That
ated with metacognition, including careful planning and
Are Present in Typical Metacognitive
monitoring of success. At the same time, there were some
Interventions Disappear
unique qualities to her adaptive metacognition.
Lin (2001a) conducted a study that illuminates how the con- One quality was that the teacher did most of her
text of teaching introduces task characteristics unlike those metacognition offline, when she was not caught up in the
found in many successful metacognitive interventions. The time-sensitive, complex, and novel issues that occurred during
setting of the study is not representative of all teaching, but it the class. After each class period, she engaged in intense
nicely brings to light the situation faced by many teachers, in self-questioning. A second quality was the character of this
a more modest fashion, when they choose to try new methods self-questioning. It was less about the “effectiveness” of her
of teaching. The study documented how a fifth-grade teacher solutions and more about her values and identity as a teacher.
in Hong Kong responded to the introduction of a new tech- For example, she worried that letting the students pursue their
nology artifact into her classroom—a 20-min, video-based own problem-solving inclinations would erode her authority
math problem, called the Adventures of Jasper Woodbury in class. She also wondered whether letting students work to-
(Cognition and Technology Group at Vanderbilt [CTGV], gether to sort things out before her lectures would indicate to
1997). The teacher and the school principle were interested them that she did not know the material and was not a good
in finding new ways to use technology in the school to help teacher. A third quality was that she had to think of how to
spruce up the school’s reputation. Jasper problems use a story adapt the classroom and the Jasper problem so that it met the
narrative complete with characters to present a complex emerging values of the students, yet still allowed her to com-
planning task that depends on a number of related mathemat- municate her values and knowledge about mathematics. As
ical calculations and numerical information embedded in the she pointed out during an interview, the most difficult thing for
narrative. At first the teacher attempted to fit the novel Jasper her as a teacher was how to handle different goals from the
into her usual routines, but as she quickly found out, it is hard class and to decide the valuable actions to take. For instance,
to embrace novelty without adapting. This was when the she had to ask herself several times about her goals as a teacher,
unique qualities of her metacognition began to appear. what a good teacher should do in the difficult situations she en-
When the teacher started, she worked to maintain her countered, and how to make sure that students learn.
usual classroom structure. This structure flowed through a As a consequence of her reflections, the Hong Kong
set sequence of activities every day: a review of homework, teacher made a series of justified decisions to adapt some of
an introductory lecture, a period of practice, a brief quiz, and the affordances of the new artifact (focus on subproblems so
 TEACHERS’ METACOGNITION 249

students can deal with smaller problems one at a time) and re- did not do the homework assignment. Those students who
ject others. She also adapted her role as a teacher by provid- listed learning characteristics tended to say “The boy or the
ing lectures on a need-to-know basis instead of using girl would not learn.” The students who listed behavioral
prelecturing at the start of each class period. She let go of her characteristics tended to say “The boy or the girl would get in
desire to give in-class memorization quizzes at the end of trouble.”
each day. As a result of these changes, she also discovered We do not intend to claim that these results are representa-
new insights about herself and her students as revealed in her tive of all public or private schools—this will require a much
comments, “students were a lot more creative than I have larger sample of schools. Rather, the importance involves the
ever imagined” or “I did not know that I was kind of control- relation between the students’ ideal students and the teachers’
ling actually.” ideal students. Figure 1 presents the percentage of students
The level of metacognition and adaptation was intense for and teachers from each nation and school type who men-
this teacher as she began the transition from a routine to an tioned learning and/or behavioral properties for their ideal
adaptive teacher (Hatano & Inagaki, 1986). Metacognition students. The key pattern to notice is how the American pub-
served as a mechanism for problem finding, for setting adap- lic school results differ from the other school types. In partic-
tive goals, for identity building, and for value clarification. ular, one should notice that the teachers in the American pub-
These uses of metacognition seem quite different from what lic schools, and not just the students, emphasized behavioral
appears in moment-to-moment logical problem solving and qualities at a very high rate.
