Psychological Tools in Problem–based Learning

Cindy E. Hmelo-Silver

Ellina Chernobilsky

Maria Carolina DaCosta

Theoretical Framework

In this chapter, we will examine how cultural and psychological tools are integral parts of the Problem–based learning (PBL) process. Using both medical school and pre-service teacher education as contexts, we have examined students' ways of knowing—both what they learn and the learning processes they engage in. Examination of cognitive outcomes addresses how well students meet the three goals of PBL: (1) constructing a solid knowledge base, (2) becoming better reasoners, and (3) becoming self-directed learners. Cognitive outcomes were examined through the lenses of two major theoretical positions: transfer-appropriate processing theory and psychological tools theory. This chapter's goal is to show how both of these theories have provided useful frameworks for understanding cognition and learning in PBL contexts. As we shall discuss, the transfer-appropriate processing theory is a useful tool for understanding learning outcomes, but the psychological tools perspective has been more fruitful in understanding learning processes and outcomes.

Our initial work was guided by the transfer-appropriate processing theory, which states that people who learn in a problem-solving context should be able to retrieve that information when they need to (Adams et al., 1988). Spontaneous transfer of knowledge and strategies is generally hard to achieve; but with increasing practice and expertise, the likelihood of transfer is improved (Novick, 1988; Novick & Holyoak, 1991). Transfer often fails because problem solvers fail to retrieve an appropriate analog. Since in PBL the knowledge is encoded in a problem-solving context, students are more likely to retrieve that knowledge when faced with future problems, which is especially important in professional education. Students of professional education are often learning foundational disciplines (e.g., basic sciences for medicine and psychology in teacher education), and the goal of learning is often not to learn these disciplines in isolation but to be able to apply this knowledge to problem solving. As PBL students learn domain knowledge (e.g., basic biomedical sciences in the case of medical students), hypothesis-driven reasoning strategies,¹ and self-directed learning strategies in the context of solving problems, it is reasonable to expect transfer-appropriate processing mechanisms to come into play. In illstructured domains such as medicine and teaching, transfer is a particularly thorny issue because concepts apply irregularly across different problem situations (Spiro, Coulsen, Feltovich, & Anderson, 1988).

A more general cognitive analysis of PBL suggests that, as students are presented with problems, they access their prior knowledge, establish a problem space, search for new information to help reach their problem-solving goals, and in the process may construct new mental representations or restructure existing representations that include the conditions in which the knowledge might be used. This process involves developing metacognitive awareness of one's progress in both learning and problem solving (Hmelo & Lin, 2000).

In studies by Hmelo (1998) and Hmelo and Lin (2000), 75 first-year medical students (from both PBL and traditional medical curricula) were studied at three points over their first year of medical school and were given pathophysiological explanation problems to solve. Hmelo (1998) demonstrated that the PBL students were more likely to transfer their knowledge, reasoning, and learning strategies to new problems than the students in traditional medical education. This study also showed that the PBL students became more accurate in their diagnostic hypotheses on the new problems than their traditional counterparts. They also learned to produce more coherent explanations. Moreover, the study provided evidence that the PBL students were more likely to use science concepts in their problem solving—supporting a transfer-appropriate processing interpretation of the efficacy of PBL. Another finding that supported this interpretation is that the PBL students were more likely to transfer the reasoning strategies that were modeled to new problems. Hmelo and Lin (2000) compared these same students on their self-directed learning strategies. After each problem that the students solved, they were asked what else they would need to know in order to better solve the problem and how they would go about getting that information. This analysis found that the PBL students used their hypothesis-driven reasoning strategies to guide their self-directed learning. They were more likely to identify holes in their knowledge that were related to their hypotheses, whereas the “traditional” students were more likely to focus on clinical signs and symptoms. Moreover, the PBL students had better-defined plans for how to proceed with their learning than students from the traditional curriculum. At the end of the last problem-solving session, all the students were asked to read several paragraphs relevant to the case they were working on and then to generate a new explanation. The analysis of these explanations indicated that the PBL students integrated more of the new knowledge into their explanations than the traditional students, providing additional support that they had learned how to learn. Thus, this study demonstrated that the PBL students were able to apply their reasoning strategies to their self-directed learning—in other words, they approached their learning with their hypotheses in mind. They were able to transfer the metacognitive strategies of planning their learning to novel problems.

