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

net/publication/226053277

Problem-Based Learning: What and How Do Students

Learn?

Article  in  Educational Psychology Review · January 2004

DOI: 10.1023/B:EDPR.0000034022.16470.f3

CITATIONS READS

2,800 53,024

1 author:

Cindy Hmelo-Silver
Indiana University Bloomington
121 PUBLICATIONS   7,634 CITATIONS   

SEE PROFILE

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

Meta-Synthesis of Computer Supported Collaborative Learning in STEM education View project

HOWARD View project

All content following this page was uploaded by Cindy Hmelo-Silver on 13 March 2014.

The user has requested enhancement of the downloaded file.

P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Educational Psychology Review, Vol. 16, No. 3, September 2004 (°

C 2004)

Problem-Based Learning: What and How Do

Students Learn?
Cindy E. Hmelo-Silver1,2

Problem-based approaches to learning have a long history of advocating

experience-based education. Psychological research and theory suggests that
by having students learn through the experience of solving problems, they can
learn both content and thinking strategies. Problem-based learning (PBL) is
an instructional method in which students learn through facilitated problem
solving. In PBL, student learning centers on a complex problem that does
not have a single correct answer. Students work in collaborative groups to
identify what they need to learn in order to solve a problem. They engage
in self-directed learning (SDL) and then apply their new knowledge to the
problem and reflect on what they learned and the effectiveness of the strate-
gies employed. The teacher acts to facilitate the learning process rather than
to provide knowledge. The goals of PBL include helping students develop 1)
flexible knowledge, 2) effective problem-solving skills, 3) SDL skills, 4) effec-
tive collaboration skills, and 5) intrinsic motivation. This article discusses the
nature of learning in PBL and examines the empirical evidence supporting
it. There is considerable research on the first 3 goals of PBL but little on the
last 2. Moreover, minimal research has been conducted outside medical and
gifted education. Understanding how these goals are achieved with less skilled
learners is an important part of a research agenda for PBL. The evidence
suggests that PBL is an instructional approach that offers the potential to help
students develop flexible understanding and lifelong learning skills.
KEY WORDS: problem-based learning; constructivist learning environments; learning pro-
cesses; problem solving; self-directed learning.

1 Department of Educational Psychology, Rutgers, The State University of New Jersey, New
Brunswick, New Jersey.
2 Correspondence should be addressed to Cindy E. Hmelo-Silver, 10 Seminary Place, New
Brunswick, New Jersey 08901-1183; e-mail: chmelo@rci.rutgers.edu.

235

1040-726X/04/0900-0235/0 °
C 2004 Plenum Publishing Corporation
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

236 Hmelo-Silver

Problem-based approaches to learning have a long history. They are

one of many instructional approaches that situate learning in a meaning-
ful task, such as case-based instruction and project-based learning. In the
traditions of Kilpatrick (1918, 1921) and Dewey (1938), these approaches
argue for the importance of practical experience in learning. Problem-based
learning (PBL) is part of this tradition of meaningful, experiential learning.
In PBL, students learn by solving problems and reflecting on their experi-
ences (Barrows and Tamblyn, 1980). PBL is well suited to helping students
become active learners because it situates learning in real-world problems
and makes students responsible for their learning. It has a dual emphasis on
helping learners develop strategies and construct knowledge (Cognition and
Technology Group at Vanderbilt [CTGV], 1997; Collins et al., 1989; Hmelo
and Ferrari, 1997; Kolodner et al., 1996).
A review of PBL is timely because issues of flexible thinking and lifelong
learning have come to the fore in discussions of classroom reform (Bransford
et al., 2000; Greeno et al., 1996). PBL is of increasing interest to K–16 ed-
ucators as demonstrated by widespread publication of books written about
PBL (e.g., Duch et al., 2001; Torp and Sage, 2002). Educators are interested
in PBL because of its emphasis on active, transferable learning and its poten-
tial for motivating students. This article first describes PBL and distinguishes
it from other experiential approaches to learning. Second, it discusses the
goals of PBL. Third, the PBL tutorial process is discussed in detail. Next, the
article examines what we have learned about PBL. In particular, this sec-
tion examines the research regarding the goals of PBL. Finally, the article
discusses the research limitations and avenues for the future.

PBL AND OTHER EXPERIENTIAL APPROACHES

PBL is focused, experiential learning organized around the investiga-

tion, explanation, and resolution of meaningful problems (Barrows, 2000;
Torp and Sage, 2002). In PBL, students work in small collaborative groups
and learn what they need to know in order to solve a problem. The teacher
acts as a facilitator to guide student learning through the learning cycle de-
picted in Fig. 1. In this cycle, also known as the PBL tutorial process, the
students are presented with a problem scenario. They formulate and ana-
lyze the problem by identifying the relevant facts from the scenario. This
fact-identification step helps students represent the problem. As students
understand the problem better, they generate hypotheses about possible
solutions. An important part of this cycle is identifying knowledge deficien-
cies relative to the problem. These knowledge deficiencies become what are
known as the learning issues that students research during their self-directed
learning (SDL). Following SDL, students apply their new knowledge and
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 237

Fig. 1. The problem-based learning cycle.

evaluate their hypotheses in light of what they have learned. At the comple-
tion of each problem, students reflect on the abstract knowledge gained. The
teacher helps students learn the cognitive skills needed for problem solv-
ing and collaboration. Because students are self-directed, managing their
learning goals and strategies to solve PBL’s ill-structured problems (those
without a single correct solution), they also acquire the skills needed for
lifelong learning. PBL was originally developed in medical schools and has
been used in a variety of settings from middle school to professional educa-
tion (Barrows, 2000; Barrows and Kelson, 1995; Barrows and Tamblyn, 1980;
Gallagher et al., 1992; Hmelo et al., 2000; Hmelo-Silver, 2000; Torp and Sage,
2002; Williams, 1992).
PBL is one of a family of approaches that include anchored instruc-
tion and project-based science (CTGV, 1997; Krajcik et al., 2000; Linn and
Hsi, 2000; White and Frederiksen, 1998). As shown in Table I, all three ap-
proaches use a common problem and rely on the teacher to help guide the
learning process. They differ in terms of the type and role of the problem,
the problem-solving process, and the specific tools that are employed. For
example, PBL uses realistic, ill-structured problems such as medical diagno-
sis or lesson design (Barrows, 2000; Hmelo-Silver, 2000). In this setting, the
problem is the focus for acquiring knowledge and reasoning strategies. In
anchored instruction, the problem is a video-based story that presents learn-
ers with a challenge at the end such as determining if a certain school project
can be profitable. The problem has two roles in anchored instruction. First,
it provides an apportunity for learners to apply their shared knowledge to a
 P1: FLT

238

Table I. Approaches to Learning Situated in Problem-Solving Experiences

PBL Anchored instruction Project-based science
Educational Psychology Review [jepr]

Problem Realistic ill-structured problem Video-based narrative ending with Driving question
complex problem
Role of problem Focus for learning information and Provide shared experience so students can Focus for scientific inquiry process
reasoning strategies understand how knowledge can support leading to artifact production
problem solving
Video supports problem comprehension
Process Identify facts, generate ideas and Guided planning and subgoal generation Prediction, observation, explanation
learning issues, SDL, revisit, and cycles
pp1224-edpr-487801

reflect
Role of teacher Facilitate learning process and model Engage students’ prior knowledge, model Introduce relevant content before
reasoning problem-solving strategies, provide and during inquiry
content instruction when needed by
students Guides inquiry process
Collaboration Negotiation of ideas Negotiation of ideas and strategies within Negotiation of ideas with peers and
May 22, 2004

Individual students bring new small groups and whole class local community members
knowledge to group for application
to problem
0:50

Tools Structured whiteboard Video controller Computer-based tools that support

planning, data collection and
Student-identified learning resources Problem-specific tools (e.g., maps, analysis, modeling, and
compasses) information-gathering
Hmelo-Silver
Style file version June 4th, 2002
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 239

relevant problem. Second, the video supports ongoing problem comprehen-

sion as problems often require 15–20 steps for solution. In project-based sci-
ence, the problem is a driving question such as “where does my garbage go?”
The problem is the focus for scientific inquiry as students must determine
how they will answer their question. In terms of process, PBL uses the specific
PBL learning cycle shown in Fig. 1. Anchored instruction emphasizes plan-
ning and subgoal generation so the students can make the problem-solving
task manageable. In project-based science, students engage in scientific in-
quiry cycles as they design experiments, make predictions, observations, and
then construct explanations of why their prediction was or was not correct.
In terms of tools, PBL uses a simple tool, a structured whiteboard with
lists of facts, ideas (i.e., hypotheses), learning issues, and action plans (dis-
cussed later) to help structure the students’ problem solving and learning.
In contrast, project-based science uses a variety of computer-based tools
to scaffold students problem solving. These include metacognitive scaffold-
ing, planners, and modeling tools (Krajcik et al., 2000; Linn and Hsi, 2000).
Anchored instruction uses tools to help learners control which aspects of
the video they want to review and includes a variety of problem-specific
assessment tools (CTGV, 1997). Collaborative learning is an important as-
pect of all three methods. All of these methods have proven effective (e.g.,
CTGV, 1997; Hmelo, 1998; Krajcik et al., 2000; Linn and Hsi, 2000; White
and Frederiksen, 1998).
There are at least two key issues that go to the heart of all of these
approaches to learning through problem solving. First, all the approaches
emphasize that learners are actively constructing knowledge in collaborative
groups. Second, the roles of the student and teacher are transformed. The
teacher is no longer considered the main repository of knowledge; she is the
facilitator of collaborative learning. The teacher helps guide the learning pro-
cess through open-ended questioning designed to get students to make their
thinking visible and to keep all the students involved in the group process.
In anchored instruction and project-based science, the teacher does some
direct instruction, often when students need information for the problem-
solving activities. The SDL emphasis is a distinguishing feature of PBL. In
PBL, students become responsible for their own learning, which necessi-
tates reflective, critical thinking about what is being learned (Bereiter and
Scardamalia, 1989). In PBL, students are asked to put their knowledge to
use and to be reflective and self-directed learners.

