JITE v37n1 - The Effect of Thinking Aloud Pair Problem Solvings (TAPPS) on the Troubleshooting Ability of Aviation Technician Students

Volume 37, Number 1
Fall 1999


The Effect of Thinking Aloud Pair Problem Solving (TAPPS) on the Troubleshooting Ability of Aviation Technician Students

Scott D. Johnson
University of Illinois at Urbana-Champaign
Shih-Ping Chung
USF Holland, Inc.

The teaching of cognitive skills is becoming more important as technology in the workplace becomes increasingly complex and ever-changing. Development of the cognitive skill of problem solving is of particular importance in workforce training. Generally speaking, there are two schools of thoughts for teaching problem solving: (a) teaching domain-specific knowledge, and (b) teaching general problem-solving strategies. Studies on the nature of human expertise in various disciplines tend to support the idea that specific domain knowledge is a primary factor in problem-solving skill ( Anzai, 1984 ; S.D. Johnson, 1988 ; Pfau & Murphy, 1988 ). Research shows that personal experience and knowledge accumulated in the field strongly affect an individual's capability to solve problems ( Bransford, Sherwood, Vye, & Rieser, 1986 ; Chi, Glaser, & Farr, 1988 ). In practice, learning and action are highly connected ( J.S. Brown, Collins, & Duguid, 1989 ; Ferguson & Hegarty, 1991 ; Lave, 1988 ; Rogoff, 1984, 1990 ; Scribner, 1984 ). The context in which a problem occurs must relate to the solver's knowledge domain in order for the problem to be recognized and solved properly. Several educational approaches, such as situated learning ( J.S. Brown et al., 1989 ) and cognitive apprenticeship ( Collins, 1989 ; Collins, J.S. Brown, & Newman, 1989 ; Rogoff, 1990 ), strive to connect context with domain-specific knowledge.

The second school of thought focuses on teaching general problem-solving strategies to students. Educators in this camp express concerns about the rigidity of domain-specific problem-solving skills and suggest that instructors teach flexible strategies and decontextualized general principles ( Andre, 1986 ; Feltovich, Spiro, & Coulson, 1989 ; Perkins & Salomon, 1989 ; Pestel, 1993 ; Spiro, Vispoel, Schmitz, Samarapungavan, & Boerger, 1987 ). The notion of transfer is a main theme for this group of educators. They advocate the importance of learning general problem-solving skills to deal with problems in a wide variety of knowledge domains ( Larkin, 1989 ). Methods of teaching general strategies in fields where the specific knowledge domain is highly relevant include analogy instruction ( Royer, 1986 ), examples learning ( Chi & Bassok, 1989 ), and case-based instruction ( Kolodner, 1993 ; Spiro et al., 1987 ).

In addition to the contrasting views on specific versus general approaches to problem-solving instruction, there are differing views about the best learning environment for acquiring problem-solving skills. Problem solving can be learned either individually or with a group. Although much problem-solving instruction is individual-based, instruction in cooperative, group problem-solving appears to provide additional benefits that enable learners to successfully deal with relatively complex tasks. Empirical studies on cooperative learning across various groups of learners and subject areas have shown it to be an effective instructional method for managing student learning ( Bossert, 1988 ; D.W. Johnson & R.T. Johnson,1989, 1994 ). Reciprocal teaching is a form of instruction that provides students with the opportunity to become familiar with the process of cooperation, while simultaneously stimulating their social reasoning skills and individual problem-solving skills ( Palincsar, 1986 ; Palincsar & A.L. Brown, 1984 ; Rubin & Forbes, 1984 ).

The Problem

Although there is an abundant literature base on problem solving, few studies address instructional methods for enhancing the unique problem-solving approach of technical troubleshooting. Within any industry, competent maintenance technicians can save their employers considerable money and time by keeping machines running. As service industries continue to expand, the competence of after-sale technical support personnel plays a critical role associated with a product's quality and the company's image. Troubleshooting is, therefore, a valuable skill, which shows not only the competence of an industry's workforce, but also the ability of a company to support its products.

A quasi-experimental study was designed to examine the effectiveness of an instructional strategy called Thinking Aloud Pair Problem Solving (TAPPS) for enhancing troubleshooting skill. TAPPS was used as an experimental treatment with university-level aviation technician students as they practiced troubleshooting common aircraft electrical system faults generated by a computerized tutoring system. The purpose of this study was to determine whether TAPPS improved the learners' ability to troubleshoot. The following research questions were developed to assess the impact of TAPPS on troubleshooting ability:

  1. Does TAPPS improve ability to troubleshoot electrical system faults?
  2. Is there a posttreatment difference between the TAPPS and control groups in the use of cognitive skills while troubleshooting?
  3. What ancillary training benefits result from the use of the TAPPS strategy?