is not normally covered in the strategies that have been found Once the implicit values and implied practices associated
in conventional metacognitive interventions that use scripted with those values were brought to the teachers’ attention, this
problems. Characterizing successful metacognitive strate- triggered adaptive metacognition. They first expressed sur-
gies for adaptive situations appears to be a fruitful area for prise at their students’ complete emphasis on behavior rela-
further research. tive to other schools. They also noticed that they too were
emphasizing behavior more than teachers from other
schools. The teachers began to reflect on their own assump-
Case Two: The Challenge of Triggering Adaptive
tions about schooling, what they had been doing, how to im-
Metacognition in Everyday Classrooms
prove their own practice and reintroduce learning as a central
Unlike our first case involving a novel curriculum, teachers part of the students’ ideal. This finding suggests that one im-
may not recognize that something needs adaptation in more portant component of metacognitive training for teachers
subtle situations. Teachers already possess a great deal of ex- should help them see the values that they and their students
perience and knowledge about teaching (after all, they were hold so they can begin to reflect on whether changes are ap-
students for at least 16 years). This leads to the problem of as- propriate. As this example suggests, providing contrasting
similation, where people understand things in terms of famil- cases is one way to help.
iar habits of seeing and doing things instead of recognizing
there is something that requires metacognitive reflection Possibilities of being dismissive of new models for
(Bransford, Brown, & Cocking, 1999). We provide two ex- doing things. Our second example involves the challenge
amples that highlight two variants on this problem. One is the of recognizing the possibility of doing things in new ways.
problem of not recognizing that a situation could use im- Like the Hong Kong teacher, this requires receptiveness to
provement, and the second is being dismissive of new models new ways of doing things. It also requires seeing that there is
for doing things.

The difficulties in recognizing situations that warrant

adaptation. Our first example comes from a study with
280 American and Chinese fifth-grade students and their
teachers in both public and private schools (key schools in
China; Lin & Schwartz, 2003). The students and teachers
were asked to list the five most important characteristics of
an ideal student. The Chinese students in both public and pri-
vate schools, and the American students in private schools,
emphasized learning-oriented characteristics (e.g., explains
and understands deeply, knows when one makes mistakes).
In contrast, the American public school students were con-
cerned mostly with good behavior in class (e.g., does not
fight, sits still during lectures). To validate our ad hoc catego- FIGURE 1 Percentage of students and teachers from United States
ries of learning versus behavioral characteristics, we subse- and Chinese private and public schools who mentioned learning or
quently asked students what would happen if a boy or a girl behavioral properties for their ideal students.
250 LIN, SCHWARTZ, HATANO

something new and of potential value in other ways of doing had not completed the intervention, the role of detached ob-
things, rather than being dismissive. People have a strong server is not ideal for triggering thoughts about personal
tendency to assimilate potentially new situations to their rou- adaptation. This finding is echoed by studies of teachers
tine ways of living. Instead of seeing what is new and adapt- learning from multimedia cases (Hewitt, Pedretti, Bencze,
ing, they often reject what is new (Cohen, 1991; Lin, 2001a). Vailiancourt, & Yoon, 2003). In these studies, preservice
Presumably, an adaptive metacognitive teacher would keep teachers reflected more deeply about their decision-making
an eye open for what was new so he or she could begin the processes when they were asked to make personal choices
process of reflection and adaptation. But this also requires about instructional situations compared to when they simply
both the willingness and the ability to search for what is novel watched and analyzed somebody else’s teaching. Thus, we
beneath the surface level. We have encountered many situa- think it is valuable for promoting metacognition to put people
tions in which teachers’ routinized beliefs prevent them from in a position in which they have to make decisions about a sit-
seeing and being receptive to new ideas. It is only when we uation of their own actions. This seems obvious. Traditional
have increased the variability by introducing new technolo- metacognitive strategy training would hardly work if stu-
gies or by providing explicit contrasts that teachers have had dents only observed other people and never had to use
a chance to see the need for reflection and adaptation. metacognition coupled with active problem solving. At the
In one study, we asked in-service, American teachers to same time, teacher training often asks teachers to reflect on
view videotapes of Chinese teachers (Lin & Schwartz, 2003). other people’s cases without making decisions about their
The American teachers tended to dismiss the instructional own potential actions in the same situation. So, perhaps the
techniques used by the Chinese teachers. For example, one point is not that obvious.