GOALS OF PBL

Problem-based curricula provide students with guided experience in

learning through solving complex, real-world problems. PBL was designed
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

240 Hmelo-Silver

with several important goals (Barrows and Kelson, 1995). It is designed to

help students

1) construct an extensive and flexible knowledge base;

2) develop effective problem-solving skills;
3) develop self-directed, lifelong learning skills;
4) become effective collaborators; and
5) become intrinsically motivated to learn.

Constructing extensive and flexible knowledge goes beyond having students

learn the facts of a domain. It involves integrating information across mul-
tiple domains. Such knowledge is coherently organized around the deep
principles in a domain (Chi et al., 1981). It is also flexibly conditionalized
to the extent that it can be fluently retrieved and applied under varying
and appropriate circumstances (Bransford et al., 1990). Increasingly flexi-
ble knowledge develops as individuals apply their knowledge in a variety of
problem situations (CTGV, 1997; Kolodner, 1993).
Common sense suggests that to encourage students to develop flexible
knowledge and effective problem-solving skills we must embed learning in
contexts that require the use of these skills. Laboratory experiments have
demonstrated that this is indeed the case (e.g., Needham and Begg, 1991;
Perfetto et al., 1983). Classroom-based research supports these findings as
well (Gallagher et al., 1992; Hmelo, 1998; Hmelo et al., 2000; Schwartz and
Bransford, 1998). Discussing problems in a PBL group (before beginning to
research learning issues) activates relevant prior knowledge and facilitates
the processing of new information (Schmidt et al., 1989). Students are better
able to construct new knowledge when they can relate it to what they already
know (Bransford and McCarrell, 1977).
Achieving the second goal, the development of effective problem-
solving skills, includes the ability to apply appropriate metacognitive and
reasoning strategies. For example, hypothetical–deductive reasoning is an
appropriate strategy for medical problem solving whereas analogical or case-
based reasoning may be more appropriate in many design domains such as
architecture (Kolodner, 1993; Kolodner et al., 1996). Metacognitive skills
refer to the executive control processes of planning one’s problem solving,
monitoring one’s progress, and evaluating whether one’s goals have been
met (Schoenfeld, 1985).
Metacognitive strategies are also important for the third goal of devel-
oping self-directed, lifelong learning skills. These are the skills that enable
autonomous learning. There are several subskills involved in SDL (Hmelo
and Lin, 2000; Zimmerman, 2002). First, learners must have a metacognitive
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 241

awareness of what they do and do not understand. Second, they must be able
to set learning goals, identifying what they need to learn more about for the
task they are engaged in. Third, they must be able to plan their learning and
select appropriate learning strategies. In other words, they must decide on a
course (or courses) of action to reach these goals. Finally, as they implement
their plan, learners must be able to monitor and evaluate whether or not
their goals have been attained.
The fourth goal of being a good collaborator means knowing how to
function well as part of a team. This encompasses establishing common
ground, resolving discrepancies, negotiating the actions that a group is going
to take, and coming to an agreement (Barron, 2002). These tasks require
an open exchange of ideas and engagement by all members of the group
(Cohen, 1994; Wenger, 1998). Explaining one’s ideas is important for produc-
tive collaboration and also serves to enhance learning (Webb and
Palincsar, 1996). The goal of becoming a good collaborator and the process
of learning collaboratively are often woven together.
The final goal of PBL is to help students become intrinsically moti-
vated. Intrinsic motivation occurs when learners work on a task motivated
by their own interests, challenges, or sense of satisfaction. Creating an en-
gaging problem is simple for medical students because they all share the
intrinsic goal of becoming physicians. Similarly, gifted high school students
tend to be highly motivated and have the cognitive skills that allow them
to be confident in tackling some complex task. Determining an appropriate
problem for less skilled students requires that the problem designers under-
stand what is developmentally appropriate, interesting to a heterogeneous
group of students, and moderately challenging without being overwhelming.
Several features of PBL support increased motivation for learning. Students
are more motivated when they value what they are learning and when their
educational activity is implicated in personally meaningful tasks (Ferrari and
Mahalingham, 1998; Leontiev, 1978). Students are also more motivated when
they believe that the outcome of learning is under their control (Bandura,
1997; Dweck, 1991).
To be intrinsically motivating, problems should provide students with
the proximal and tangible goal of applying their knowledge to solve a con-
crete problem. This type of goal is more motivating than are more distant,
abstract goals that may seem insurmountable (Bandura, 1997). Classroom
contexts that reward students for deep understanding, independent thought,
and action are also more motivating than many traditional classroom struc-
tures that reward comparative performances (Ames, 1992; Biggs, 1985;
Ramsden, 1992). PBL instruction techniques assume that all these goals
are achieved as part of the PBL learning cycle.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

242 Hmelo-Silver

THE PBL TUTORIAL PROCESS

The PBL learning cycle (shown in Fig. 1) is enacted through the tutorial
process that begins with the presentation of a problem and ends with student
reflection. A PBL tutorial session begins by presenting a group of students
with minimal information about a complex problem (Barrows, 2000). From
the outset, students must question the facilitator to obtain additional prob-
lem information; they may also gather facts by doing experiments or other
research (Torp and Sage, 2002). For example, when middle-school children
were asked to build artificial lungs, they performed experiments to deter-
mine how much air the lungs had to displace (Hmelo et al., 2000). At several
points during their problem solving, students typically pause to reflect on the
data they have collected so far, generate questions about those data, and hy-
pothesize about underlying causal mechanisms that might help explain the
data. Students also identify concepts they need to learn more about in or-
der to solve the problem, labeling these concepts as “learning issues.” After
considering the problem with their naı̈ve knowledge, students independently
research the learning issues they have chosen. They then regroup to share
what they have learned, reconsider their hypotheses, and/or generate new
hypotheses in light of their new learning. When completing the task, learners
deliberately reflect on the problem to abstract the lessons learned about the
problem and about their SDL and collaborative problem-solving processes.
While working through the problem, students use whiteboards to record
their evolving ideas. Figure 2 shows an example of how engineering students
used whiteboards while determining the cause of a chemical spill (Hmelo
et al., 1995). The whiteboard is divided into four record-keeping columns to
facilitate problem solving. The Facts column holds information that the stu-
dents gleaned from the problem statement such as what the problem is and
where it occurred. The Ideas column serves to keep track of their evolving
hypotheses about solutions, such as reducing the storage of hazardous chem-
icals. The students place their questions for further study into the Learning
Issues column. In this example, students identify issues related to the tech-
nology for storing hazardous chemicals. They use the Action Plan column
to keep track of plans for resolving the problem or obtaining additional in-
formation such as calling a government agency. The four columns provide
scaffolding communicating the problem-solving process in PBL (Hmelo and
Guzdial, 1996). The whiteboard serves as a focus for negotiation of the prob-
lem and as a forum for students to co-construct knowledge. The whiteboard
helps students externalize their problem solving and allows them to focus on
more difficult aspects of the problem-solving process. It provides a model of
a systematic approach to problem solving and supports student planning and
monitoring as they identify what needs to be recorded on or later removed
from the board.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 243

Fig. 2. An example of PBL whiteboard.

The PBL tutorial process helps support knowledge construction as stu-

dents are guided through their learning and problem-solving processes. As
students begin to discuss a problem with a naı̈ve understanding, they acti-
vate their prior knowledge, which helps prepare them for learning (Schmidt
et al., 1989). It also facilitates the social construction of knowledge as learn-
ers work in small groups using their inquiry skills to solve real-world prob-
lems (Greeno et al., 1996). For example, medical students learn in the
context of authentic patient problems, using the inquiry and discourse skills
of medical practice. From a purely cognitive perspective, learning expe-
riences are organized to foster the students’ understanding of concepts
through problem-solving activities (Greeno et al., 1996). However, from a
sociocultural perspective, knowledge is constructed through social interac-
tions. Increasingly, this perspective acknowledges that social practices of
inquiry and learning support the development of students as capable learn-
ers and knowers (Lampert, 2001). Thus, curricula are organized to support
the development of disciplinary patterns of discourse and representations
of a domain, emphasizing the importance of formulating and evaluating
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

244 Hmelo-Silver

questions, problems, arguments, and explanations. To further explore the

components of the tutorial process, the following sections focus on the role
of the problem, the role of the facilitator, collaborative learning in PBL, and
reflection in PBL.

The Role of the Problem

Cognitive research and practical experience with PBL have made im-
portant strides in identifying the characteristics of a good problem (Barrows
and Kelson, 1995; Gallagher et al., 1992; Kolodner et al., 1996). To foster flexi-
ble thinking, problems need to be complex, ill-structured, and open-ended; to
support intrinsic motivation, they must also be realistic and resonate with the
students’ experiences. A good problem affords feedback that allows students
to evaluate the effectiveness of their knowledge, reasoning, and learning
strategies. The problems should also promote conjecture and argumentation.
Problem solutions should be complex enough to require many interrelated
pieces and should motivate the students’ need to know and learn. As students
generate hypotheses and defend them to others in their group, they publicly
articulate their current state of understanding, enhancing knowledge con-
struction and setting the stage for future learning (Koschmann et al., 1994).
Such problems should help students become engaged in the learning process
based on their initial understanding. The problems in a PBL curriculum are
chosen so that concepts are visited in a number of problems across the entire
curriculum (Hmelo-Silver, 2000; Koschmann et al., 1994). For example, pre-
service teachers might apply information-processing concepts to a problem
on individual assessment, another on group assessment, and yet again as part
of an instructional redesign task (Derry et al., 2002; Hmelo-Silver, 2000). It
is also likely that aspects of motivation would be considered in some or all
of these problems.
Good problems often require multidisciplinary solutions. For example,
planning a long trip on the Appalachian Trail requires using knowledge
and skills from several content areas to solve the problem (e.g., math and
life sciences). The necessity of gathering knowledge from a wide range of
sources allows students to see how knowledge is a useful tool for problem
solving. Good problems also foster communication skills as students present
their plans to the rest of their class. Multidisciplinary problems should help
build extensive and flexible knowledge because information is not learned
in isolation.