Problem Solving Instruction and TAPPS

The important role of thinking, problem solving, and reasoning in individual learning and performance has been a priority of training for decades. The activity of problem solving consists of both a passive, reflective response to the context in which the problem occurs, and the active decision making that occurs when the problem solver is deliberately evaluating a situation.

Many instructional programs that emphasized problem solving in the 1950s through 1980s took a process model approach. These model-based programs taught problem solving by modifying a stage model of problem solving, such as Polya's four-step model. Most of these programs emphasized instruction on preestablished procedures for solving specific problems such as puzzles, games, or invented scenarios. Other instructional programs promoted the use of generally familiar knowledge from well-structured domains (e.g., mathematics, physics, engineering) to teach problem-solving procedures ( Bransford & Stein, 1984 ). Although these latter programs relied on a process model of problem solving (e.g., Bransford and Stein's IDEAL approach), they differed from the process-oriented programs in that they avoided the focus on puzzles and games and tended to emphasize application of general problem-solving principles to everyday life situations. By relying on set principles or rules, the problem solver could know what had been completed and what else needed to be done. However, since the emphasis was on following set rules, the problem solver failed to gain an awareness of the strategies used to solve the problems.

Rather than emphasize a linear problem-solving model that relies on specific procedures to be followed, current views on learning suggest that problem solving should be taught in a realistic context, with an awareness of the cognitive processes that are used to arrive at a solution. TAPPS is a form of reciprocal teaching that engages learners in deeper cognitive processing ( Glass, 1992 ; Lochhead & Whimbey, 1987 ; Pestel, 1993 ; Stice, 1987 ; Swartz, 1989 ). TAPPS has been used in a variety of disciplines, from mathematics and engineering to reading comprehension and general thinking skills. Following a dyadic-learning procedure, two students work as a team and take turns playing the role of the problem solver. The non-solving student assumes the role of monitor, and observes, critiques, and evaluates the problem-solving performance ( A.L. Brown & Palincsar, 1989 ; Greenfield, 1987 ). The monitor does not help the solver solve the problem. Instead, he or she observes the problem-solving process. The monitor reminds the problem solver to verbalize what he or she is thinking and doing while solving the problem and asks questions to clarify the process being described verbally by the problem solver. The goal is to make the problem solvers aware of what they know, what they can do, whether they are doing it correctly, and whether the process is reasonable. Even when the problem solver fails to solve the problem, both students can benefit by monitoring the thought processes involved. Through this technique, tacit thought processes can be brought into the open and observed by both learners. By taking turns solving different problems, the students not only learn to solve problems, but also learn to become more reflective problem solvers.

Verbalization is the key feature of TAPPS. The purpose of verbalization during the process of problem solving is to make the individual's inner thoughts explicit. The verbalization of inner thoughts reveals thought patterns and brings subconscious thought to consciousness, allowing the problem solver to monitor his or her chain of reasoning and identify errors. Vygotsky's ( 1934 ) inner speech theory of cognitive development provides a theoretical foundation for the TAPPS methodology. In Vygotsky's view, a child's self-directed talk helps to plan and direct activities; it helps the child solve problems. As a child grows, the egocentric speech will turn into nonverbal inner speech. Thus, gradually, people develop the ability to give themselves verbal instructions inwardly and silently, through inner speech. Evidence of this theory can be easily observed in the common occurrence of people murmuring to themselves while doing various tasks, without noticing they are making sounds.

Method

The TAPPS technique was used while students solved technical problems provided by a computer-coached practice environment called the Technical Troubleshooting Tutor (TTT). TTT is a computer-based expert system that utilizes several pedagogical principles discussed in cognitive science literature. TTT simulates realistic workplace problems and incorporates components of cognitive apprenticeship, such as coaching and fading, into its design. The TTT was originally used in a 1993 study that explored the effect of a computer-coached practice environment on individual learners' electrical troubleshooting skills ( S.D. Johnson, Flesher, Jehng & Ferej, 1993 ). That study was a quasi-experimental field study which involved 34 university students enrolled in a second level aircraft electrical systems course. Members of the experimental group (N=18) individually used the TTT, in addition to participating, with the control group (N=16), in the regular laboratory activities. At the end of the research period, all students were given realistic troubleshooting problems to solve, verbal protocols were collected, and observations made. The results showed that students who worked on the TTT became more effective and efficient troubleshooters. The TTT group showed a 78% improvement in actual troubleshooting success over the control group, with only 19% more troubleshooting practice.