By applying a transfer-appropriate processing framework to these findings, we can conclude:

• Students who learn knowledge in a problem-solving context such as PBL are likely to retrieve and transfer their knowledge to new problems.
• Similarly, students who learn reasoning and self-directed learning strategies in a problem-solving context and have extensive practice in applying them are likely to retrieve and apply these strategies to new problems.

Both of these findings suggest that students are achieving the goals of constructing a solid knowledge base and becoming good reasoners and effective self-directed learners. This work provided a great deal of information about what students learn but not about how they learn, and it provided only limited information about how they use knowledge as a psychological tool to solve problems. Because of these limitations, we turned to another framework: that of psychological tools for learning. This allowed us to focus on the processes of learning as well as the outcomes. This has paralleled our theoretical evolution from a transfer-appropriate processing view (e.g., Adams et al., 1988; Perfetto, Bransford, & Franks, 1983) to more sociocultural theories that emphasize the role of tools in mediating activity (Engström, 1993; Lave & Wenger, 1991). The notion of mediated activity is a critical concept in sociocultural theories of learning (Engström, 1999; Kozulin, 1998). A key aspect of mediation is that people control their behavior by using and creating artifacts and other tools. Thus, studying the role of such tools is central to understanding learning. This evolution in thinking led us to reinterpret our earlier results by considering that the PBL students were more successful in their problem solving because they were able to use their science knowledge and the strategies that they developed as tools for their thinking. Kozulin (1998) defines psychological tools as “those symbolic artifacts—signs, symbols, tests, formulae, graphic—symbolic devices—that help individuals master their own ‘natural’ psychological functions of perception, memory, attention, and so on. Psychological tools serve as a bridge between individual acts of cognition and the symbolic sociocultural prerequisites of these acts” (p. 1). These tools mediate both individual and collective activities. It is important to note that tools are involved in a dialectic in which learners both transform and are transformed by the tools they use in their activities (John-Steiner & Mahn, 1996). In our studies of learning processes, we examined the role of several kinds of psychological tools, which are classified as conceptual tools (e.g., knowledge, strategies, language) or representational tools that students construct (and that may be used to scaffold and guide their learning). By scaffolding, we mean providing support that learners could not do without in order to accomplish tasks (Hmelo & Guzdial, 1996; Vygotsky, 1978). Such support might help model and communicate a process, elicit articulation and reflection, or provide other forms of coaching.

Conceptual Tools in PBL

Knowledge, strategies, and language help mediate goal-directed activity by helping one make inferences and reason about one's activities. Students use language as a tool to help them construct meaning. Among the various psychological tools, language plays a critical role in the development of cognition (Vygotsky, 1978). The use of language allows its further development and subsequent mastery. As Lave and Wenger (1991) pointed out, this mastery allows students to progress in becoming participants in the sociocultural practices of communities.

PBL provides many opportunities for students to engage with conceptual tools such as language and domain knowledge. Adequate language practice is essential for being a part of a community of practitioners—a group of people who share goals, ideas, and interests in order to solve similar problems. Through participation and discussion, practitioners have a chance to appropriate and manipulate newly acquired vocabulary, negotiate word meanings, and interact with other members of the community (Brown et al., 1993). PBL allows for such talk to take place. As students work in small comfortable settings, they have a chance to both share what they have learned and find out what they still need to learn. Such talk is an ideal environment for students to appropriate the conceptual language of a discipline as they practice it and have a chance to learn from their mistakes.