The Role of the Facilitator

Having good problems is a necessary but not sufficient condition for ef-
fective PBL. The facilitator role is critical to making PBL function well. With
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 245

its emphasis on learning through problem solving and on making key aspects
of expertise visible, PBL exemplifies the cognitive apprenticeship model
(Collins et al., 1989). In PBL, the teacher/facilitator is an expert learner,
able to model good strategies for learning and thinking, rather than an ex-
pert in the content itself. The facilitator scaffolds student learning through
modeling and coaching, primarily through the use of questioning strategies
(Hmelo-Silver and Barrows, 2003). Facilitators progressively fade their scaf-
folding as students become more experienced with PBL until finally the
learners adopt many of the facilitators’ roles. The facilitator is responsi-
ble both for moving the students through the various stages of PBL and
for monitoring the group process. This monitoring assures that all students
are involved and encourages them both to externalize their own thinking
and to comment on each other’s thinking (Hmelo-Silver, 2002; Koschmann
et al., 1994). The PBL facilitator (a) guides the development of higher or-
der thinking skills by encouraging students to justify their thinking and (b)
externalizes self-reflection by directing appropriate questions to individu-
als. The facilitator plays an important role in modeling the problem solving
and SDL skills needed for self-assessing one’s reasoning and understanding.
Although the facilitator fades some of his or her scaffolding as the group
gains experience with the PBL method, s/he continues to monitor the group,
making moment-to-moment decisions about how best to facilitate the PBL
process. The facilitator directly supports several of the goals of PBL. First,
s/he models the problem solving and SDL processes. Second, the facilitator
helps students learn to collaborate well. An underlying assumption is that
when facilitators support the learning and collaboration processes, students
are better able to construct flexible knowledge.
There are several important issues in understanding how to facilitate
and why it is so difficult for a facilitator who is comfortable with one group
and a small number of students to then monitor a typical classroom with sev-
eral groups and many students. Facilitation is a subtle skill. It involves know-
ing when an appropriate question is called for, when the students are going
off-track, and when the PBL process is stalled. In a study of an expert PBL
facilitator, Hmelo-Silver (2002) found that he accomplished his role largely
through metacognitive questioning and questioning that focused students’
attention and elicited causal explanations. The facilitator used a variety of
strategies to support his goal of getting medical students to construct causal
models of a patient’s illness. He asked students to explain their reasoning to
the point where they realized that the limitations of their knowledge neces-
sitated creating a learning issue. Another strategy was to ask students how
hypotheses related to the patient’s signs and symptoms in order to encour-
age the students to elaborate causal mechanisms. This study demonstrated
that an expert facilitator has a flexible set of strategies that can be tailored
to different stages of the PBL process.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

246 Hmelo-Silver

An important issue in moving beyond this model of PBL is one of scale.

The role of the facilitator is extremely important in modeling thinking skills
and providing metacognitive scaffolding. The medical school environment is
privileged in being able to provide a facilitator for each small group. It is less
clear how this might translate into other environments. Hmelo-Silver (2000)
has successfully managed to facilitate multiple groups, using a wandering
facilitation model. In this model, the facilitator rotates from group to group,
adjusting the time spent with each of the groups in the classroom according
to their needs. By looking at large poster sheets created by each group and
hung on the classroom walls, she was able to dynamically assess the progress
of each of the groups and adjust her facilitation efforts accordingly. In ad-
dition, students rotated through the facilitator role with the help of prompt
cards that gave examples of different techniques that could be used at dif-
ferent stages of the PBL process. This is a lower level of scaffolding than is
possible in a one-facilitator-per-group model so some adaptations of PBL
are needed to accomplish some of the facilitation functions. For example,
reflection rarely happens in groups without a facilitator and so alternative
mechanisms, such as structured journals, are needed to ensure reflection
(Hmelo-Silver, 2000). This wandering facilitation strategy was used with un-
dergraduate students who are a more varied group than medical students
but are still more mature than elementary and secondary students. Further
research is needed to explore strategies that can be used to facilitate PBL
with less mature learners in a typical classroom of 25 or more students.

Collaborative Learning in PBL

Collaborative problem-solving groups are a key feature of PBL. One

assumption of PBL is that the small group structure helps distribute the
cognitive load among the members of the group, taking advantage of group
members’ distributed expertise by allowing the whole group to tackle prob-
lems that would normally be too difficult for each student alone (Pea, 1993;
Salomon, 1993). The notion of distributed expertise is particularly relevant
in PBL because as the students divide up the learning issues they become
“experts” in particular topics. Furthermore, research suggests that the small
group discussions and debate in PBL sessions enhances problem solving
and higher order thinking and promotes shared knowledge construction
(Blumenfeld et al., 1996; Brown, 1995; Vye et al., 1997).
In PBL groups, the students often work together to construct collabo-
rative explanations. Most PBL groups need some help to collaborate effec-
tively. In the traditional PBL model (i.e., Barrows, 2000), a facilitator helps
ensure that all students are involved in the discussion. In the absence of a
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 247

dedicated facilitator, there are a variety of techniques that might help estab-
lish productive collaboration. For example, scripted cooperation, reciprocal
teaching, and the use of student roles have all been used to support effec-
tive collaborative learning with K–16 students (O’Donnell, 1999; Palincsar
and Herrenkohl, 1999). Although none of these techniques have been tested
in a PBL classroom, the technique of assigning students to different cogni-
tive roles was used in a fourth-grade inquiry science classroom (Palincsar and
Herrenkohl, 1999). These cognitive roles included predicting and theorizing,
summarizing results, and relating predictions and theories to results. These
roles, much like a facilitator, help ensure that all group members are cogni-
tively engaged (Hmelo-Silver, 2002). Use of some of these well-researched
techniques might facilitate effective collaboration within PBL settings.

Reflection in PBL

Reflecting on the relationship between problem solving and learning is

a critical component of PBL and is needed to support the construction of ex-
tensive and flexible knowledge (Salomon and Perkins, 1989). This reflection
should help learners understand the relationship between their learning and
problem-solving goals. Thus, each problem-solving task is not an end in it-
self but rather a means to achieve a self-defined learning goal (Bereiter and
Scardamalia, 1989). Reflection helps students (a) relate their new knowl-
edge to their prior understanding, (b) mindfully abstract knowledge, and
(c) understand how their learning and problem-solving strategies might be
reapplied. PBL incorporates reflection several times throughout the tutorial
process and when completing a problem. Students periodically reflect on the
adequacy of the Ideas they have recorded on the whiteboard and their own
knowledge relative to the problem. At the completion of a problem, students
reflect on what they have learned, how well they collaborated with the group,
and how effectively they directed their learning. As students make inferences
that tie the general concepts and skills to the specifics of the problem that
they are working on, they should construct a more coherent understanding
(Chi et al., 1989). Key to developing such understanding is reflection on their
knowledge and strategies. The reflection process in PBL is designed to help
students make these inferences; identify gaps in their thinking; and trans-
fer their problem-solving strategies, SDL strategies, and knowledge to new
situations. The literature on transfer shows that individuals have a difficult
time transferring general principles from one task to another, even when
the knowledge is perfectly relevant to someone who understands both tasks
(Gick and Holyoak, 1980, 1983; Novick and Holyoak, 1991; Williams et al.,
1993). Reflection should increase the probability of transfer (Salomon and
Perkins, 1989).
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

248 Hmelo-Silver

Critical reflection can give students a basis for improvement. This next
example demonstrates that students can be critical about themselves and
others during the end of a problem reflection session. They freely acknowl-
edged both their strengths and their weaknesses as Delia, a 2nd-year medical
student, did in this excerpt:
Well, I’ll go first because we’ll just go around . . . I think I did a pretty good job with
this case. I think for my knowledge, I think I covered . . . a good breadth of issues,
and I think I covered the issues on the learning issues list well, I think I could have
gotten a little bit more in depth into some of them, especially, you know . . . vitamin
B12 deficiency was a thing I ran across, you know we kind [of] ruled it down, but I
ran across it when I was looking at some differentials, and I studied it a little bit but
didn’t get all the way down into the details of it, so I’m gonna want to go back and
do that, especially the biochemistry . . ., I think also with the neurosyphilis I’d like to
look a little bit more into that because I’m really weak in understanding how that
works . . . but I think, I did a good job with the differential diagnosis and figuring out
what we needed to know in order to rule things in and out. I thought I did a good
job with that when I was studying, and it worked well for the case . . .. (Hmelo-Silver
and Barrows, 2002)

Delia noted that she needed to go back and look up some of the biochemistry
concepts and she reflected on her SDL. The other students offered their
specific feedback. For example, Carol pointed out that
One thing I thought you did really well was that you understood which tests were
needed and why they were needed and how they were done and what they would
tell us. I thought that was really one of your strengths in this case. (Hmelo-Silver and
Barrows, 2002)

Jim, however, noted that

I did feel you seemed to anchor a little bit on the olivopontocerebellar atrophy, you
seemed to bring that up a lot, but at the same time you did generate several hypothe-
ses and you seemed to listen to other people’s explanations of why they thought
something else was more likely, or whatever . . . I didn’t notice you interrupting that
much. (Hmelo-Silver and Barrows, 2002).

These student reflections refer to specific behaviors exhibited by Delia, both

those that she did well and those that needed improvement. Both positive
and negative comments from students provide others in their group with
information they can use to improve their effectiveness as collaborators and
their self-directed learning, such as using a particular general textbook to
get an overview before consulting a more specialized text.
Most groups need help to reflect on their learning (Hmelo-Silver, 2000).
A dedicated facilitator can support student reflection, but in larger groups
and with younger learners there are other techniques that may be help-
ful. One approach to improving student reflection is the use of structured
journals. In the context of using PBL with middle-school students, Puntam-
bekar and Kolodner (1998) developed structured diaries to support student
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 249

reflection. Approaches such as these need further evidence of their effec-

tiveness before advocating their widespread incorporation into PBL models.