American teacher stated, “These methods are great. But, that
can only happen in Asia. We cannot do that in America.” An-
other stated, “Typical Chinese. Too strict.” The teachers TOWARD ADAPTIVE METACOGNITION
tended to discount the video of alternative teaching practices,
because it simply represented another culture. As a conse- In this section, we describe an approach—Critical Event In-
quence, they did not reflect on how they might adapt their struction—that we have developed to help teachers appreci-
own practices. ate the need for metacognitive adaptation, particularly in sit-
These results can be contrasted with American teachers uations that appear routine on the surface level. One of our
who had completed an intervention to make them more re- primary goals is to prepare preservice teachers for common
ceptive to new ways of doing things. The teachers who com- events that are likely to appear in the course of their teaching.
pleted the intervention were much more reflective about the At the same time, we want to help them see the novelty in
sample of Chinese teaching, and they related it to their own these familiar events so that they do not apply assumptions of
classroom practices. For example, one teacher saw the appar- “routine” metacognition to teaching situations where adap-
ent strictness in the Chinese methods in another light. She tive metacognition is more appropriate. For example, in our
stated, “Maybe I have let my standards drop too low.” critical-event-based instruction, we provide teachers with ex-
The intervention put American and Chinese teachers to- amples of recurrent problem situations (e.g., poor attitudes
gether online. It used a three-dimensional interactive virtual toward school assignments). We then help teachers learn
learning space (VLS) with an activity about conducting ex- ways to avoid the tendency to schematize these recurrent
periments to find out about an insect’s preferred habitat. We problems so they do not seek out the particular, but often hid-
asked each intervention teacher to coteach a lesson in the den, features that require differentiated solutions through re-
VLS with a Chinese teacher. We also arranged to have six flection and adaptation. We do so by having people with con-
students (three from America and three from China) in the trasting values, goals, and experiences look at the same
VLS at the same time, so it was a real lesson with implica- recurrent event. They offer their unique perspectives and dis-
tions for student learning. The teachers and students ap- cuss what additional information they would like to help
peared as avatars in the VLS, and they moved from location solve the problem and determine what actions they might
to location to conduct their experiments. take. Our goal of asking people to search for additional infor-
The VLS intervention is not easily scaled up for working mation is to help people learn that all situations do not have a
with many teachers (there is a great deal of logistical over- one-size-fits-all solution, and to attune themselves to impor-
head in arranging times across nations and so forth). At the tant types of information that they should seek. This adds
same time, it had a feature that we think was particularly im- novelty and complexity to what appears as simple and famil-
portant for supporting the onset of adaptive metacognition iar problem situations. By doing so, we hope to turn the re-
when the teachers subsequently watched the sample videos. current problem situations into critical event-based situations
During the intervention, the teachers had been involved in the where the novelty and hidden properties are made explicit
process of making decisions about their own practices rather along with the need to seek additional information. It is our
than observing other people and commenting. As indicated hope that teachers are able to develop the habits of seeing
by the dismissive responses of the American teachers who past their taken-for-granted view of the recurrent classroom
 TEACHERS’ METACOGNITION 251

problems to begin the process of seeking for other sources of

information in new situations and adapting to different values
and social norms that many classes and students bring.