The current study extended the 1993 TTT study by comparing the effect of individual TTT problem-solving practice to TTT practice that involved pairs of students using TAPPS. An intact student group, similar to the previous TTT group, served as the experimental TAPPS group. The TAPPS group of students received the same lecture and laboratory instruction, completed all course requirements, and used the same TTT system. The only difference between the TAPPS group in this study and the TTT group from the previous study was the use of TAPPS while the students solved system fault problems on the TTT. At the end of the study, each TAPPS subject individually participated in the same transfer task given to the subjects of the previous study.

Subjects

Data for this study were collected from two intact classes of students who were enrolled in a second level electronic systems course in the Institute of Aviation at the University of Illinois at Urbana-Champaign. The students enrolled in this course were working toward their Airframe and Powerplant certification from the Federal Aviation Administration. The course consisted of three hours of lecture and four hours of laboratory activity each week. Due to small enrollments and the fact that the course was offered only once each year, it was necessary to use an intact class. The first set of data came from the original TTT group ( S.D. Johnson et al., 1993 ), which served as the control group in this study. The original TTT group consisted of 18 students who worked on the TTT as a part of their lab exercises. This control group practiced troubleshooting by using the TTT on an individual basis. The second group of data was collected from the experimental TAPPS group, a similar group of students who were enrolled in the same class in a later semester. The TAPPS group consisted of 10 students who used the TAPPS strategy while working on the TTT. Students in both groups had completed prerequisite courses in fundamental aircraft electrical systems and powerplant systems. All subjects participated in the regular laboratory exercises, in which they were to troubleshoot various subsystem malfunctions with a electrical system simulator. The transfer task administered later in the study was performed on that same electrical system simulator.

Since the electrical systems troubleshooting course was only offered once each year by the Institute of Aviation, it was necessary to have the control and treatment groups participate in different years. To address the concern about the equivalency of the two student groups, demographic and aptitude indicators (i.e., age, previous electrical experience, prerequisite course grade, ACT scores, and high school rank) were compared. Statistical comparisons on these domain indicators revealed no significant difference between the groups. Based on these comparisons, it was assumed that the two groups were sufficiently equivalent. The equivalency of the two groups is further supported by lack of difference in the performance of the groups on a domain knowledge test that was administered late in the semester. Statistical comparisons of the total test scores and each of the subscales (i.e., items related to system structure, system function, and system behavior) showed no difference between the two groups.

Procedure

The TAPPS strategy was used while pairs of students worked on the computerized troubleshooting tutor. Two Macintosh computers containing the TTT software were available for student use during all scheduled laboratory periods. Prior to the start of the treatment, a one-hour demonstration and explanation of the TTT and TAPPS was provided for the subjects. A graduate research assistant observed each session when students worked on the TTT. The observer's role was to answer any questions related to the operation of the TTT and to take notes regarding the effectiveness of each pair in implementing the TAPPS procedure.

Each member of the TAPPS group was taught how to play the roles of problem solver and monitor. The problem solvers were instructed to talk through the complete troubleshooting process by saying everything they were thinking and doing while troubleshooting the fault provided by the TTT. They knew their partner would remind them to keep talking, would sometimes ask for clarification of their actions or decisions, but would not solve the problem for them.

Students who played the role of problem solver were asked to use the TTT to collect the information related to the problem (e.g., pilot compliant, problem symptoms, operating conditions, and system history) and read that information aloud. The problem solver would then identify as many potential faults as possible on a concept map provided by the TTT, and attempt to isolate the fault by checking each item on the potential fault list. Students who played the role of monitor were instructed to keep track of the problem solver's troubleshooting path. They were to follow the problem solver's logic, but not interrupt thinking processes. If the problem solver was stuck, the monitor could provide directions by asking the solver to recheck his or her previous assumptions or by reminding the solver of overlooked items on the potential fault list. After each problem was solved, the pair debriefed each other on the procedures that were used. This allowed the problem solvers to reflect on what they had just done, and the monitors to clarify the problem solvers' actions and suggest alternative solutions.

The TAPPS subjects had approximately six weeks during the semester to work through the TTT using TAPPS. Each pair had about 90 minutes each week to work on TAPPS during their scheduled laboratory work time.

Transfer Task

After completing the treatment, each subject participated in the same troubleshooting performance transfer task applied in the S.D. Johnson et. al. study ( 1993 ). To maximize the instructional effects of the course, the transfer task was conducted in the last four weeks of the sixteen-week semester. Each student was individually presented with an aircraft electrical system simulator in which four independent faults were inserted. The students were not told how many faults existed in the system, only that there were multiple faults inserted by the researcher. Verbal protocols were collected and analyzed to identify the cognitive processes used during troubleshooting. Treatment effects were examined by comparing performance on the transfer task.