In addition to language, one's knowledge and strategies can serve as important tools for problem solving. In PBL, students appropriate new knowledge and strategies as they engage with problems. This is different from acquisition of content knowledge because implicit in the notion of tools is the instrumental value of this knowledge (Kozulin, 1998). Knowledge is only a tool if “it is appropriated as a generalized instrument capable of organizing individual cognitive and learning processes in different contexts and in applications to different tasks” (p. 86). A psychological tools perspective suggests that some reinterpretation of the Hmelo (1998) and Hmelo and Lin (2000) results is needed. From this perspective, the notion that students in the PBL curriculum are using their science concepts and hypothesis-driven reasoning strategies to produce good-quality explanations on a variety of problems suggests that these concepts and strategies are serving this instrumental purpose. In addition, the hypothesis-driven strategies that they use in their reasoning also serve to mediate their self-directed learning. This occurs because learners can use their hypotheses as a way to make principled judgments about the relevance of new information for the problem at hand. In other words, this approach helps by constraining the set of possible concepts that might be explored as well as allowing learners to evaluate the value of the information with respect to the hypotheses that they are considering (Hmelo & Lin, 2000). For example, a learning issue related to a disease leads students to consider abnormal laboratory values in the context rather than as an isolated feature, as in this example: “the physiology of the adrenal gland: what are the compounds which it synthesizes, and what are the systemic effects of their release into blood in abnormally elevated levels?” (p. 237). This example refers to a patient with an adrenal tumor. We can also see the effect of using hypothesis-driven reasoning as a tool in another example:

If Ms. Dupree does indeed have MG (and she did improve with anticholinesterase, which is used to determine a diagnosis of MG), most of her symptoms are characteristic of this disease, including ptosis, difficulty swallowing, and respiratory distress. The pathophysiological process that accounts for her symptoms is that insufficient amounts of Ach are binding to neural receptors (Hmelo & Lin, 2000, 244).

This student was basically testing her hypothesis of MG (myasthenia gravis, a neuromuscular disorder) using textual information and then inferring that the causal mechanism accounted for this patient's problems. At the same time, she was engaging in knowledge construction that connected what she already knew to new information. Thus, she had used her reasoning and learning strategies as tools for knowledge construction.

Representational Tools in PBL

In addition to conceptual psychological tools, representations can serve as tools for students' thinking. Constructing representations is an important social practice for learning. Different representations afford and constrain social knowledge construction in several ways (Pea, 1993; Roth, 1998). First, representations serve as a shared concrete referent for all the members of a group and provide common ground for negotiation. Second, the structure of the representation can guide student discussions (Suthers & Hundhausen, 2001). In PBL, several representational artifacts may be constructed. One representation is the formal structured PBL whiteboard with its facts, ideas or hypotheses, learning issues, and action plan. This helps guide the discourse to consider certain issues and not others. The whiteboard serves as an external memory for the students—it reminds them of their ideas, both solidified and tentative, as well as hypotheses that they need to test. One ritualized aspect of the PBL tutorial is “cleaning up the boards” (Hmelo-Silver, 2002a). This occurs several times, but in particular after students have discussed the resources they used for their self-directed learning. Here students evaluate each of their hypotheses, look at the fit to data, and determine how that maps onto what they have gleaned from their self-directed learning. The discussion of what hypotheses are viable, which ones are more or less likely, leads to substantive discussions that are centered around what needs to be filled in on the whiteboard (see example in Figure 1). Students often discuss how hypotheses should be ranked or when they should be added or deleted. Similar discussions revolve around learning

issues (Hmelo-Silver & Barrows, 2003). The formal whiteboards serve as a focus for students to negotiate their ideas and identify those that need to be held in abeyance. When students mark something as needing to be entered onto the whiteboard, it also signifies agreement by the group that the item is worth attending to. The use of the whiteboard is a fluid part of the tutorial that supports reasoning, knowledge construction, and self-directed learning as students use it to remind them of what they are considering, what they know, and what they still need to learn.

Other representational tools students may construct are less formal representations, such as flowcharts, concept maps, and diagrams. For example, students may draw a flowchart or anatomic diagram as shown in Figure 2. In the next section, we will demonstrate how these tools support reasoning and knowledge construction.

Illustrations of Psychological Tools in PBL

PBL in Medical Education

In medical education, we have conducted detailed analyses of students' representational activity and discourse as key components of social knowledge construction. In a fine-grained analysis of a second-year PBL tutorial group, we examined how the act of creating inscriptions served as a tool for both representing and transforming the group's understanding (Hmelo-Silver, 2002a, b). To examine this, we attempted to identify focal representational events in the five hours of video that we analyzed of medical students and an expert facilitator. We found that there were very few that stood out as independent events but rather they were clearly an integral part of the discussion, serving as a tool in their deliberations.