We developed a multimedia-based learning shell, criti-
cal-event based learning environment (CEBLE) that contains
a series of video vignettes to capture recurrent and authentic
classroom events or challenges. For instance, one event is
about students’ negative attitudes toward science assign-
ments. Other instances present disengagement during group
learning, and still others focus on students’ difficulties in en-
gaging an inquiry approach to learning math, science, and
history. Here is a partial transcript of the critical-event video
for negative science attitudes. The critical event comes from
videos created for the CTGV Challenge Series (CTGV, 2000;
Vye et al., 1998): FIGURE 2 Learning cycle for the critical event-based instructional
learning environment.
A student named Billy and his high school class team
just returned from monitoring the water quality of a lo- the problem. Ideally, this can become a habit of mind for
cal river, and they were ready to turn in their report. A teachers. We also ask teachers to adopt a perspective that
dialogue between Billy and his teammate Sally is pre- contrasts with other perspectives on how to solve the prob-
sented: lem. This exemplifies that any given adaptation depends on
one’s values; hopefully, it can help teachers clarify what val-
Billy: That’s it, Sally. We are done. ues they deem most important. Finally, CEBLE offers teach-
Sally: I am not so sure, Billy. I think we should recheck our ers opportunities to interact with people of different back-
conclusions. We are saying the river isn’t polluted, grounds, values, and goals through its multiple perspectives,
and if we were wrong. It’s like we are letting all the and through sharing thoughts with other members of the
fish die. communities. These social interactions allow teachers to gain
Billy: Enough, Sally. It’s just a school assignment and our the first-hand experience that the same problem can have
report is 5-pages long, single-spaced! We will pass many different solutions depending on the goals and values
for sure. people bring with them.
We conducted a study with 30 preservice teachers to see if
The CEBLE shell uses a cycle evolved from a STAR.Leg- CEBLE holds any instructional potential. We presented the
acy software shell (see Schwartz, Lin, Brothy, & Bransford, negative attitude scenario to all the participants. They were
1999). The shell includes (a) meet the event; (b) generate re- then randomly assigned to two different groups. The adaptive
sponses to questions including “Did you notice anything new metacognition group completed the aforementioned cycle.
in this event?” “Do you need any additional information in The problem-solving group generated solutions without de-
order to solve the problem embedded in the event? If so, what ciding what additional information is needed and they did not
is it? If not, why not?”; (c) listen to multiple perspectives of- see the multiple perspectives. However, they were permitted
fered by people of different backgrounds, goals, and values; to talk with another. The conditions were meant to maximize
(d) act on the perspectives by generating solutions for se- the chances of seeing whether CEBLE leads to adaptive out-
lected perspectives—the decision about which perspective to comes, rather than showing that CEBLE is better than all
act on depends on the individual teachers’ needs and their other forms of teacher training. We also administered a trans-
values; (e) reflect on the effectiveness of one’s solutions and fer posttest to see if the students learned differently. For the
share their choices of perspectives, solutions and legacies posttest, both groups saw another critical event about student
with other members in the community who explore a similar disengagement during group activities. Everyone received a
event and topic. In this way, legacies for learning from recur- worksheet that asked them to analyze the event and generate
rent events grow over time with multiple uses (see Figure 2 three solutions, and to state whether they would have liked
for CEBLE learning cycle). any additional information to make decisions, and if so, what
We believe CEBLE has a potential for supporting adaptive kind of questions they would ask to get the information. All
metacognition, because it not only provides experiences with participants said that these events were familiar and that they
specific and recurrent events, but it also asks teachers to think had encountered both kinds of events during their student
past the application of their schematic responses. We ask teaching or other teaching experiences.
teachers what additional information they need before solv- We coded the student solutions for both the original and
ing the problem, so they consider potential sources of hidden transfer case along two dimensions. One dimension captured
variability in the situation and can better locate the source of whether a student proposed distinctly different solutions.
252 LIN, SCHWARTZ, HATANO

The other dimension captured the specificity of the solutions. task, where there was no special support, is very heartening
The interrater reliability was 85% for the negative attitude and warrants further research into whether these effects hold
event and 90% for the transfer, disengagement event. up after a longer delay or in the teachers’ own classes.