The electrical system simulator consisted of 10 distinct independent subsystems found in a small aircraft's electrical system. Common aircraft components such as circuit breakers, switches, relays, terminal strips, conductors, a rotating beacon, and a power inverter were mounted on the board. Subjects were allowed to use common troubleshooting tools (e.g., multimeter and screwdriver), job-aids (i.e., schematic), and some spare parts (e.g., light bulbs and wires) during the transfer task. Audio recording equipment was used during the transfer task to collect each subject's verbalizations. Finally, an open-ended survey was used to collect opinions and feelings about the use of the TTT and TAPPS.

Data Analysis

Each subject's verbal protocol from the transfer task was transcribed and quantified with the data analysis procedure used in the S.D. Johnson et al. study ( 1993 ). The verbal protocols were first segmented into episodes, and then coded based on a seven-category coding system ( S.D. Johnson, 1989 ). Each episode consisted of (a) the selection of at least one potential fault, (b) the acquisition of information for the purpose of evaluating the proposed fault, (c) the interpretation of acquired information for the purpose of evaluating the fault, and (d) a decision to accept or reject the fault as the real fault. The coding scheme covered numerous behavioral and cognitive events: Action (ACT), Acquisition of Information (ACQ), Desired Information Not Obtained (DES), Interpretation (INT), Hypothesis (HYP), Decision (DEC), and Irrelevant (IRR). To test the validity of the coding system, a second coder used the same coding process to randomly select and code 10% of the total collected verbal protocols from the TAPPS group. The agreement coefficient rate between the two coders reached 88% consensus on coding the HYP statements, which were the primary factor used in this study for the measurement of the troubleshooter's cognitive activity.

After organizing each episode into a performance data matrix, troubleshooting performance was analyzed by fault and subject. Quantitative analysis provided indices of general performance, including problem recognition and solution data. A microanalysis examined episodic data such as hypothesis interpretation, strategy usage, and individual information acquisition efforts. Qualitative analysis focused on four types of information: (a) initial problem formulation, (b) development of problem space representations, (c) sequence through the problem space, and (d) identification of patterns within protocol episodes. Individual t-tests were conducted to examine differences between the groups on the variables of Problem Recognition, Correct Solution, Practice Time, and Simulated Work Time. All statistical tests reported in this paper were conducted with a significance level of a = .05.

Results

Impact of TAPPS on Troubleshooting Ability

To examine the impact of TAPPS on troubleshooting ability, each subject's troubleshooting performance, in terms of ability to locate the transfer task faults, was analyzed. This analysis revealed the accuracy of the troubleshooting that was performed by each group.

Ability to Recognize the Existence of Faults

A review of the verbal protocols revealed that the majority of the subjects in both groups began their problem finding activity by operating the toggle switches located on the control panel of the transfer task simulator. This initial general search typified the system operation checks that are common in aircraft maintenance, and represented appropriate troubleshooting behavior because no other symptomatic information had been provided. Although each subject had experience using the simulated electrical board during their laboratory sessions, several TTT subjects failed to recognize all four faults in the transfer task.

There was a significant difference in the ability of the TAPPS group (M = 4.00, SD = 0) and the TTT group (M = 3.61, SD = .50) to recognize that four faults existed in the electrical system, t(26) = 2.431, p < .05. Eleven of the eighteen TTT group subjects recognized all four faults, while seven (39%) failed to notice that the landing gear circuit was faulty. The inability of TTT subjects to recognize the fault in the gear indicator subsystem is likely due to the fact that the subjects' initial problem finding activity focused on the operation of the control panel switches and the gear indicator does not have a switch on the control panel. Overall, the TTT group recognized 90% of the faults. All of the TAPPS subjects recognized that there were four faults in the electrical system (100%). This difference was unexpected because the TTT did not require the subjects to engage in problem finding activity. At the start of each scenario, the Tutor group subjects knew that a fault existed and the symptomatic information that was available on request provided sufficient clues to narrow the fault down to a specific subsystem.

Ability to Locate Faulty Components

Data from the verbal protocols were also used to assess the ability of the subjects to locate the actual faulty components in the aircraft electrical system used for the transfer task. Correct solutions occurred when the exact faulty component was discovered and reported verbally in the protocols. Incorrect solutions resulted when a subject identified a component other than the true fault.