WHAT HAVE WE LEARNED ABOUT PBL

This section considers the evidence about each of the purported goals
of PBL. Much of the evidence comes from research in medical schools and
gifted education although there are a few studies involving other populations.
There are many innovative descriptions of using PBL in various settings:
educational administration, business, educational psychology, engineering,
chemistry, various undergraduate disciplines, and K–12 education (Boud and
Felletti, 1991; Bridges, 1992; Duch et al., 2001; Hmelo-Silver, 2000; Hmelo
et al., 1995; Ram, 1999; Stepien and Gallagher, 1993; Torp and Sage, 2002),
but there is less empirical evidence as to what students are learning and
how. The questions that need to be addressed in such a literature include
the following: Do students construct an extensive and flexible knowledge
base, do they become better problem-solvers, what kind of self-directed
learners do they become, how do they collaborate (and what factors affect
their collaboration), and are they intrinsically motivated? The next sections
address the evidence as to whether students actually achieve these goals.

Constructing Extensive and Flexible Knowledge

The results on what students learn from PBL are mixed. Much eval-
uation of PBL has examined traditional academic outcome measures such
as examination scores. In general, several meta-analyses have demonstrated
that PBL students scored slightly lower than traditional medical students
on multiple-choice measures of academic achievement such as the National
Board of Medical Examiners (NBME) Part I, which examines basic science
knowledge (Albanese and Mitchell, 1993; Goodman et al., 1991; Mennin
et al., 1993; Vernon and Blake, 1993). However, meta-analyses also showed
that PBL students performed slightly better than traditional medical stu-
dents at tasks related to clinical problem solving such as NBME II and
on ratings and tests of clinical performance (Albanese and Mitchell, 1993;
Vernon and Blake, 1993). The effect sizes were small on the NBME II and
moderate when performance ratings were used. Although multiple-choice
tests measure knowledge, they may not get at the type of extensive and flexi-
ble knowledge aligned with the goals of PBL. In a more recent meta-analysis,
Dochy et al. (2003) found that there was no effect of PBL on measures of
factual knowledge, however studies of knowledge application demonstrated
a moderate effect size favoring PBL students compared with traditional
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

250 Hmelo-Silver

students. More complex measures of flexible knowledge application may be

more sensitive to the effects of PBL.
Recent studies have examined how PBL students performed on a
problem-solving task (Hmelo, 1998; Hmelo et al., 1997; Patel et al., 1991,
1993). Patel et al. (1991, 1993) asked traditional and PBL students to provide
diagnostic explanations of a clinical problem. The PBL students’ explana-
tions, although more error-prone, were also more elaborated than those of
the medical students in traditional curricula. This result is consistent with
research that demonstrated that as people are first attempting to apply new
knowledge, they do not always do it well (Lesgold et al., 1988; Novick and
Hmelo, 1994). Chi et al. (1994) suggest that errors are a necessary step in
learning to apply new knowledge. By articulating incorrect knowledge, learn-
ers have the opportunity to revise their false beliefs when they are confronted
with correct knowledge. The research suggests an advantage in having a well-
elaborated knowledge structure that contains some errors instead of having
little elaborated knowledge that cannot be applied.
Problem-based and traditional students were studied longitudinally
over the 1st year of medical school as they constructed causal explanations
of six different problems, covering a variety of different diseases (Hmelo,
1998). This was a quasi-experimental study that compared students who self-
selected into traditional and PBL tracks at a midwestern medical school. The
students generated causal explanations for each of two problems at each of
three testing sessions that occurred during the 1st week of classes, and again
after 3 and 7 months. The students’ explanations were scored for accuracy,
coherence, and use of science concepts. The students did not differ on any
of these measures at the initial testing session. The students in the PBL
curriculum were more likely to produce accurate hypotheses and coherent
explanations than students in the traditional curriculum. In addition, they
were more likely to use science concepts in their explanations. This suggests
that the PBL students were constructing knowledge that they could bring to
bear in accurately solving problems. Their science knowledge was flexible in
that they were able to transfer it to new problem situations.
The accuracy effect appears robust. When students in PBL and tradi-
tional curricula were compared in terms of diagnostic accuracy for 30 case
vignettes, PBL students were more accurate than students in a traditional
curriculum (Schmidt et al., 1996). Thus, these quasi-experimental studies in
medical schools tend to support the hypothesis that PBL students are able
to construct knowledge, provided that the assessments measure knowledge
in problem-solving contexts rather than in the context of multiple-choice
examinations.
PBL has been applied in undergraduate education as well. In an inno-
vative engineering course in sustainable technology, students used PBL in
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 251

multidisciplinary teams (Hmelo et al., 1995). Students were given pretests

and posttests assessing both factual knowledge and problem-solving skills.
The students demonstrated increases on both measures. When PBL was used
to teach statistical reasoning to undergraduates, Derry et al. (2000) demon-
strated that students showed learning gains for some, but not all, of the
course content using a pre–post design. In a course for preservice teachers
using video problems and a web-based information resource, Derry et al.
(2002) found evidence that their PBL approach led to transfer and flexible
use of course concepts. Learning outcomes were measured using a problem-
based assessment. Students viewed a video that showed a student being
interviewed before and after instruction along with some excerpts from the
instruction. Their task was to explain why the student featured in the video
failed to learn. The results demonstrated that students applied more relevant
concepts and produced more sophisticated explanations at posttest than at
pretest. Schwartz and Bransford (1998) conducted a controlled study of stu-
dents in an undergraduate psychology course. They compared students in
three groups: (a) students who just solved problems, (b) students who read
a textbook chapter prior to attending a lecture, and (c) students who solved
problems prior to attending a lecture. They found that students who solved
problems prior to the lecture performed better on a problem-solving task
than students who read the chapter or those who just solved problems. This
finding suggests that attempting to solve a problem helps create a readiness
to learn from a lecture.
Rather than looking at outcomes, Hmelo-Silver (2000) examined the
artifacts that students produced in a problem-based educational psychology
course and found that students identified appropriate course concepts and
used them in multiple problems. The students’ initial understanding of a
concept was often a vague knowledge display with little connection to the
problem they were working on as in this example:

Basic knowledge is the main focus of another group’s approach to teaching but
it can only be accomplished through memory. By going back to basic concepts and
incorporating new ideas, memory is a necessary attribute in the success of knowledge
based learning. Memory is the processes [sic] by which information is encoded, stored,
and retrieved. Long term memory becomes the goal of the students in this teaching
approach. Information in long term memory is practically unlimited in both capacity
and duration but on the other hand, long term memory is hard to achieve [sic].
Through constant use of information, long term memory is possible because schemas
are formed about how situations will occur based on past experiences. Schemas,
therefore, become a template of knowledge on which information is organized . . ..
(Hmelo-Silver, 2000, p. 51)

In this excerpt, the students provided a limited and abstract description of

basic information-processing concepts. Later in the course, the same students
applied information-processing concepts in a more sophisticated way.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

252 Hmelo-Silver

The first thing that Mr. Johnson should have done was to introduce a unit on static
electricity by asking the students what they already knew about static electricity. We
suggest that Mr. Johnson create a concept map using what the students already know
about static electricity from their other classes or everyday lives. “Prior knowledge
is stored in the form of schemas. Teachers can activate these schemas in a number
of ways including: reviewing, questioning, or developing with the students a concept
map of prior knowledge” (Knowledge Web, the Prior Knowledge Use [sic]). A con-
cept map is extremely important because teachers use students’ prior knowledge to
explain and discuss increasingly more sophisticated concepts. Prior knowledge be-
comes a platform upon which new understanding is constructed ” (Knowledge Web,
the Prior Knowledge Use [sic]). . . . When the students are done explaining to Mr.
Johnson what they previously knew about static electricity, we suggested he give a
brief lecture to fill in the gaps and add to the concept map what the students missed.
The new knowledge that students learn from the lecture provides them with an inte-
gral tool that will allow them to make more meaningful connections when they see
the experiment. (Hmelo-Silver, 2000, p. 53)

In this excerpt, the students did not describe information-processing theory

directly but they were using it to show the importance of prior knowledge
for new learning, indicating that they were thinking about memory at a
deeper level. The concepts in the latter excerpt are connected to the problem
solution that the students proposed. That was not evident in the earlier
excerpt. The analyses of group whiteboards and projects demonstrated that
as students gained experience with concepts (e.g., information processing),
they were able to distinguish the aspects of a concept that were appropriate
for use in solving particular problems.
The results from these studies of undergraduate students provide ad-
ditional evidence that problem-based curricula can help students construct
flexible knowledge. PBL appears to support learning in undergraduate and
professional educational contexts, but what about with less mature learners?
There has been little work with younger students. Related approaches such as
the work with anchored instruction have shown positive learning outcomes
for students using problem-solving videos (CTGV, 1997). In work with gifted
high school students, Gallagher and Stepien (1996) compared student scores
in a problem-based American studies course with those from students in a
traditional class. The PBL students scored higher on a multiple-choice test
than traditionally instructed students. In an action research study with gifted
high school students, students tended to retain information presented in PBL
units better than information from traditional units, despite the fact that the
students thought they learned more in lecture-based units (Dods, 1997). In a
study with a heterogeneous population of sixth-grade students, Hmelo et al.
(2000), in collaboration with classroom teachers, developed a PBL unit that
involved students designing artificial lungs. This design experiment demon-
strated that PBL students showed greater gains on both short-answer tests
and a drawing task than students in comparison classrooms. However, the
students in the PBL class had some misunderstandings at the end of the
3-week unit. The results of this research suggest that adaptations might be
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 253

needed to use PBL in developmentally appropriate ways and that varying

kinds of scaffolding might be needed to help children learn while tackling
complex problems. Thus there may be a place for direct instruction once
students have developed the need to know, particularly with less uniformly
skilled learners.