The major representational episode that we identified occurred late in the tutorial process. One strategy that a facilitator may use to help students integrate their knowledge is asking them to draw a diagram (Hmelo-Silver, 2002a). The students in the video decided on a particular epistemic form, a flowchart, that subsequently guided and helped consolidate their theory construction. An epistemic form is a structure that supports knowledge construction through the use of a general structure that guides inquiry (Collins & Ferguson, 1993). The particular form that the students chose helped guide their construction of a coherent causal explanation. In this 20-minute episode, the students drew both a flowchart of the causal processes involved in a patient's problem (pernicious anemia) and a diagram of the nervous system to help them locate the source of the patient's physical complaints (shown in Figure 2). As the students made connections between different levels of explanation (e.g., between biochemical and clinical phenomena), they engaged in a great deal of causal explanation and metacognitive thinking.

In the following excerpt, the group was moving back and forth in their drawing activity between representing anatomy and physiology and clinical signs and symptoms. At this point, they were considering how what is happening in the neuron could cause paralysis. The scribe (Jeff) was drawing the lower left part of the flowchart and beginning the anatomic diagram on the right. As he was doing this, Mary noted: “And you get neuronal also, um, various things that happen. I believe you get neuronal cell swelling within the membrane and then you can get neuronal death. And that's when you get the paralysis and once it progresses to that stage, as we know, neurons will regenerate.” Thus, as Jeff was drawing those two parts of the representation, Mary was noting how what is happening at the anatomic level leads to an effect observed at the clinical level. The students continued to make these bridges between the causal mechanisms and the symptoms. The need to complete an aspect of the drawing often led to substantive discussions. It also served as a referent that the students gestured toward in their negotiations. For example, as the students were trying to locate the source of a problem at the cellular level, they frequently referred to and pointed at the diagram:

So here Jim was trying to understand where the nucleotides were with respect to the flowchart and Jeff was trying to get a handle on how that was related to the neurons (nerve cells). There is some ambiguity about what was going on in the blood cells versus the nerve cells. Mary and Jim used their gestures and the flowchart to help disambiguate these two structures. The students continued to use their collaborative representational activity to make their understanding coherent and to fill in most of the blanks. As the students were drawing, they elaborated their understanding. When the scribe was switching levels in the drawing, for example between the biochemistry and the clinical signs, the students engaged in causal reasoning. In addition, the parts of the anatomic drawing (on the right side of Figure 2) became issues that the students realized they needed to learn more about. This demonstrates how the representational tools provided a concrete reference for negotiation that helped guide the students' thinking and provided an external memory to guide and support problem solving.

PBL in Teacher Education

Our discussion of tools in PBL switches context here from medical education to teacher education. Using PBL in teacher education provides special, additional challenges. Rather than having one facilitator for every group, there were one or two facilitators for six to seven groups. With the facilitator not available at all times, some of the facilitation function was offloaded onto the representations that students used—in particular, the whiteboards (or in this case, large 2.5 x 2 foot stick-on notes). Thus, in this context, the external tools served to scaffold the development of the internal tools. In preparing pre-service teachers, examination of the artifacts that they produced has provided a great deal of insight into their thinking and collaborative activity, as well as how the artifacts served as a mechanism for knowledge diffusion (Hmelo-Silver, 2000).

The context for this study was an undergraduate educational psychology class. Analysis of student artifacts (the group whiteboards and papers) provided evidence that the students crisscrossed the landscape of educational psychology as they deepened their understanding of the psychological concepts. The group whiteboards were divided into four columns, for facts (about the case), ideas (about the cause of the problem and potential solutions), learning issues, and action plan. The students initially used the whiteboards in small-group discussions; then the groups were brought together to share their work. The students used the whiteboard in their own group as a scaffold—a reminder of what they needed to attend to. It provided an external memory so that important ideas did not get lost. It also provided a mechanism for ideas to spread out among the groups as the whiteboards were always visible to every group. During the whole-class reporting, ideas were diffused among the groups. For example, during the fourth problem of the semester, only one group had “information processing” as a learning issue. After the large-group discussion, this idea spread to all the groups and appeared in their papers. Thus, the whiteboards and the ideas contained on them helped support social knowledge construction. This is an example of how representational tools support the development of conceptual tools.

Moreover, the students' knowledge became increasingly differentiated as they applied it to subsequent problems. The first time the students applied a concept, it was often used as a way to display their knowledge, 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 (Hmelo-Silver, 2000, 51).