Although all the subjects were quite good at recognizing that faults existed within the system, several subjects had difficulty locating the actual faulty components. There was a statistical difference between the two groups in their ability to locate the actual faults within the system, t(26) = 2.632, p < .05. The TAPPS group (M = 3.70, SD = .48) outperformed the TTT group (M = 2.89, SD = .90) by locating significantly more of the faulty components. In terms of the percentage of problems solved, the TAPPS group solved 93% of the problems (37/40) while the TTT group solved 72% of the problems (52/72). The improved performance of the TAPPS group over the TTT group is impressive, especially when one considers that individual use of the TTT (without TAPPS) has already been shown to result in significantly improved troubleshooting performance over the more traditional form of laboratory training ( S.D. Johnson et al.,1993 ). As shown in the 1993 study, students who individually used the TTT showed a 78% improvement in actual troubleshooting success over the control group that received the same instruction, but practiced troubleshooting in a lab instead of on the TTT. With the simple pedagogical change from individual practice to the paired problem solving technique, an additional learning gain in troubleshooting performance has been demonstrated.

Impact of TAPPS on Use of Cognitive Skills

The process of troubleshooting is a series of cognitive processes. Captured in the subjects' verbal reports were fragments of their individual thoughts. This descriptive information was converted into quantitative data by assigning each fragment a meaningful code. Thinking skills can be roughly classified as lower order (retrieving superficial information) and higher order (combining or managing acquired information with existing knowledge). In this study, the higher order cognitive activities of interest involved hypothesis generation and testing, that is, a sequence of interpreting information, making assumptions, collecting information, and evaluating the current assumption ( Elstein, Shulman, & Sprafka, 1978 ; S.D. Johnson, 1989 ). Relevant cognitive activity data were gathered from the verbal protocols collected from the subject as they searched for faults during the transfer task. Analysis of the group performance included the examination of the group's percentage of correct hypothesis evaluation, percentage of incorrect hypothesis evaluation, and percentage of no hypothesis evaluation.

The TAPPS subjects (M = .92, SD = .04) outperformed the TTT group (M = .82, SD = .10) in their ability to correctly evaluate the fault hypotheses they generated. The TAPPS group correctly evaluated 90% of their fault hypotheses while the TTT group correctly evaluated 82%. This difference was statistically significant, t(26) = 2.842, p < .05. Incorrect hypothesis evaluations can mislead troubleshooters, which results in reaching an incorrect solution or wasting time by making extra troubleshooting attempts. The TAPPS group (M = .02, SD = .03) also had fewer incorrect hypothesis evaluations than the TTT group (M = .11, SD = .07), 2% vs. 22%. This difference was also statistically significant, t(26) = -4.058, p < .05. There were cases in the protocols where the subjects did not evaluate a potential fault hypothesis they had considered. This situation appeared to happen for several reasons: (a) the hypothesis was too trivial to be tested, (b) the subjects changed their focus to another more plausible hypothesis, or (c) the subjects were incapable of collecting sufficient information to support an evaluation. This occurred in seven percent of the hypotheses for the TAPPS group and eight percent for the TTT group. There was no statistical difference in the percentage of cases where a subject did not formally evaluate a generated hypothesis.

The examination of the impact of TAPPS on the use of cognitive skills suggests that TAPPS had a positive impact on the subject's ability to correctly evaluate troubleshooting hypotheses. These findings show that the TAPPS subjects were better able to evaluate the potential faults they considered, than were the TTT group subjects. The use of TAPPS seems to be an effective means to help learners improve their ability to evaluate potential faults in a system.

The Efficiency of TAPPS as a Training Strategy

Based on the results of this study, it appears that the TAPPS strategy adds to the learning experience of students. It also appears that TAPPS adds a measure of efficiency to the training environment. There were major differences between the two groups in their performance on the TTT. These differences appeared in the amount of time that it took them to complete the TTT and the number of problems each group solved.

The TTT was designed to take students through three categories of problems. The first category involved device failures (e.g., burned bulb, faulty voltmeter, broken switch). The second category of problems contained open circuits, while the third category involved short circuits. Within each of these categories was a hierarchy of problems, organized by level of difficulty. Students were given easier problems to solve when attempting new problem types. Each time a student solved a troubleshooting problem, the TTT assessed and compared the student's overall performance to a norm group based on 25 performance indicators. The total number of problems completed by each student before being advanced to the next problem type depended on the level of their individual performance. For example, if a student successfully solved five open circuit problems at a high level of performance, the TTT advanced the student to problems that involved short circuits. Students who displayed stronger problem solving skills needed to solve fewer problems on the TTT, and ultimately took less time to complete the TTT. Those students who experienced difficulties were cycled through additional problems by the computer. Students completed the TTT when they demonstrated competency on all problem types.