A student received credit for different solutions when all On the transfer posttest, we also asked the participants
the solutions depended on different information sources. The from both conditions if they needed more information, and if
following three solutions yielded credit for being different: so what would they like to know. Figure 4 shows the differ-
ences between the conditions. Twice as many participants in
Solution 1: “I would talk to Billy to see if my assignment is the adaptive condition said they would like more information
not interesting and motivating to him … then I compared to the problem-solving condition. This should not
would … ” be too surprising, because they had just completed an inter-
Solution 2: “I will think about whether I give them (the stu- vention that emphasized asking for more information. What
dents) clear criteria for how their report will be is more impressive is the type of questions that people in both
evaluated. If not, I would … ” conditions generated. Of those participants who generated
Solution 3: “Have a talk with Billy’s parents to see how information questions in both conditions, participants in the
they think about science … ” adaptive condition asked metacognitive How/Why and
If/Then questions twice as often as the problem-solving par-
A student did not receive credit for differentiation when ticipants, who tended to ask “What” questions or questions
all the solutions were based on the same source of informa- require only “Yes or No” answers (see King, 1992; Lin &
tion. The following examples all depend on the information Lehman, 1999). An example of a How/Why question is
that Billy had not organized his experiments well enough. “How did the teacher arrange the group work, based on abil-
ity or randomly?” and “Why did the teacher decide to use
Solution 1: ”Set up a scientifically accountable method to group work rather than use individual seat work?” An exam-
help Billy report the results … ” ple of a “What” question is “ What are the details of the activ-
Solution 2: “Make him repeat the experiment 3 times for ity in the group?”
the same results … ” In debriefing, we asked participants to rate the usefulness
Solution 3: “Conference w/both students until they run the of the experience of working with critical events on a 5-point
experiments enough times to reach a conclu- scale ranging from 1 (least useful) to 5 (most useful) and ex-
sion” plain their rating. The average rating for the adaptive group
was 4, whereas the average rating for the problem-solving
For the second coding dimension of specificity, students group was 2.5. Their reasons reflected our suppositions about
received credit if all the solutions provided procedural de- the tendency of people to assimilate events. For example, one
tails. An example of a specific solution is, “I will conduct student from the problem-solving group wrote, “You guys
clinical interviews with Billy and find out why he thinks that are making a big deal out of nothing. Disengagement during
getting C or B is enough.” A similar but vague solution was, group work is so common. There is nothing new about it.”
“Be sure to ask for prior knowledge.” Moreover, the problem-solving participants stated that they
For both events, the groups generated equal numbers of so- had trouble generating alternative solutions on the posttest,
lutions. However, Figure 3 shows that the individuals in the because there was nothing new so they kept saying the same
adaptive metacognition condition generated more distinct so- thing in different ways. In contrast, the adaptive
lutions that were also more specific. The fact that the differ- metacognition group seemed to appreciate the value of look-
ence between the conditions does not diminish on the transfer ing past the surface. As one adaptive metacognition partici-

FIGURE 3 Percentage of participants from CEBLE and comparison groups who were able to generate both differentiated and specific solutions to both
instructional and transfer events.
 TEACHERS’ METACOGNITION 253

the kinds of adaptive metacognition needed by the teachers.

First, instead of assuming that a given task should induce al-
most the same activity from everybody (a series of actions
aimed at a taken-for-granted goal), we emphasize that varied
teaching or learning goals can be set, and thus different activ-
ities can be derived for apparently the same educational tasks
and materials, depending on the values of the participants.