They went on to provide more details about long-term memory, but all of this discussion was abstract—they were not putting the knowledge to work as a psychological tool to help them solve their problem (in this case, choosing among several theory-based approaches to instruction). In their final problem in the course, the students actually applied the information processing concepts to the problem:

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]).² A concept map is extremely important because teachers use students' prior knowledge to explain and discuss increasingly more sophisticated concepts. “Prior knowledge becomes 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 integral tool that will allow them to make more meaningful connections when they see the experiment (Hmelo-Silver, 2000, 53).

Here the students were using the concepts as a tool for thinking about how they would redesign a lesson. They were applying information processing theory to describe the importance of prior knowledge in learning, indicating that they were thinking deeply about memory. These results suggest that for students to use knowledge as a tool for problem solving they need to engage with it.

In a subsequent study, we examined the artifacts for evidence that students' use of professional vocabulary became richer and more flexible as they engaged in a Problem–based course in educational psychology (Chernobilsky & Hmelo-Silver, 2002). Language is an important conceptual tool because it helps shape thinking. Our approach was to examine language changes using written narratives that students constructed, both in groups and individually. The decision to focus on narrative artifacts was based on the ideas of Bakhtin (1981) and Vygotsky (1978), who viewed speech and writing as a coherent whole with a dynamic interaction between them. Examining language expressed in writing is one way to assess the level of knowledge that students constructed during learning activities (Bereiter & Scardamalia, 1987).

We examined how students developed professional language in an educational psychology course for pre-service teachers over the course of seven problems that the students worked on in a single semester. In order to examine language change over time, we concentrated on such constructs as professional jargon, definitions used, and identification of theories and theorists as indicators of professional language. We also examined students' use of sources because the skill of locating sources and critically evaluating their reliability is an indication of how students use professional sources as tools for thinking (Wineburg, 1991). For this analysis, we looked at a number of group and individual artifacts. Group artifacts included whiteboards, concept maps, and group-generated papers. Students' learning logs (reflective journals) and midterm and final examinations constituted individual artifacts. The quantitative analysis of the data showed an overall linear trend in the growth of vocabulary usage in the students' individual work. The use of professional language in the group artifacts showed a similar but nonsignificant trend. Group data also indicated that groups identified many more concepts and used more vocabulary than the students did individually. This supports the sociocultural notion that when students work in groups their work is more elaborated than when done individually (Hmelo-Silver, 2000).

To capture the dynamics of group learning in PBL and to better understand the quantitative results, we examined three representative groups for qualitative trends. This analysis indicated that strong, well-functioning groups that devoted time to scaffolding weaker students performed better in comparison to other groups. It also showed that weaker groups who spent a lot of time resolving issues that were not related to their learning task (e.g., conflicts among group members) and did not engage in peer scaffolding did not show the same increase in the sophistication of vocabulary and idea development as other groups. These analyses show that language learning is tied to group dynamics and to the situations students create while working in groups. PBL allows for maximizing of a positive environment for learning and at the same time allows students to internalize their thinking.

While looking at language learning and development during PBL activities is useful, one needs to keep in mind that language does not exist in isolation. Examining language development together with knowledge in PBL adds another dimension to our understanding of how students construct and apply psychological tools.

To examine how pre-service teachers' knowledge changed, we studied their individual and group artifacts generated during a single PBL problem (Hmelo-Silver, DaCosta, & Chernobilsky, 2002). In this problem, students observed a videocase of real-life instruction and suggested ideas for redesigning the instruction. Looking at student-generated artifacts during a PBL activity allowed us to examine how students used knowledge as a tool in their problem solving, as we did in Hmelo-Silver (2000). Because we had made some changes to the PBL model to include some initial and final individual phases in addition to the group work, we were able to examine how individual use of knowledge was related to the knowledge applied in the group phase of the activity. We had two goals for the analysis of the individual and group artifacts that students generated. The first was to analyze how concepts flowed from individual to group artifacts and vice versa. This was an attempt to understand the role of shared knowledge in developing ideas. The second goal was to examine how students dealt with relevant concepts and whether they achieved more sophisticated levels of knowledge development during the activity.