The TAPPS group was able to complete the TTT in an average of 4.3 hours, while the TTT group averaged 5.2 hours. While the TAPPS group was able to complete the TTT in less time, they were also able to solve more problems in the shorter time. The TAPPS group members solved an average of 34.4 problems apiece, with an average problem time of 7.5 minutes per problem (SD = 7.62). The TTT group members solved an average of 30.4 problems, with an average time of 10.4 minutes per problem (SD = 9.27). The ability of the TAPPS group to solve problems in a shorter amount of time than the TTT group was highly significant, t(717) = -3.656, p < .05. Although there was a difference between the two groups on the amount of time needed to complete the TTT; there was no difference in the amount of simulated work time as tracked by the computer. As the subjects attempted to solve each problem, they were "charged" for the cost of collecting problem-related information (e.g., the pilot's complaint, the maintenance history of the aircraft, operating conditions, access to service manuals or schematic diagrams, etc.), conducting various tests (e.g., sensory checks, voltage or continuity tests, etc.), and performing various actions (e.g., replacing parts or consulting an expert). The average simulated work time for the TAPPS group was 29.0 minutes per problem (SD = 56.95), while the TTT group averaged 28.6 minutes per problem (SD = 49.09). This difference was not significant, t(717) = -0.092, p > .05. This finding suggests that learning how to become a better troubleshooter can be accomplished in a shorter amount of time when working in a TAPPS pair, even though the problem solving time, as determined by simulated work time on the computer, was no different than that achieved by working alone on the TTT.

These data suggest that the TAPPS strategy adds a measure of efficiency to the learning process. The TAPPS subjects were able to complete the TTT in less time, and yet gain more experience solving problems than the TTT group. This finding is important for technical instruction in situations where there are limited workstations for students and limited time available for troubleshooting practice. The use of TAPPS allows an instructor to group students into pairs, which reduces the need for workstations by 50%. Also, the shorter time needed to complete the TTT problems could provide more time for other laboratory learning experiences. It is important to note that the efficiency gains of TAPPS is not a result of two students working together to solve the problems, an approach that would logically result in faster problem solution. When TAPPS is used correctly, only one student actually attempts to solve the problem, while the other student serves as an observer of the process.

Implementation Issues for Using TAPPS

Why is TAPPS a powerful pair-learning technique? The monitor is able to learn from the problem solver by listening and observing, and the problem solver receives help and advice from the monitor during and after solving the problem. By taking turns being the problem solver and the monitor, each member of the TAPPS pair can learn from the other's problem solving approach and strategies and also become more conscious of her or his own thinking process.

There was some concern prior to the start of the study that the subjects might be hesitant to talk aloud when a partner is observing them. This was not the case. According to the researcher's observations, the pairs who were most enthusiastic about working on the TTT system had the least amount of difficulty talking through a problem. Even those who at first seemed uncomfortable thinking aloud, quickly became accustomed to playing the role of problem solver. The surprising difficulty in the use of TAPPS involved the role of monitor. Being a monitor was not a natural task for most, if not all, of the subjects. The TAPPS monitor needs to be trained to be a listener and observer who attempts to interpret all of the problem solver's verbal cues. It was hard for the monitors to be an actively involved observer without attempting to solve the problem, especially when they thought they could do a better job themselves.

Discussion

TAPPS is a form of cooperative learning that emphasizes the impact of social interaction within pairs of learners on individual learning. Cooperative problem solving is common in the world of work. Teamwork, social interaction, professional cooperation, and inter-disciplinary collaboration are necessary characteristics of productive work environments. As such, these ideas need to be integrated into the technical curriculum, yet are easily ignored by instructors and trainers. The extra time needed to develop adequate instructional materials, the energy required to undertake group activity, and students' unpredictable attitudes toward different approaches may keep many instructors from experimenting with alternative instructional strategies, even though they may realize that their students could benefit from the change. Especially for technical instructors, insufficient information regarding effective instructional strategies that are directly related to the subject area makes the adoption of new strategies even more rare.

Does the use of TAPPS enhance troubleshooting skill? It seems to be the case. By utilizing a combination of cognitive learning theory and an alternative instructional strategy, the TAPPS approach appears to be an effective technique for courses that prepare technicians. The effectiveness was supported by the TAPPS group's better ability to make correct hypothesis evaluations and fewer misinterpretations of collected information. Supported by its empirical findings, this study provides a simple-to-implement approach that can directly improve technical students' analytic ability and enhance their troubleshooting performance. Ultimately, TAPPS seems to help learners become more confident and competent troubleshooters.