Whereas has emphasized whether the participant has an ef-
fective metacognitive strategy to pursue the activity or how to
teach that strategy, we propose that clarifying the potential
goals is the initial critical step of adaptive metacognition,
though choosing or inventing an effective metacognitive
strategy to achieve a selected goal is also an important com-
FIGURE 4 Percentage of participants from the CEBLE and the
comparison groups who expressed need for additional information
ponent (see Pintrich, 2000).
for the transfer event (Event 2) and who were able to generate Second, we propose that planning appropriate activities
metacognitive questions. and observing activities engaged by others with different val-
ues or sociocultural backgrounds give a very good opportunity
pant wrote, “The idea of asking for additional information to for participants to reflect on their own goals and also learn how
a problem that is so familiar is quite interesting to me.” Many best to set goals under given constraints. We believe that the
commented that the multiple perspectives were very helpful traditional way of metacognitive training has been unduly
because they let them see new insights to an event. cognitive in its assumption of a common set of goals. For ex-
All told, the results suggest that this instructional ap- ample, an attempt to directly teach strategies for solving prob-
proach is promising. People tend to be more reflective lems may be effective for monitoring and controlling learners’
around unexpected situations (Flavell, 1979). CEBLE in- behaviors, but only when the trainer and the trainee share the
creases the “unexpectedness” of “typical” events by having same goal. Another training procedure is to let a participant
people with different experiences and backgrounds offer observe others’ problem solving, but its effectiveness presup-
their perspectives to an apparently familiar situation. This poses that the observer can see the problem from the perspec-
helped people consider more specific alternatives and want tive of the solver’s metacognition. We would suggest that these
more information during training; in turn, this helped them training procedures can be used more successfully by consid-
avoid dismissing a second event as typical. CEBLE enhanced ering participants’ goals explicitly.
the benefit of these multiple perspectives by first asking par- Third, we claim that social, collaborative situations might
ticipants to make their own thoughts and likely actions ex- be used more often in the assessment and training of
plicit before the arrival of multiple perspectives. Too often, metacognition. Many important situations for teachers are
experts’ perspectives are used as modeled behaviors for the highly social in nature. So, rather than solving bookish prob-
teacher to follow, rather than being used as resources for lems, they need to think about how to arrange social matters.
teachers to self-assess their habitual ways of seeing and do- We believe that the emphasis of metacognitive research on
ing things. Our hypothesis is that the timing for providing individual’s “inward looking” processes could be broadened
multiple perspectives matters (Schwartz & Bransford, 1999). to include a consideration of social behaviors designed to al-
CEBLE was also designed to help students hear one an- ter one’s social environment. It is important, however, to
other’s perspective more clearly, by providing the models of avoid an optimistic assumption that any interactions and cul-
multiple perspectives before they had a discussion. Many of tural exposure to new information or new forms of practices
the participants in the adaptive metacognition condition lead to productive reflection (Lin & Schwartz, 2003). When
wrote of their surprise to discover that the same event, espe- people are in the role of a detached observer, they tend to be
cially an event that almost everybody has had experiences less analytic and reflective about their own and other peo-
with, was not reacted to in the same way by other members of ple’s teaching. This finding was echoed by studies of teach-
the class, which led to reflective discussions afterward. ers learning from multimedia cases (Hewitt et al., 2003).
Fourth and finally, we believe that computers can help
teachers’ metacognition by giving them both (a) a set of ex-
periences with specific and recurrent events where personal
CONCLUSION decision making is required and (b) opportunities to appreci-
ate what other sources of information are important to con-
We conclude our article by summarizing the arguments and sider and to reflect on. This kind of metacognition is useful
findings presented so far in terms of suggestions for the because, in many situations, especially in complex teaching
metacognitive development of teachers. Four sets of issues situations, teachers often lack background information to
arise by contrasting conventional metacognitive tasks with know what solution can be sought and which strategies will
254 LIN, SCHWARTZ, HATANO

work. People usually do not detect absence of the important Bransford, J. D., & Darling-Hammond, L. (2005). Preparing teachers for a
changing world: what teachers should learn and be able to do. San Fran-
information very well. Therefore, it is important to develop
cisco, CA: Jossey-Bass.
the habits of gathering more information so that teachers can Dweck, C. S. (1999). Self-theories: Their role in motivation, personality,
determine what strategies and solution to apply. and development. Ann Arbor, MI: Edwards Brothers.