We coded individual and group artifacts for the educational psychology concepts represented in the case and for the level of knowledge development they engaged when applying such concepts. Knowledge development ranged from an undeveloped telling level to a more advanced transforming level, with elaborated telling being an intermediate level between the two (Bereiter & Scardamalia, 1989). Concepts coded at the transforming level were those that were properly elaborated and applied to the case. Knowledge transformation implies that students are using knowledge as a tool in their problem solving. Unelaborated concept naming with no connections to the case was coded as just telling. The intermediate level of elaborated telling occurred when students provided conceptual explanations but still no connection to the case. The results of our analysis demonstrated that groups differed in levels of shared knowledge and knowledge development while solving the problem. Greater proportions of shared knowledge (i.e., transfer of concepts across individual and group artifacts) seemed to have contributed to higher levels of knowledge development during the activity. For example, the group that engaged mostly in knowledge transforming was also the group with the highest proportion of shared knowledge. At the same time, groups without sufficient shared knowledge before solving the problem often engaged in knowledge telling, just listing or mentioning the concepts without applying them to the problem. More effective groups were those in which members were able to establish a common and rich knowledge base, which allowed them to go beyond telling to transforming. The sooner a common knowledge base was established in a learning community, the more opportunities students had to elaborate and deepen their understanding. The shared knowledge served as a tool by creating common ground that the students could build on and apply toward solving the problem.

To summarize, the structural representations in PBL pre-service teacher education take an even bigger role than in other PBL settings. As Pea (1993) pointed out, language (which is a conceptual tool in itself) and representational tools are the means through which intelligence gets distributed and discussed. By making ideas and thought visible to students, such tools help extend working memory capacity, which now can be occupied with higher-order thinking and more knowledge construction. As a result, knowledge, which is also a conceptual tool, is growing, and this allows students to build their professional base and to engage in active social participation in the community of practitioners.

Discussion and Implications

Psychological tools are a key component of PBL. They work for students as well as helping students accomplish their problem solving and understanding. Initially, students treat concepts as isolated entities that have names to be learned and defined. Our data from both medical students and pre-service teachers demonstrate that, as students participate in PBL, they learn to put their knowledge to work and the language they appropriate allows them to engage with that knowledge in productive ways. They learn to use professional language in their discourse to help them understand and solve meaningful problems. This helps them to see the big picture and to engage in constructing a solid knowledge base, and it provides opportunities for them to see how knowledge can be applied. The feedback they get in their PBL tutorials helps them refine their language and subsequently their ideas. In addition, using appropriate professional language supports their self-directed learning, as they are better able to find what they need from their learning resources. It is important for teachers using PBL to understand that part of their role is modeling professional language, at times revoicing students' ideas in the appropriate terminology. Although it is important not to be focused just on vocabulary, this more subtle demonstration of how language is used can support students' knowledge construction and enhance their effectiveness as self-directed learners.

As our research has demonstrated, these tools help students achieve the goals of PBL. In addition to conceptual and linguistic tools, student-generated artifacts serve as psychological tools to help remind students where they have been and where they are going. The whiteboard structure can help guide the discourse in productive ways that support social knowledge construction. The whiteboard reminds students of their hypotheses and serves as a focus for generating and testing ideas, a process that supports the development of reasoning skills. As students try to generate and evaluate ideas with each other, they are forced to justify their thinking and explain their reasoning. The “learning issues” column makes it salient that gaps in their knowledge are valued as opportunities to learn, thus supporting the development of self-directed learning skills. In addition, the structured whiteboards and other representations that students construct become tools for them to externalize their thinking and can become a focal point for them to negotiate meaning and revise their understanding, as we saw in the case of the medical students. Finally, because these representations are public and available to other groups, they promote social knowledge construction as ideas diffuse across a classroom. Other representations, such as concept maps and flowcharts, can also be used to help synthesize a coherent understanding.

This research has implications for practice. It is important to realize that merely posing a problem does not in and of itself guarantee learning. In the example of medical education, there was one facilitator for every five to seven students. In the case of the pre-service teachers, we used a combination of large-group and small-group discussions and the teacher served as a wandering facilitator. This model makes it harder to encourage students to think deeply and to model appropriate professional language. An important issue is to consider how the use of psychological tools can be modeled and scaffolded in such settings. Providing procedural facilitations is one approach that we have used (Hmelo-Silver, 2002a). In this approach, students rotate the facilitator role. The student-facilitators are given prompt cards, as shown in Figure 3, with questions that they might ask at different points in the PBL process, as well as explaining the reason for asking the question.