Another important finding was that the paired learning process enhances training efficiency by engaging two students in learning at the same time. The comparison of the TTT group's average of 30 problem solutions in just over five hours for each individual subject, with the TAPPS group's average of 34 problems in just over four hours for each pair demonstrates the efficient power of the technique. Not only did the pair learning process demonstrate its effectiveness in enhancing subjects' troubleshooting performance; it also seemed to be effective in exposing students to more problems in less time.

Because it is limited by its small sample size, the use of intact groups during different semesters, and the lack of control over the subjects' prior technical experience, the findings of the study cannot be generalized to other learning settings without further research. Although most technical training programs or formal technical education courses do not have large numbers of technician students in a class, it would be helpful to understand whether the TAPPS approach is manageable for larger classes. Further research is needed to explore the effect of TAPPS in different technical domains. It is also important to examine the impact of TAPPS when the technique is applied in a traditional instructional setting, instead of a supportive computerized practice environment. Longitudinal studies are also needed to understand how the skills learned through TAPPS are transferred and influence later performance in work settings.

Author

Johnson is an Associate Professor and Graduate Programs Coordinator, Department of Human Resource Education, University of Illinois at Urbana-Champaign.

Chung is a MIS Programmer/Analyst for USF Holland, Inc., Holland, Michigan.

References

Andre , T. (1986). Problem solving and education. In G. D. Phye & T. Andre (Eds.), Cognitive classroom learning: Understanding, thinking and problem solving (pp. 169-204). New York: Academic Press.

Anzai , Y. (1984). Cognitive control of real-time event-driven systems. Cognitive Science , 8, 221-254.

Bossert , S. T. (1988). Cooperative activities in the classroom. In E. Z. Rothkopf (Ed.), Review of research in education (Vol. 15, pp. 225-250). Washington, DC: American Educational Research Association.

Bransford , J., Sherwood, R., Vye, N., & Rieser, J. (1986). Teaching thinking and problem solving. American Psychologist , 41 (10), 1078-1089.

Bransford , J., & Stein, B. S. (1984). The IDEAL problem solver . New York: W. H. Freeman.

Brown , A. L., & Palincsar, A. S. (1989). Guided, cooperative learning and individual knowledge acquisition. In L. B. Resnick (Ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser (pp. 393-451). Hillsdale, NJ: Erlbaum.

Brown , J. S., Collins, A., & Duguid, P. (1989). Situated cognition and the culture of learning. Educational Researcher , 18 (1), 32-42.

Chi , M. T. H., & Bassok, M. (1989). Learning from examples via self-explanations. In L. B. Resnick (Ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser (pp. 251-282). Hillsdale, NJ: Erlbaum.

Chi , M. T. H., Glaser, R., & Farr, M. J. (1988). The nature of expertise . Hillsdale, NJ: Lawrence Erlbaum.

Collins , A. (1989). Cognitive apprenticeship and instructional technology (Tech. Rep. No. 474). Urbana-Champaign: University of Illinois, Center for the Study of Reading.

Collins , A., Brown, J. S., & Newman, S. E. (1989). Cognitive apprenticeship: Teaching the craft of reading, writing, and mathematics. In L. B. Resnick (Ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser (pp. 453-494). Hillsdale, NJ: Erlbaum.

Elstein , A. S., Shulman, L. S., & Sprafka, S. A. (1978). Medical problem solving: An analysis of clinical reasoning . Cambridge, MA: Harvard University Press.

Feltovich , P. J., Spiro, R. J., & Coulson, R. L. (1989). The nature of conceptual understanding in biomedicine: The deep structure of complex ideas and the development of misconceptions. In D. Evans & V. Patel (Eds.), The cognitive sciences in medicine (pp. 113-172). Cambridge, MA: MIT Press (Bradford Books).

Ferguson , E. L., & Hegarty, M. (1991, April). A hands-on advantage for mechanical understanding: Effects of real objects and diagrams on learning and problem solving . Paper presented at the meeting of the American Education Research Association, Chicago.

Glass , A. R. (1992). The effects of thinking aloud pair problem solving on technology education students' thinking processes, procedures, and problem solution . Unpublished doctoral dissertation, University of Minnesota, St. Paul.

Greenfield , L. B. (1987). Teaching thinking through problem solving. In J. E. Stice (Ed.), New directions for teaching and learning, no 30. Developing critical thinking and problem solving abilities (pp. 5-22). San Francisco: Jossey-Bass.

Johnson , D. W., & Johnson, R. T. (1989). Cooperation and competition: Theory and research . Edina, MN: Interaction Book.