Our approach to metacognitive learning tries to integrate Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area
both specific cognitive skills (e.g., making decisions for spe- of cognitive-developmental inquiry. American Psychologist, 34,
906–911.
cific problem situations) and general adaptive and social
Flavell, J. H. (1987). Speculations about the nature and development of
abilities (e.g., reflecting deeply on what types of information metacognition. In Weinert, F. E., & Kluwe, R. H. (Eds.), Metacognition,
are needed or useful to make these types of decisions). Our motivation, and understanding. Hillsdale, NJ: Lawrence Erlbaum Associ-
hope is that the example of critical-event-based instruction ates, Inc.
provides one concrete instance that can help the field move in Graesser, R. C., & McNamara, D. (2005). Scaffolding deep comprehension
strategies through AutoTutor and iSTART. Educational Psychologist, 40,
the direction of increasingly integrative approaches toward 225–234.
metacognitive interventions that use both cognitive and so- Hacker, D. J., Dunlosky, J., & Graesser, A. C. (1998). Metacognition in edu-
cial information to help teachers develop proactive cational theory and practice. Hillsdale, NJ: Lawrence Erlbaum Associ-
metacognitive capabilities. ates, Inc.
Hatano, G., & Inagaki, K. (1986). Two courses of expertise. In H. A. H.
Stevenson, & K. Hakuta (Ed.), Child development and education in Japan
(pp. 262–272). New York: Freeman.
Hewitt, J., Pedretti, E., Bencze, L., Vaillancourt, B. D., & Yoon, S. (2003).
ACKNOWLEDGMENT New applications for multimedia cases: Promoting reflective practice in
preservice teacher education. Journal of Technology and Teacher Educa-
The preparation of this article and some of the research stud- tion, 11, 483–500.
King, A. (1992). Facilitating elaborative learning through guided stu-
ies reported in this article were made possible by the Carne- dent-generated questioning. Educational Psychologist, 27(1), 111–126.
gie Scholar 2003 to Xiaodong Lin, Carnegie Corporation of Lampert, M. (2001). Teaching Problems and the Problems in Teaching. New
New York City. The corporation does not take responsibility Haven, CT: Yale University Press.
for any views expressed in this article. Lin, X. D. (2001a). Reflective adaptation of a technology artifact: A case
We thank John Bransford, Allan Collins, Chuck Kinzer, study of classroom change. Cognition & Instruction, 19, 395–440.
Lin, X. D. (2001b). Designing metacognitive activities. Educational Tech-
Nancy Vye, Hank Clark, and other members from the group nology Research & Development, 49(2), 23–40
formerly known as Cognition Technology Group at Lin, X. D., Hmelo, C., Kinzer, C., & Secules, T. (1999). Designing technol-
Vanderbilt. ogy to support reflection. Educational Technology Research & Develop-
ment, 47(3), 43–62.
Lin, X. D., & Lehman, J. (1999). Supporting learning of variable control in a
computer-based biology environment: effects of prompting college stu-
dents to reflect on their own thinking. Journal of Research In Science
REFERENCES Teaching, 36(7), 1–22.
Lin, X. D., & Schwartz, D. L. (2003). Reflection at the crossroad of cultures.
Berardi-Coletta, B., Buyer, L. S., Dominowski, R. L., & Rellinger, E. R. Mind, Culture & Activities, 10(1), 9–25.
(1995). Metacognition and problem solving: A process-oriented ap- Mathan, S., & Koedinger, K. R. (2005). Benefits of Tutoring Error Detection
proach. Journal of Experimental Psychology, 21(1), 205–223. and Self-Correction Skills. Educational Psychologist, 40, 257–265.
Biswas, G., Schwartz, D. L., Leelawong, K., Vye, N., & TAG-V. (2005). Pea, R. D. (1994). Seeing what we build together: Distributed multimedia
Learning by teaching: A new agent paradigm for educational software. learning environments for transformative communications. The Journal
Applied Artificial Intelligence, 19, 363–392. of the Learning Sciences, 3(3), 285–301.