Developing Effective Problem-Solving Skills

One indicator of effective problem-solving skills is the ability to transfer

reasoning strategies to new problems. Patel et al. (1991, 1993) asked tradi-
tional and PBL students to provide diagnostic explanations of a clinical prob-
lem. They showed that students in the PBL curriculum were more likely to
use hypothesis-driven reasoning (as they were taught) than were students in
a traditional curriculum. The students in the traditional medical school used
predominately data-driven reasoning, a form of reasoning that is more char-
acteristic of experts, but only on familiar problems (Norman et al., 1994). Al-
though experts often use this type of reasoning, it is not likely to be effective
for novices. Data-driven reasoning relies on having an extensive knowledge
base, which is unlikely for novices. In a laboratory experiment designed to
test the notion that hypothesis-driven reasoning supports learning, Norman
et al. (1998) found that having students learn how to read electrocardiograms
using a hypothesis-driven strategy led to greater accuracy on a transfer task
than using a data-driven strategy. This finding suggests that a hypothesis-
driven reasoning strategy is a mechanism for learning. In two studies, stu-
dents in PBL curricula transferred the hypothesis-driven reasoning strategy
to unrelated problems and generated more coherent explanations than stu-
dents without PBL experience (Hmelo, 1998; Hmelo et al., 1997).
Another aspect of problem-solving skills is being able to define what
the problem actually is, especially with ill-structured problems. This is called
problem finding and is the aspect of problem solving that refers to identi-
fying the problem. When comparing gifted students who were traditionally
instructed with students in a PBL class on problem-solving skills, Gallagher
et al. (1992) found that PBL students were more likely to include problem
finding as a step when presented with a novel ill-structured problem. Al-
though research on the influence of PBL on strategy transfer is limited, it
does provide some evidence that students in PBL learn problem-solving and
reasoning strategies that are transferable to new problems.

Developing SDL Skills

One of the purported benefits of PBL is its claim to prepare life-

long learners because of its emphasis on self-directed learning (SDL).
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

254 Hmelo-Silver

Understanding SDL is a complex phenomenon so this section has two parts.

The first part examines of what it means to become a self-directed learner by
reviewing qualitative studies of SDL. The second part reviews the literature
that examines quantitative indicators of SDL. Some of the indicators of SDL
include planning one’s own learning, developing and applying strategies, and
appropriately using learning resources.
Becoming a self-directed learner is a multifaceted process. When stu-
dents in PBL curricula in three different disciplines were interviewed,
Abrandt Dahlgren and Dahlgren (2002) found that students in two of three
disciplines felt a great sense of uncertainty about what to study. Students’
SDL strategies evolve over time as Evensen (2000; Evensen et al., 2001)
demonstrated in qualitative analyses. In both analyses, Evensen interviewed
medical students from a PBL group and analyzed oral learning logs and writ-
ten notes that the students kept. The first analysis used a grounded theory ap-
proach and focused on two of these students as it examined the interactions of
academic self-concept, learning strategies, learning opportunities provided
by the program, and evaluation mechanisms (Evensen, 2000). Both students
developed strategies for coping with challenges to their self-efficacy and de-
scribed the reflection on their learning and information-seeking strategies.
This research provides a glimpse into the lives of two students learning to
adapt to the SDL demands of a PBL program. For example, one of the
students, Kyle, a successful undergraduate premed major, began the PBL
program highly confident that he could easily learn anything that was “in
the book,” and believed that was what he needed to be successful in medical
school. He chose PBL because he thought it would allow him to learn on his
own without attending lectures. He enjoyed the problems but initially found
it annoying that topics already studied were revisited in multiple problems.
Revisiting topics also caused him to doubt whether he had adequately stud-
ied. His major strategy was to rely on reading to learn. He saw the group
meetings as a good way to get the big picture of a case, to collectively de-
cide on what was important, and to review. As time went on, Kyle began
encountering material that was difficult so he needed to adjust his strategies.
For the first time in his life, he had to reread text. This took more time and
he realized that he needed to take good notes because he would not have
time to reread the texts before exam. He also learned which topics could
be skimmed and which topics needed more in-depth research. Early in the
year, he was pleased to be an information resource for his group. Later, he
shifted to become more of a collaborator. He realized the benefits of shar-
ing resources. The case studies demonstrate that students bring strategies
and beliefs about learning to a new situation. This is not surprising because
medical students have generally been successful learners throughout their
school careers. However, prior ways of learning were not transported intact
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 255

to this new learning environment; rather, good self-directed learners were

proactive in achieving their goals. That is, these successful learners adapted
their personal strategies to the situational demands.
To build a model of how students developed as self-directed learners,
additional case studies were constructed (Evensen et al., 2001). These case
studies suggested that the more reflective learners were about SDL, the more
likely they could modify or invent strategies for SDL. For example, Beth, an
undergraduate theater major, learned to compensate for her lack of science
background by starting with a “baby bio book” (p. 663). The simpler text
helped provide the background knowledge needed to tackle the medical
textbooks. Beth’s strength was her desire to master the content and her con-
fidence in her ability to catch up to other students. Beth’s knowledge of her
weaknesses allowed her to develop adaptive strategies to cope with her lim-
ited science background. Initially, Beth missed the traditional benchmarks
of exams but over the course of the semester she compensated by purchasing
an NBME exam review book and reviewing with her group. Beth built on
her strengths and overcame her weaknesses as she adjusted her SDL to meet
the demands of the PBL curriculum. Another student in the group, Hope,
had always been successful in school by reading and taking abundant notes.
She chose the PBL curriculum because it would give her the opportunity to
do something different. She was initially enthusiastic and liked the idea of
learning for understanding rather than memorizing. She was not confident
in the group discussion, though, which challenged her self-efficacy. Rather
than adapting to the demands of PBL, Hope went back to her old strategies.
Evensen concluded that SDL may well be an individual characteristic but
it is one that changes over time, for better or worse. As these case studies
show, students understood the need for SDL but had different ideas of what
this meant. For some, it was the same as independent study. Other students
recognized the contrast between PBL and previous educational experiences
and realized that they would need to adopt new strategies to attune to the
program. The strategies they adopted interacted with their previous learning
histories, self-regulated learning strategies, self-efficacy, and the features of
the PBL environment.
Students’ approaches to learning from problems differ qualitatively de-
pending on their degree of self-regulation. Ertmer et al. (1996) conducted
a qualitative study of how veterinary students approached learning from
problems. Students who were low self-regulated learners (SRL) had diffi-
culty adapting to the kind of learning required in problem-based instruction.
They fluctuated in their perception of the value of learning from problems.
High SRL students valued learning from problems and tended to focus on
the problem analysis and reflection process. In contrast, low SRL students
tended to focus on fact acquisition. Ertmer et al.’s results suggest that low
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

256 Hmelo-Silver

SRL students may have difficulty dealing with the SDL demands of a PBL
curriculum.
A commonly held assumption in the PBL approach is that problem
content and ensuing discussions direct SDL. In questioning this assumption,
Dolmans and Schmidt (2000) asked what other curricular elements might
direct SDL. The PBL students at their institution received course objectives,
a limited number of lectures, tests, and suggested references. Dolmans and
Schmidt reasoned that over reliance on these external elements might im-
pede the development of SDL skills. To examine this, they gave first through
4th-year students a questionnaire with items relating to the different curricu-
lar elements. They found that both problem discussion and course objectives
had the greatest positive influence on SDL. Tests and lectures had the least
positive influence. Furthermore, over the 4 years of the curriculum, students
increased their emphasis on the functional knowledge they would need as
physicians and decreased their reliance on external elements such as tests.
These results indicate that over time, students in a PBL curriculum become
more self-reliant.
Developing SDL skills is a difficult and multifaceted process for students
in PBL curricula. This begs the question of whether PBL students develop
better SDL skills than traditional medical students. One way to study this
is by examining component processes in SDL such as how students gener-
ated learning issues, planned their learning, and integrated new knowledge
in problem solving. Hmelo and Lin (2000) examined these component pro-
cesses by comparing medical students in traditional and PBL curricula who
had completed a pathophysiological explanation task. They found that the
PBL students transferred the hypothesis-driven strategies from their prob-
lem solving into their SDL as they used their hypotheses to plan their learn-
ing. Moreover, they were more likely to integrate new information into a
revised explanation than traditional medical students.
In a study comparing traditional and PBL medical students in terms of
the learning resources they used, Blumberg and Michael (1992) found that
PBL students were more likely to use self-chosen learning resources whereas
students in the conventional curriculum used faculty-chosen resources. PBL
students were more likely to report selecting the material to study them-
selves, whereas conventional curriculum students reported reading specific
teacher-generated assignments. Engineering students in a PBL course in sus-
tainable technology increased their use of expertise other than that provided
by course instructors as the course progressed and tended to use a variety
of student-selected resources throughout the course (Shikano and Hmelo,
1996).
The research on SDL in PBL has largely been confined to professional
students so its wider application is ripe for additional research. It is clear
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 257

that all students need to grapple with the uncertainty involved in directing
their own learning. For students who are poor self-regulated learners, PBL is
likely to pose difficulties without appropriate scaffolding for students trying
to develop SDL skills. Becoming self-directed learners is not a given as the
Evensen et al. (2001) and Ertmer et al. (1996) studies demonstrate. It is not
at all certain how to structure PBL for less mature learners. Scaffolding SDL
is likely to be especially important for younger learners but as the research
has noted, there are considerable individual differences even among adult
learners.

Becoming Effective Collaborators

Another goal of PBL is to help students become effective collaborators,

but there is little research that examines this directly. Rather, the research has
focused on factors that affect how well students learn collaboratively. Group
function is particularly important because it affects learning outcomes and
intrinsic motivation (Schmidt and Moust, 2000). Schmidt and Moust’s path
model demonstrates that group functioning is affected by the quality of the
problem and facilitator functioning but, surprisingly, not by prior knowledge.
They do not include collaboration skills as part of this model.
Students in PBL curricula do attend to collaboration as DeGrave et al.
(1996) have demonstrated. In their study, they analyzed videotape of a tu-
torial group and had students engage in stimulated recall while watching
the video. In the group meeting, students did not appear to attend to col-
laboration explicitly but their stimulated recall indicated that students were
sensitive to collaborative process and their own part in the collaboration.
Effective collaboration can lead to knowledge construction as students
construct joint explanations. An analysis of two PBL tutorial sessions found
that student discourse often focused on responding to and refining ideas
that had been proposed (Hmelo-Silver, 2002). Students constructed many
collaborative explanations as in this example after the facilitator asked the
group to clarify one of their hypotheses:

Delia: Like pernicious anemia is a big one.