Johnson , D. W., & Johnson, R. T. (1994). Learning together and alone: Cooperative, competitive, and individualistic learning (4th Ed.). Needham Heights, MA: Allyn & Bacon.

Johnson , S. D. (1988). Cognitive analysis of expert and novice troubleshooting performance. Performance Improvement Quarterly , 1 (3), 38-54.

Johnson , S. D. (1989). A description of expert and novice performance differences on technical troubleshooting tasks. Journal of Industrial Teacher Education , 26 (3), 19-37.

Johnson , S. D. (1994). Implications of cognitive science for technological problem solving. In D. Blandow, & M. J. Dyrenfurth (Eds.), NATO ASI Series F. Advanced educational technology in school-industry link projects (pp. 157-177). Berlin, Germany: Springer-Verlag.

Johnson , S. D., Flesher, J. W., Jehng, J., & Ferej, A. (1993). Enhancing electrical troubleshooting skills in a computer-coached practice environment. Interactive Learning Environments , 3 (3), 199-214.

Kolodner , J. (1993). Case-based reasoning (pp. x-xx). San Mateo, CA: Morgan Kaufmann.

Larkin , J. H. (1989). What kind of knowledge transfers? In L. B. Resnick (Ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser (pp. 283-305). Hillsdale, NJ: Erlbaum.

Lave , J. (1988). Cognition in practice: Mind, mathematics and culture in everyday life . Boston: Cambridge.

Lochhead , J., & Whimbey, A. (1987). Teaching analytical reasoning through thinking aloud pair problem solving. In J. E. Stice (Ed.), New directions for teaching and learning, No 30. Developing critical thinking and problem solving abilities (pp. 73-92). San Francisco: Jossey-Bass.

Palincsar , A. S. (1986). Metacognitive strategy instruction. Exceptional Children , 53 (2), 118-124.

Palincsar , A. S., & Brown, A. L. (1984). Reciprocal teaching of comprehension-fostering and monitoring activities. Cognition and Instruction , 1, 117-175.

Perkins , D. N., & Salomon, G. (1989). Are cognitive skills context-bound? Educational Researcher , 18 (1), 16-25.

Pestel , B. C. (1993). Teaching problem solving without modeling through "thinking aloud pair problem solving." Science Education , 77 (1), 83-94.

Pfau , H. E., & Murphy, M. D. (1988). Role of verbal knowledge in chess skill. American Journal of Psychology , 101, 73-86.

Rogoff , B. (1984). Instruction: Thinking and learning in social context. In B. Rogoff & J. Lave (Eds.), Everyday cognition: Its development in social context (pp. 1-8). Cambridge: Harvard University Press.

Rogoff , B. (1990). The cultural context of cognitive activity (Chapter 3). Apprenticeship in thinking: Cognitive development in social context . New York: Oxford.

Royer , J. M. (1986). Designing instruction to produce understanding: An approach based on cognitive theory. In G. D. Phye & T. Andre (Eds.), Cognitive classroom learning: Understanding, thinking, and problem solving (pp. 83-113). New York: Academic Press.

Rubin , D., & Forbes, D. (1984). Children's reasoning and peer relations. In B. Rogoff & J. Lave (Eds.), Everyday cognition: Its development in social context (pp. 220-237). Cambridge: Harvard University Press.

Scribner , S. (1984). Studying working intelligence. In B. Rogoff & J. Lave (Eds.), Everyday cognition: Its development in social context (pp. 9-40). Cambridge: Harvard University Press.

Spiro , R. J., Vispoel, W. P., Schmitz, J. G., Samarapungavan, A., & Boerger, A. E. (1987). Knowledge acquisition for application: Cognitive flexibility and transfer in complex content domains. In B. K. Gritton & S. M. Glynn (Eds.), Executive control processes in reading (pp. 177-199). Hillsdale, NJ: Lawrence Erlbaum.

Stice , J. E. (1987). Further reflections: Useful resources. In J. E. Stice (Ed.), Developing critical thinking and problem solving abilities. New directions for teaching and learning, No. 30. (pp. 101-109). San Francisco: Jossey-Bass.

Swartz , R. J. (1989). Making good thinking stick: The role of metacognition, extended practice, and teacher modeling in the teaching of thinking. In D. M. Topping, D. C. Crowell, & V. N. Kobayashi (Eds.), Thinking across cultures: The third international conference on thinking (pp. 417-436). Hillsdale, NJ: Erlbaum.

Vygotsky , L. S. (1934). Thought and language (A. Kozulin, Trans.). Cambridge: MIT Press, 1986.