The whiteboards themselves have served as useful tools in the wandering facilitation model. Using large pieces of newsprint or whiteboards for each group serves to make students' thinking visible, quite literally. With a quick scan about the room, the facilitator can determine from the boards how groups are progressing in their problem solving and what their knowledge gaps are. This allows wandering facilitators to scaffold more effectively by allocating their effort as needed and adapting to the needs of different groups.

Varying the structure of the whiteboard is another approach that can help scaffold the development of psychological tools. In our current work, we have modified the whiteboard and structured the PBL process to specifically support analysis and redesign of classroom cases. Students are prompted to think about particular aspects of instruction in the way that a skilled facilitator might, if one were constantly available (Steinkuehler, Derry, Hmelo-Silver, & DelMarcelle, 2002). Approaches such as this offer

promise by using representations as psychological tools to help support the development of conceptual tools. Teachers might think about how the labels on whiteboard columns might model a problem-solving process or support the development of conceptual tools in specific subjects.

In summary, the various psychological tools discussed in this chapter have direct applications to practice in achieving the goals of PBL. The teacher has an important role in asking questions that help students understand how conceptual tools can support their problem solving. The whiteboards can be used to extend the facilitation—by providing a model of problem solving that can support the development of reasoning, knowledge construction, and self-directed learning. The generic whiteboard structure with its hypotheses, facts, and learning issues makes it applicable in a variety of contexts. On the other hand, when trying to facilitate many groups, the facilitator may wish to adapt the whiteboard for a particular domain. In other words, teachers may want to specialize the representational tools to help support the development of conceptual tools in particular subjects. Psychological tools have many important roles in the PBL process, but these may need adaptation to one's unique context.

Acknowledgments

This research was partially funded by a National Academy of Education/Spencer Foundation Postdoctoral Fellowship to the first author and a National Science Foundation ROLE grant #107032 to Sharon J. Derry and Cindy E. Hmelo-Silver. Any opinions, findings, and conclusions or recommendations expressed in this chapter are those of the authors and do not necessarily reflect the views of the National Science Foundation.

References

Adams, L., Kasserman, J., Yearwood, A., Perfetto, G., Bransford, J., & Franks, J. (1988). The effect of fact versus problem oriented acquisition. Memory and Cognition, 16, 167–75.

Bakhtin, M. (1981). Discourse in the novel (C. Emerson & M. Holquist, Trans.). In M. Holquist (Ed.), The dialogic imagination: Four essays by M. M. Bakhtin (pp. 252–422). Austin: University of Texas Press.

Bereiter, C., & Scardamalia, M. (1987). The psychology of written composition. Hillsdale, NJ: Erlbaum.

Bereiter, C., & Scardamalia, M. (1989). Intentional learning as a goal of instruction. In L. B. Resnick (Ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser (pp. 361–92). Hillsdale, NJ: Erlbaum.

Brown, A. L., Ash, D., Rutherford, M., Nakagawa, K., Gordon, A., & Campione, J. C. (1993). Distributed expertise in the classroom. In G. Salomon (Ed.), Distributed cognitions: Psychological and educational considerations (pp. 188–228). New York: Cambridge University Press.

Chernobilsky, E., & Hmelo-Silver, C. E. (2002). Learning to talk the educational psychology talk through a Problem–based course in educational psychology. Paper presented at the PBL2002 conference. Baltimore, MD, June.

Collins, A., & Ferguson, W. (1993). Epistemic forms and epistemic games: Structures and strategies to guide inquiry. Educational Psychologist, 28, 25–42.

Engström, Y. (1993). A developmental study of work as a test bench of activity theory: The case of primary care medical practice. In S. Chaiklin & J. Lave (Eds.), Understanding practice: Perspectives on activity and context (pp. 64–103). Cambridge: Cambridge University Press.

Engström, Y. (1999). Activity theory and individual and social transformation. In Y. Engström, R. Miettinen, & R. Punamaki (Eds.), Perspectives on activity theory (pp. 19–38). New York: Cambridge University Press.

Hmelo, C. E. (1998). Problem–based learning: Effects on the early acquisition of cognitive skill in medicine. Journal of the Learning Sciences, 7, 173–208.