Mindy: Right. That must be the vitamin, the B.
Facilitator: What, what’s pernicious anemia?
Delia: Uh, it’s a deficient, deficiency of cobalamine.
Mindy: Vitamin B12, cobalamine or . . .
Jim: Or folate.
Mindy: Or folate.
Delia: Yeah, but it’s not, that’s not pernicious anemia. That’s a, also another
macrocytic anemia.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

258 Hmelo-Silver

Mindy: Pernicious anemia is specifically.

Jim: Oh. You’re right. That’s right.
Delia: And you get anemia and you can also get peripheral . . .
Mindy: Neuropathies.
Delia: . . . neuropathies.
Facilitator: Down there too?
Carol: Technically pernicious, pernicious anemia is technically just the loss,
the lack of intrinsic factor.
Delia: The loss of intrinsic factor. So you don’t absorb.
Carol: And that’s [unintelligible].
Delia: You don’t absorb.
Carol: Right.
Mindy: Right. That’s a good distinction. You see, we just . . .
Carol: As opposed to like somebody who had part of their intestine removed
and can’t absorb.
Mindy: Right.
Carol: But their ileum is gone and they can’t absorb the B12. That’s different
than pernicious anemia, to vit[amin], intrinsic factor. (Hmelo-Silver, 2002,
pp. 203–204)

This is collaborative because students all contributed different parts of the

explanation. The facilitator triggered the explanation but then different stu-
dents offered different parts of the explanation about pernicious anemia,
what some signs might be (neuropathies), and what alternative explanations
they can rule out. Although this case study provides an example of what is
possible in well-functioning groups, helping students become effective col-
laborators is still an important issue for research.
In PBL, students are encouraged to attend to collaboration processes
through their reflection and through the interdependence of learning within
the group, but they do not necessarily know how to deal with the collabo-
rative aspects of PBL effectively (Abrandt Dahlgren and Dahlgren, 2002;
Evensen et al., 2001). To help students learn to be better collaborators,
Faidley, et al. (2000) developed observational and self-report instruments
that could be used to provide information on group processing. The Learning
Team Survey (LTS) is a self-report instrument designed to focus on behav-
iors and attitudes that are important to group learning. The observational
checklist focuses on substantive and group processing behaviors. In a pilot
study, Faidley et al. used the observational checklist to determine whether
group differences in LTS could be explained by attending to the behavioral
variables. The results demonstrated that there was a relationship between
the two instruments. There was great variability in the groups studied, pro-
viding further evidence that not all groups are effective. The facilitator had
a large effect on how well the group worked, consistent with the model
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 259

of Schmidt and Moust (2000). Faidley et al. hypothesized that instruments

such as LTS provide information that can lead to productive reflection and
improvement in the group process.
There is not yet evidence that supports the hypothesis that PBL helps
students become better collaborators (broadly defined). There is evidence
that students do work together to provide collaborative explanations as
demonstrated in the example from Hmelo-Silver (2002). There is also evi-
dence that demonstrates that collaboration in tutorial groups is a key factor
in student learning and motivation, but that not all groups collaborate well.
Further research is needed to examine whether PBL environments help all
learners become better collaborators.

Becoming Intrinsically Motivated

Enhancing student motivation is purported to be a major advantage

of PBL. Because learning issues arise from the problem (in response to
students’ need to know), intrinsic motivation should be enhanced. Unfor-
tunately there is little research that bears directly on this issue. Most of the
research has instead examined student satisfaction or confidence. Students’
reaction to a PBL course in statistical reasoning was mixed (Derry et al.,
2000). Some students really enjoyed the class but others resisted changing
their way of learning or did not like working collaboratively. In interviews
with veterinary students, Ertmer et al. (1996) found evidence of intrinsic mo-
tivation associated with PBL but the nature of students’ motivation in PBL
may depend on their academic or professional discipline (Abrandt Dahlgren
and Dahlgren, 2002). In the medical school environment, across many stud-
ies, students in PBL curricula report being more satisfied with their learn-
ing and confident in their understanding than those in traditional curricula
(Albanese and Mitchell, 1993; Hmelo, 1994; Vernon and Blake, 1993). There
has been little if any work examining motivation among K-12 PBL students.
It is important to note that in medical schools, the students are a fairly se-
lect group and the PBL curricula are well established. Moreover, PBL is
used throughout the entire curriculum. In other studies, the instructional in-
tervention was a single experimental course within a larger curriculum with
competing demands from other well-structured courses. A single course may
not provide the opportunity to become acclimated to a new way of learning.
This makes motivation in PBL a complex issue. The results for medical stu-
dents are consistent—they enjoy PBL and feel confident about their learning
(Albanese and Mitchell, 1993; Hmelo, 1994; Vernon and Blake, 1993). There
is little empirical data about motivation in undergraduate and K-12 educa-
tion making it difficult to draw conclusions about motivation in these other
contexts.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

260 Hmelo-Silver

DISCUSSION

This article describes PBL, one of a family of experiential learning meth-

ods, and how students learn using this method. There is some strong evi-
dence about the nature of knowledge construction and the development of
problem-solving skills in certain settings, but there are some cautionary notes
to consider. The claims of PBL advocates are not all supported by an exten-
sive research base, and much of the research has been restricted to higher
education, predominantly in medical schools. There is little research with K-
12 populations. Much of the research has used case study, pre–post test, or
quasi-experimental designs rather than controlled experiments. However,
these different designs provide converging evidence about learning with
PBL. The majority of the research on PBL examines knowledge construc-
tion, problem solving, and SDL. There is little work in the areas of motivation
and collaboration. A research agenda for PBL should examine these facets
of PBL more carefully. It would be naı̈ve to believe that the medical school
model of PBL could be imported into other settings without considering
how to adapt it to the local context, goals, and developmental level of learn-
ers. Another barrier to using PBL in K-12 education is the constraints of
classroom organization. The model of PBL in medical school involves an
integrated, interdisciplinary curriculum organized around problems rather
than subject domains. In most K-12 situations, teachers must assess students
in specific subject areas and problems often do not map neatly onto these
subject area divisions. Moreover, it requires careful planning to engage in
PBL in 50-min class periods.
Certain aspects of the PBL model should be tailored to the devel-
opmental level of the learners. SDL may prove particularly difficult for
younger learners who tend to have difficulty applying metacognitive strate-
gies. Scaffolding that helps support student metacognition and reflection
is particularly important at this level and has been used profitably with
younger learners in project-based science (Linn and Hsi, 2000; White and
Frederiksen, 1998). Understanding the nature of these adaptations and how
aspects of PBL may be scaffolded for different learners are important is-
sues for research. Moreover, there may be a place for direct instruction on a
just-in-time basis. In other words, as students are grappling with a problem
and confronted with the need for particular kinds of knowledge, a lecture at
the right time may be beneficial. As Schwartz and Bransford (1998) demon-
strated, PBL may create a “time for telling” (p. 475). How this can be incor-
porated into a student-centered learning environment remains an empirical
question. There is a great need for evidence-based instructional strategies
that demonstrate which facets of PBL are important for particular kinds of
outcomes so that educators can make informed choices in adapting PBL to
their particular contexts.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 261

One barrier to using PBL in more diverse settings is the lack of a suffi-
cient number of skilled facilitators in many settings. Classrooms have more
students than one person can easily facilitate, and learning to facilitate well is
a challenge (Derry et al., 2001). Some techniques such as procedural facilita-
tion, scripted cooperation, and structured journals may prove useful tools in
moving PBL to other settings. In addition, the PBL activity structure might
be modified to support PBL for specific teaching goals. Technology may also
play an important role in adapting PBL for specific disciplines. For example,
in the STEP system (Secondary Teacher Education Program), the activity
structure was modified to help preservice teachers engage in instructional
design as they learn about educational psychology (www.estepweb.org). The
STEP system is an integrated system with videocases of classroom instruc-
tion, a hypermedia textbook, a set of problems, a virtual whiteboard, and an
activity structure that guides students through PBL and the instructional de-
sign process (Steinkuehler et al., 2002). The activity is divided into a sequence
that starts with individual problem analysis, moves on to group SDL and
problem solving, and ends with individual explanation and reflection. The
whiteboard structure has been modified to include students’ ideas about ob-
jectives, assessment, and activities. A space is provided that reminds students
to include the psychological justification for their ideas. The same three cat-
egories emphasized in the whiteboard are the focus of the students’ problem
solving. The problems are video based and indexed to educational psychol-
ogy concepts to help support the students’ SDL. Approaches to distributing
facilitation using various tools and technology need further research and
evaluation to understand their strengths and limitations.
In conclusion, PBL is a pedagogical technique that situates learning
in complex problem-solving contexts. It provides students with opportuni-
ties to consider how the facts they acquire relate to a specific problem at
hand. It obliges them to ask what they need to know. PBL offers the po-
tential to help students become reflective and flexible thinkers who can use
knowledge to take action. Although the roots of PBL go back to Kilpatrick
(1918) and Dewey (1938), PBL has the advantage of suggesting a method
to promote active and reflective knowledge-building-for-action (Hmelo and
Guzdial, 1996). Still, careful research is needed to understand if and how
these potentials might be realized.

ACKNOWLEDGMENTS

This work has been partially supported by a National Academy of

Education/Spencer postdoctoral fellowship and NSF ROLE grant # 0107032.
Any opinions, findings, and conclusions or recommendations expressed in
this material are those of the author and do not necessarily reflect the views
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

262 Hmelo-Silver

of the National Science Foundation. I thank Ellina Chernobilsky, Angela

O’Donnell, the editor, and two anonymous reviewers for their helpful com-
ments on earlier versions of this article. I also thank Leah Smith for her
editorial assistance.