Bransford, J. D., Brown, A. L., & Cocking, R. R. (1999). How people learn: Pintrich, P. R. (2000). The role of goal orientation in self-regulated learning.
Brain, mind, experience, and school. Washington, DC: National Acad- In M. Boekaerts, P. Pintrich, & M. Zeidner (Eds.), Handbook of self-regu-
emy. lation (pp. 451–502). San Diego, CA: Academic.
Brown, A. L., Bransford, J. D., Ferrara, R. A., & Campione, J. C. (1983). Pressley, M., Etten, S. V., Yokoi, L., Freebern, G., & Meter, P. V. (1998). The
Learning, remembering, and understanding. In J. H. Flavell & E. M. metacognition of college studentship: A grounded theory approach. In H.
Markman (Eds.), Handbook of child psychology: Vol. 3. Cognitive devel- Dunlosky & A. Graesser (Eds.), Metacognition in educational theory and
opment (4th ed., pp. 77–166). New York: Wiley. practice (pp. 347–367). Hillsdale, NJ: Lawrence Erlbaum Associates,
Chi, M. T. H., DeLeeuw, N., Chiu, M. H., & LaVancher, C. (1994). Eliciting Inc.
self-explanations improves understanding. Cognitive Science, 18, Quintana, C., Zhang, M., & Krajcik, J. (2005). Scaffolded software environ-
439–477. ments for supporting metacognitive aspects of online inquiry. Educa-
Cognition and Technology Group at Vanderbilt. (1997). The Jasper project: tional Psychologist, 40, 235–244.
Lessons in curriculum, instruction, assessment, and professional develop- Scardamalia, M., & Bereiter, C. (1991). Higher levels of agency for children
ment. Mahwah, NJ: Lawrence Erlbaum Associates, Inc. in knowledge building: A challenge for the design of new knowledge me-
Cognition and Technology Group at Vanderbilt. (2000). Adventures in an- dia. Journal of the Learning Sciences, 1, 37–68.
chored instruction: Lessons from beyond the ivory tower. Advances in In- Schoenfeld, A. H. (1987). What’s all the fuss about metacognition? In A. H.
structional Psychology (Volume V. pp. 35–100). Mahwah, NJ: Lawrence Schoenfeld (Ed.), Cognitive science and mathematics education (pp.
Erlbaum Associates, Inc. 189–215). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Cohen, D. K. (1991). Revolution in one classroom. American Educator, Schwartz, D. L., & Bransford, J. D. (1998). A time for telling. Cognition &
15(2), 16–23, 44–48. Instruction, 16, 475–522.
 TEACHERS’ METACOGNITION 255
Schwartz, D. L., Bransford, J. D., & Sears, D. L. (2005). Efficiency and in- Vye, N. J., Schwartz, D. L., Bransford, J. D., Barron, B. J., Zech, L., & Cog-
novation in transfer. In J. Mestre (Ed.), Transfer of learning from a mod- nition and Technology Group at Vanderbilt. (1998). SMART environ-
ern multidisciplinary perspective (pp. 1–51). CT: Information Age. ments that support monitoring, reflection, and revision. In D. Hacker, J.
Schwartz, D. L., Lin, X., Brophy, S., & Bransford, J. D. (1999). Towards the Dunlosky, & A. Graesser (Eds.), Metacognition in educational theory and
development of flexibly adaptive instructional design. In C. M. Reigeluth practice (pp. 305–346). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
(Ed.), Instructional-design theories and models: A new paradigm of in- White, B., & Frederiksen, J. (2005). Cognitive models and instructional en-
structional theory (Vol. 2, pp. 183–213). Mahwah, NJ: Lawrence vironments that foster young learners’ metacognitive development. Edu-
Erlbaum Associates, Inc. cational Psychologist, 40, 211–223.