Hmelo, C. E., & Guzdial, M. (1996). Of black and glass boxes: Scaffolding for learning and doing. In D. C. Edelson & E. A. Domeshek (Eds.), Proceedings of ICLS 96 (pp. 128–34). Charlottesville, VA: Association for the Advancement of Computing in Education.

Hmelo, C. E., & Lin, X. (2000). Becoming self-directed learners: Strategy development in Problem–based learning. In D. H. Evensen & C. E. Hmelo (Eds.), Problem–based learning: A research perspective on learning interactions (pp. 227–50). Mahwah, NJ: Erlbaum.

Hmelo-Silver, C. E. (2000). Knowledge recycling: Crisscrossing the landscape of educational psychology in a Problem–based learning course for preservice teachers. Journal of Excellence in College Teaching, 11, 41–56.

Hmelo-Silver, C. E. (2002a). Collaborative ways of knowing: Issues in facilitation. In G. Stahl (Ed.), Proceedings of CSCL 2002 (pp. 199–208). Hillsdale, NJ: Erlbaum.

Hmelo-Silver, C. E. (2002b). Getting the big picture: Discourse, representation, and reflection in a tutorial group. Unpublished manuscript.

Hmelo-Silver, C. E., & Barrows, H. S. (2003). Facilitating collaborative ways of knowing. Cognition and Instruction, in press.

Hmelo-Silver, C. E., DaCosta, M. C., & Chernobilsky, E. (2002). Knowledge construction using online videocases. Paper presented at the Fifth International Conference on the Learning Sciences. Seattle, WA, October.

John-Steiner, V., & Mahn, H. (1996). Sociocultural approaches to learning and development: A Vygotskian framework. Educational Psychologist, 31, 191–206.

Kozulin, A. (1998). Psychological tools. Cambridge, MA: Harvard University Press.

Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. New York: Cambridge University Press.

Novick, L. R. (1988). Analogical transfer, problem similarity, and expertise. Journal of Experimental Psychology: Learning, Memory, and Cognition, 14, 510–20.

Novick, L. R., & Holyoak, K. J. (1991). Mathematical problem solving by analogy. Journal of Experimental Psychology: Learning, Memory, and Cognition, 17, 398–415.

Pea, R. D. (1993). Practices of distributed intelligence and designs for education. In G. Salomon (Ed.), Distributed cognitions: Psychological and educational considerations (pp. 47–87). New York: Cambridge University Press.

Perfetto, G. A., Bransford, J. D., & Franks, J. J. (1983). Constraints on access in a problem-solving context. Memory and Cognition, 11, 24–31.

Roth, W.-M. (1998). Inscriptions: Toward a theory of representing as social practice. Review of Educational Research, 68, 35–60.

Spiro, R. J., Coulsen, R. L., Feltovich, P. J., & Anderson, D. K. (1988). Cognitive flexibility theory: Advanced knowledge acquisition in ill-structured domains. In Proceedings of the Tenth Annual Conference of the Cognitive Science Society (pp. 375–83). Hillsdale, NJ: Erlbaum.

Steinkuehler, C. A., Derry, S. J., Hmelo-Silver, C. E., & DelMarcelle, M. (2002). Cracking the resource nut with distributed Problem–based learning in secondary teacher education. Journal of Distance Education, 23, 23–39.

Suthers, D., & Hundhausen, C. (2001). Learning by constructing collaborative representations: An empirical comparison of three alternatives. In P. Dillenbourg, A. Eurelings, & K. Hakkarainen (Eds.), Proceedings of the First European Conference on Computer-supported Collaborative Learning (pp. 577–84). Maastricht: Universiteit Maastricht.

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press.

Wineburg, S. S. (1991). Historical problem solving: A study of the cognitive processes used in the evaluation of documentary and pictorial evidence. Journal of Educational Psychology, 83, 73–87.

1 Hypothesis-driven strategies involve reasoning based on one's theoretical framework and assumptions, such as a hypothesized disease process or psychological theory, versus a focus on situational features such as signs and symptoms or characteristics of the problem setting.

2 The students were referring to the knowledge web at www.wcer.wisc.edu/step, which is a hypermedia web site on learning science instruction. See Steinkuehler, Derry, Hmelo-Silver, and DelMarcelle (2002) for further details.