REFERENCES

Abrandt Dahlgren, M., and Dahlgren, L. O. (2002). Portraits of PBL: Students’ experiences
of the characteristics of problem-based learning in physiotherapy, computer engineering,
and psychology. Instr. Sci. 30: 111–127.
Albanese, M. A., and Mitchell, S. (1993). Problem-based learning: A review of literature on its
outcomes and implementation issues. Acad. Med. 68: 52–81.
Ames, C. (1992). Classrooms: Goals, structures, and student motivation. J. Educ. Psychol. 84:
261–271.
Bandura, A. (1997). Self-Efficacy: The Exercise of Control, Freeman, New York.
Barron, B. J. S. (2002). Achieving coordination in collaborative problem-solving groups. J.
Learn. Sci. 9: 403–437.
Barrows, H. S. (2000). Problem-Based Learning Applied to Medical Education, Southern Illinois
University Press, Springfield.
Barrows, H., and Kelson, A. C. (1995). Problem-Based Learning in Secondary Education and
the Problem-Based Learning Institute (Monograph 1), Problem-Based Learning Institute,
Springfield, IL.
Barrows, H. S., and Tamblyn, R. (1980). Problem-Based Learning: An Approach to Medical
Education, Springer, New York.
Bereiter, C., and Scardamalia, M. (1989). Intentional learning as a goal of instruction. In
Resnick, L. B. (ed.), Knowing, Learning, and Instruction: Essays in Honor of Robert Glaser,
Erlbaum, Hillsdale, NJ, pp. 361–392.
Biggs, J. B. (1985). The role of metalearning in study processes. Br. J. Educ. Psychol. 55: 185–212.
Blumberg, P., and Michael, J. A. (1992). Development of self-directed learning behaviors in a
partially teacher-directed problem-based learning curriculum. Teach. Learn. Med. 4: 3–8.
Blumenfeld, P. C., Marx, R. W., Soloway, E., and Krajcik, J. S. (1996). Learning with peers: From
small group cooperation to collaborative communities. Educ. Res. 25(8): 37–40.
Boud, D., and Feletti, G. (1991). The Challenge of Problem Based Learning, St. Martin’s Press,
New York.
Bransford, J. D., Brown, A. L., and Cocking, R. (2000). How People Learn, National Academy
Press, Washington, DC.
Bransford, J. D., and McCarrell, N. S. (1977). A sketch of a cognitive approach to comprehen-
sion: Some thoughts about understanding what it means to comprehend. In Johnson-Laird,
P. N., and Wason, P. C. (eds.), Thinking: Readings in Cognitive Science, Cambridge Univer-
sity Press, Cambridge, UK, pp. 377–399.
Bransford, J. D., Vye, N., Kinzer, C., and Risko, R. (1990). Teaching thinking and content
knowledge: Toward an integrated approach. In Jones, B. F., and Idol, L. (eds.), Dimensions
of Thinking and Cognitive Instruction, Erlbaum, Hillsdale, NJ, pp. 381–413.
Bridges, E. M. (1992). Problem-Based Learning for Administrators, ERIC Clearinghouse on
Educational Management, Eugene, OR.
Brown, A. L. (1995). The advancement of learning. Educ. Res. 23(8): 4–12.
Chi, M. T. H., Bassok, M., Lewis, M. W., Reimann, P., and Glaser, R. (1989). Self-explanations:
How students study and use examples in learning to solve problems. Cogn. Sci. 13: 145–182.
Chi, M. T. H., DeLeeuw, N., Chiu, M., and LaVancher, C. (1994). Eliciting self-explanations
improves understanding. Cogn. Sci. 18: 439–477.
Chi, M. T. H., Feltovich, P., and Glaser, R. (1981). Categorization and representation of physics
problems by experts and novices. Cogn. Sci. 5: 121–152.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 263

Cognition and Technology Group at Vanderbilt (1997). The Jasper Project: Lessons in Curricu-
lum, Instruction, Assessment, and Professional Development, Erlbaum, Mahwah, NJ.
Cohen, E. G. (1994). Restructuring the classroom: Conditions for productive small groups. Rev.
Educ. Res. 64: 1–35.
Collins, A., Brown, J. S., and Newman, S. E. (1989). Cognitive apprenticeship: Teaching the
crafts of reading, writing, and mathematics. In Resnick, L. B. (ed.), Knowing, Learning,
and Instruction: Essays in Honor of Robert Glaser, Erlbaum, Hillsdale, NJ, pp. 453–494.
DeGrave, W. S., Boshuizen, H. P. A., and Schmidt, H. G. (1996). Problem-based learning:
Cognitive and metacognitive processes during problem analysis. Instr. Sci. 24: 321–341.
Derry, S. J., Lee, J., Kim, J.-B., Seymour, J., and Steinkuehler, C. A. (2001, April). From ambitious
vision to partially satisfying reality: Community and collaboration in teacher education.
Paper presented at the Annual Meeting of the American Educational Research Association,
Seattle, WA.
Derry, S. J., Levin, J. R., Osana, H. P., Jones, M. S., and Peterson, M. (2000). Fostering students’
statistical and scientific thinking: Lessons learned from an innovative college course. Am.
Educ. Res. J. 37: 747–773.
Derry, S. J., Siegel, M., Stampen, J., and the STEP team (2002). The STEP system for collab-
orative case-based teacher education: Design, evaluation, and future directions. In Stahl,
G. (ed.), Proceedings of CSCL 2002, Erlbaum, Hillsdale, NJ, pp. 209–216.
Dewey, J. (1938). Experience and Education, Macmillan, New York.
Dochy, F., Segers, M., Van den Bossche, P., and Gijbels, D. (2003). Effects of problem-based
learning: A meta-analysis. Learn. Instr. 13: 533–568.
Dods, R. F. (1997). An action research study of the effectiveness of problem-based learning in
promoting the acquisition and retention of knowledge. J. Educ. Gifted 20: 423–437.
Dolmans, D. H. J. M., and Schmidt, H. G. (2000). What directs self-directed learning in a
problem-based curriculum? In Evensen, D. H., and Hmelo, C. E. (eds.), Problem-Based
Learning: A Research Perspective on Learning Interactions Erlbaum, Mahwah, NJ, pp. 251–
262.
Duch, B. J., Groh, S. E., and Allen, D. E. (2001). The Power of Problem-Based Learning, Stylus,
Steerling, VA.
Dweck, C. S. (1991). Self-theories and goals: Their role in motivation, personality, and develop-
ment. In Nebraska Symposium on Motivation, 1990, University of Nebraska Press, Lincoln,
pp. 199–235.
Ertmer, P., Newby, T. J., and MacDougall, M. (1996). Students’ responses and approaches to
case-based instruction: The role of reflective self-regulation. Am. Educ. Res. J. 33: 719–752.
Evensen, D. (2000). Observing self-directed learners in a problem-based learning context:
Two case studies. In Evensen, D., and Hmelo, C. E. (eds.), Problem-Based Learn-
ing: A Research Perspective on Learning Interactions, Erlbaum, Mahwah, NJ, pp. 263–
298.
Evensen, D. H., Salisbury-Glennon, J., and Glenn, J. (2001). A qualitative study of 6 medical
students in a problem-based curriculum: Towards a situated model of self-regulation. J.
Educ. Psychol. 93: 659–676.
Faidley, J., Evensen, D. H., Salisbury-Glennon, J., Glenn, J., and Hmelo, C. E. (2000). How are
we doing? Methods of assessing group processing in a problem-based learning context. In
Evensen, D. H., and Hmelo, C. E. (eds.), Problem-Based Learning: A Research Perspective
on Learning Interactions, Erlbaum, Mahwah, NJ, pp. 109–135.
Ferrari, M., and Mahalingham, R. (1998). Personal cognitive development and its implications
for teaching and learning. Educ. Psychol. 33: 35–44.
Gallagher, S., and Stepien, W. (1996). Content acquisition in problem-based learning: Depth
versus breadth in American studies. J. Educ. Gifted 19: 257–275.
Gallagher, S. A., Stepien, W. J., and Rosenthal, H. (1992). The effects of problem-based learning
on problem solving. Gifted Child Q. 36: 195–200.
Gick, M. L., and Holyoak, K. J. (1980). Analogical problem solving. Cogn. Psychol. 12: 306–355.
Gick, M. L., and Holyoak, K. J. (1983). Schema induction and analogical transfer. Cogn. Psychol.
15: 1–38.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

264 Hmelo-Silver

Goodman, L. J., Erich, E., Brueschke, E. E., Bone, R. C., Rose, W. H., Williams, E. J., and Paul,
H. A. (1991). An experiment in medical education: A critical analysis using traditional
criteria. JAMA 265: 2373–2376.
Greeno, J. G., Collins, A., and Resnick, L. B. (1996). Cognition and learning. In Berliner, D.
C., and Calfee, R. C. (eds.), Handbook of Educational Psychology, Macmillan, New York,
pp. 15–46.
Hmelo, C. E. (1994). Development of Independent Thinking and Learning Skills: A Study of
Medical Problem-Solving and Problem-Based Learning, Unpublished Doctoral Disserta-
tion, Vanderbilt University, Nashville, TN.
Hmelo, C. E. (1998). Problem-based learning: Effects on the early acquisition of cognitive skill
in medicine. J. Learn. Sci. 7: 173–208.
Hmelo, C. E., and Ferrari, M. (1997). The problem-based learning tutorial: Cultivating higher-
order thinking skills. J. Educ. Gifted 20: 401–422.
Hmelo, C. E., Gotterer, G. S., and Bransford, J. D. (1997). A theory-driven approach to assessing
the cognitive effects of PBL. Instr. Sci. 25: 387–408.
Hmelo, C. E., and Guzdial, M. (1996). Of black and glass boxes: Scaffolding for learning and
doing. In Edelson, D. C., and Domeshek, E. A. (eds.), Proceedings of ICLS 96, AACE,
Charlottesville, VA, pp. 128–134.
Hmelo, C. E., Holton, D., and Kolodner, J. L. (2000). Designing to learn about complex systems.
J. Learn. Sci. 9: 247–298.
Hmelo, C. E., and Lin, X. (2000). The development of self-directed learning strategies in
problem-based learning. In Evensen, D., and Hmelo, C. E. (eds.), Problem-Based Learning:
Research Perspectives on Learning Interactions, Erlbaum, Mahwah, NJ, pp. 227–250.
Hmelo, C., Shikano, T., Bras, B., Mulholland, J., Realff, M., and Vanegas, J. (1995). A problem-
based course in sustainable technology. In Budny, D., Herrick, R., Bjedov, G., and Perry,
J. B. (eds.), Frontiers in Education 1995, American Society for Engineering Education,
Washington, DC.
Hmelo-Silver, C. E. (2000). Knowledge recycling: Crisscrossing the landscape of educational
psychology in a Problem-Based Learning Course for Preservice Teachers. J. Excell. Coll.
Teach. 11: 41–56.
Hmelo-Silver, C. E. (2002). Collaborative ways of knowing: Issues in facilitation. In Stahl, G.
(ed.), Proceedings of CSCL 2002, Erlbaum, Hillsdale, NJ, pp. 199–208.
Hmelo-Silver, C. E., and Barrows, H. S. (2003). Facilitating collaborative ways of knowing.
Manuscript submitted for publication.
Hmelo-Silver, C. E., and Barrows, H.S. (2002, April). Goals and strategies of a constructivist
teacher. Paper presented at American Educational Research Association Annual Meeting,
New Orleans, LA.
Kilpatrick, W. H. (1918). The project method. Teach. Coll. Rec. 19: 319–335.
Kilpatrick, W. H. (1921). Dangers and difficulties of the project method and how to overcome
them: Introductory statement: Definition of terms. Teach. Coll. Rec. 22: 282–288.
Kolodner, J. L. (1993). Case-Based Reasoning, Morgan Kaufmann, San Mateo, CA.
Kolodner, J. L., Hmelo, C. E., and Narayanan, N. H. (1996). Problem-based learning meets
case-based reasoning. In Edelson, D. C., and Domeshek, E. A. (eds.), Proceedings of ICLS
96, AACE, Charlottesville, VA, pp. 188–195.
Koschmann, T. D., Myers, A. C., Feltovich, P. J., and Barrows, H. S. (1994). Using technology
to assist in realizing effective learning and instruction: A principled approach to the use of
computers in collaborative learning. J. Learn. Sci. 3: 225–262.
Krajcik, J., Blumenfeld, P., Marx, R., and Soloway, E. (2000). Instructional, curricular, and
technological supports for inquiry in science classrooms. In Minstrell, J., and Van Zee, E.
H. (eds.), Inquiring Into Inquiry Learning and Teaching in Science, American Association
for the Advancement of Science, Washington, DC, pp. 283–315.
Krajcik, J., Marx, R., Blumenfeld, P., Soloway, E., and Fishman, B. (2000, April). Inquiry-
based science supported by technology: Achievement among urban middle school students.
Paper presented at the Annual Meeting of the American Educational Research Association,
New Orleans, LA.
P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

Problem-Based Learning 265

Lampert, M. (2001). Teaching Problems and the Problems of Teaching, Yale University Press,
New Haven, CT.
Leontiev, A. N. (1978). Activity, Consciousness, and Personality (M. J. Hall, Trans.), Prentice-
Hall, Englewood Cliffs, NJ.
Lesgold, A., Rubinson, H., Feltovich, P., Glaser, R., Klopfer, D., and Wang, Y. (1988). Expertise
in a complex skill: Diagnosing x-ray pictures. In Chi, M. T. H., Glaser, R., and Farr, M. J.
(eds.), The Nature of Expertise, Erlbaum. Hillsdale, NJ, pp. 311–342.
Linn, M. C., and Hsi, S. (2000). Computers, Teachers, Peers: Science Learning Partners, Erlbaum,
Mahwah, NJ.
Mennin, S. P., Friedman, M., Skipper, B., Kalishman, S., and Snyder, J. (1993). Performances
on the NBME I, II, and III by medical students in the problem-based and conventional
tracks at the University of New Mexico. Acad. Med. 68: 616–624.
Needham, D. R., and Begg, I. M. (1991). Problem-oriented training promotes spontaneous
analogical transfer. Memory-oriented training promotes memory for training. Mem. Cogn.
19: 543–557.
Norman, G. R., Brooks, L. R., Colle, C., and Hatala, H. (1998). Relative effectiveness of in-
struction in forward and backward reasoning. Paper presented at the Annual Meeting of
the American Educational Research Association, San Diego, CA.
Norman, G. R., Trott, A. D., Brooks, L. R., and Smith, E. K. (1994). Cognitive differences in
clinical reasoning related to postgraduate training. Teach. Learn. Med. 6: 114–120.
Novick, L. R., and Hmelo, C. E. (1994). Transferring symbolic representations across noniso-
morphic problems. J. Exp. Psychol. Learn. Mem. Cogn. 20: 1296–1321.
Novick, L. R., and Holyoak, K. J. (1991). Mathematical problem solving by analogy. J. Exp.
Psychol. Learn. Mem. Cogn. 17: 398–415.
O’Donnell, A. M. (1999). Structuring dyadic interaction through scripted cooperation. In
O’Donnell, A. M., and King, A. (eds.), Cognitive Perspectives on Peer Learning, Erlbaum,
Mahwah, NJ, pp. 179–196.
Palincsar, A. S., and Herrenkohl, L. R. (1999). Designing collaborative contexts: Lessons from
three research programs. In O’Donnell, A. M., and King, A. (eds.), Cognitive Perspectives
on Peer Learning, Erlbaum, Mahwah, NJ, pp. 151–178.
Patel, V. L., Groen, G. J., and Norman, G. R. (1991). Effects of conventional and problem-based
medical curricula on problem solving. Acad. Med. 66: 380–389.
Patel, V. L., Groen, G. J., and Norman, G. R. (1993). Reasoning and instruction in medical
curricula. Cogn. Instr. 10: 335–378.
Pea, R. D. (1993). Practices of distributed intelligence and designs for education. In Salomon,
G., and Perkins, D. (eds.), Distributed Cognitions: Psychological and Educational Consid-
erations, Cambridge University Press, New York, pp. 47–87.
Perfetto, G. A., Bransford, J. D., and Franks, J. J. (1983). Constraints on access in a problem-
solving context. Mem. Cogn. 11: 24–31.
Puntambekar, S., and Kolodner, J. L. (1998). The design diary: A tool to support stu-
dents in learning science by design. In Bruckman, A. S., Guzdial, M., Kolodner, J.,
and Ram, A. (eds.), Proceedings of ICLS 98, AACE, Charlottesville, VA, pp. 230–
236.
Ram, P. (1999). Problem-based learning in undergraduate instruction: A sophomore chemistry
laboratory. J. Chem. Educ. 76: 1122–1126.
Ramsden, P. (1992). Learning to Teach in Higher Education, Routledge, New York.
Salomon, G. (1993). No distribution without individual cognition: A dynamic interac-
tional view. In Salomon, G., and Perkins, D. (eds.), Distributed Cognitions: Psycholog-
ical and Educational Considerations, Cambridge University Press, New York, pp. 111–
138.
Salomon, G., and Perkins, D. N. (1989). Rocky roads to transfer: Rethinking mechanisms of a
neglected phenomenon. Educ. Psychol. 24: 113–142.
Schmidt, H. G., DeVolder, M. L., De Grave, W. S., Moust, J. H. C., and Patel, V. L. (1989).
Explanatory models in the processing of science text: The role of prior knowledge activation
through small-group discussion. J. Educ. Psychol. 81: 610–619.
 P1: FLT
Educational Psychology Review [jepr] pp1224-edpr-487801 May 22, 2004 0:50 Style file version June 4th, 2002

266 Hmelo-Silver

Schmidt, H. G., Machiels-Bongaerts, M., Hermans, H., ten Cate, T. J., Venekamp, R., and
Boshuizen, H. P. A. (1996). The development of diagnostic competence: Comparison of
a problem-based, an integrated, and a conventional medical curriculum. Acad. Med. 71:
658–664.
Schmidt, H. G., and Moust, J. H. C. (2000). Factors affecting small-group tutorial learning: A
review of research. In Evensen, D., and Hmelo, C. E. (eds.), Problem-Based Learning: A
Research Perspective on Learning Interactions, Erlbaum, Mahwah, NJ, pp. 19–51.
Schwartz, D. L., and Bransford, J. D. (1998). A time for telling. Cogn. Instr. 16: 475–522.
Schoenfeld, A. H. (1985). Mathematical Problem Solving, Academic Press, Orlando, FL.
Shikano, T., and Hmelo, C. E. (1996, April). Students’ learning strategies in a problem-based
curriculum for sustainable technology. Paper presented at American Educational Research
Association Annual Meeting, New York.
Steinkuehler, C. A., Derry, S. J., Hmelo-Silver, C. E., and DelMarcelle, M. (2002). Cracking the
resource nut with distributed problem-based learning in secondary teacher education. J.
Distance Educ. 23: 23–39.
Stepien, W. J., and Gallagher, S. A. (1993). Problem-based learning: As authentic as it gets.
Educ. Leadersh. 50(7): 25–29.
Torp, L., and Sage, S. (2002). Problems as Possibilities: Problem-Based Learning for K–12
Education, 2nd edn., ASCD, Alexandria, VA.
Vernon, D. T., and Blake, R. L. (1993). Does problem-based learning work?: A meta-analysis
of evaluative research. Acad. Med. 68: 550–563.
Vye, N. J., Goldman, S. R., Voss, J. F., Hmelo, C., and Williams, S. (1997). Complex math
problem-solving by individuals and dyads: When and why are two heads better than one?
Cogn. Instr. 15: 435–484.
Webb, N. M., and Palincsar, A. S. (1996). Group processes in the classroom. In Berliner,
D., and Calfee, R. (eds.), Handbook of Educational Psychology, MacMillan, New York,
pp. 841–876.
Wenger, E. (1998). Communities of Practice: Learning, Meaning, and Identity, Cambridge
University Press, New York.
White, B. Y., and Frederiksen, J. R. (1998). Inquiry, modeling, and metacognition: Making
science accessible to all students. Cogn. Instr. 16: 3–118.
Williams, S. M. (1992). Putting case based learning into context: Examples from legal, business,
and medical education. J. Learn. Sci. 2: 367–427.
Williams, S. M., Bransford, J. D., Vye, N. J., Goldman, S. R., and Carlson, K. (1993). Positive and
negative effects of specific knowledge on mathematical problem solving. Paper presented
at the American Educational Research Association Annual Meeting, Atlanta, GA.
Zimmerman, B. (2002). Becoming a self-regulated learner: An overview. Theory Pract. 41,
64–71.

View publication stats