study of collaborative patterns and prociency developmentjcal_407 448..461 L.O. Hll,* T. Sderstrm,* J. Ahlqvist & T. Nilsson *Department of Education Ume University, S-901 87 Ume, Sweden Oral and Maxillofacial Radiology, Department of Odontology, Ume University, S-901 85 Ume, Sweden Abstract This article is about collaborative learning with educational computer-assisted simulation (ECAS) in health care education. Previous research on training with a radiological virtual reality simulator has indicated positive effects on learning when compared to a more conven- tional alternative. Drawing upon the eld of Computer-Supported Collaborative Learning, we investigate collaborative patterns, their causes, and their implications for learning. We investi- gate why the extent of application of subject-specic terminology differs between simulation training and more conventional training. We also investigate howthe student-simulator interac- tion affordances produce collaborative patterns and impact learning. Prociency tests before and after training, observations during training, and interviews after training constitute the empirical foundation. Thirty-six dentistry students volunteered for participation. The results showed that not only the task but also the medium of feedback impacts the application of subject-specic terminology. However, no relation to prociency development was revealed. We identied turn-taking as well as dominance patterns of student-simulator interaction but again found no relation to prociency development. Further research may give us deeper insights into if and how these collaborative patterns, in other respects, impact collaborative learning with ECAS in health care education. Keywords collaboration, computer, education, health care, learning, simulation. Introduction Applications of educational computer-assisted simula- tions (henceforth called ECAS) are spreading within health care education (Issenberg et al. 2005; Nehring & Lashley 2009). This has been described as in part an effect of a paradigm shift of educational models, from an apprenticeship model for learning based on experi- ence to a model based on expertise (Luengo et al. 2009), although other practical factors are likely to play signi- cant roles (Gaba 2004). A primary function for the ECAS is to ensure patients safety through ensuring practitioners competence when training on actual patients, and is limited or non-existent (ibid.). Previous studies have demonstrated both positive as well as nega- tive effects on skill or prociency development, often in comparison to available, more traditional alternatives (e.g. Engum et al. 2003; Quinn et al. 2003; Nilsson et al. 2007; Sderstrm et al. 2008). Newtechnologies have a tendency to attract the atten- tion of educators. Some of which read much hope into Accepted: 28 September 2010 Correspondence: Lars O. Hll, Department of Education, Ume Uni- versity, S-901 87 Ume, Sweden. Email: lars-olof.hall@pedag.umu.se doi: 10.1111/j.1365-2729.2011.00407.x Original article 448 2011 Blackwell Publishing Ltd Journal of Computer Assisted Learning (2011), 27, 448461 them for the realization of their educational ideals, while others view them as threats to the current educa- tional canon (Cuban 2001; Dillenbourg et al. 2009; Slj 2010). The tension between enthusiasts and critics can create an initial dichotomy of general statements about the technologies effect on education and black box comparisons of new versus old media. This ten- dency can be traced in the earlier research of health care simulation as well (e.g. Abrahamson et al. 1969; Ovasa- pian et al. 1988; Chopra et al. 1994; Peugnet et al. 1998). However, the tension between enthusiasts and critics can also drive the research to be more nuanced and practice-oriented studies where the technique of using the technologies enter focus. Later research on health care simulations seem to have shifted to this focus. (Gaba 2004; Issenberg et al. 2005; McGaghie et al. 2010). It is to this research that we want to contrib- ute by adding collaborative learning to the equation. From a Computer-Supported Collaborative Learn- ings (CSCL) perspective, collaborative activities may be put forward as a framework technique to improve training efciency with this type of technology (e.g. Crook 1994; OMalley 1995; Koschmann 1996; Dillen- bourg et al. 2009). Collaboration is, however, not con- sidered to support learning automatically, and a general CSCL question for Dillenbourg is thus under which conditions is collaborative learning effective? with the sub-questions under what conditions do specic inter- action patterns occur, and what interaction patterns predict learning outcomes? (ibid. p. 6). These general questions are valid also for this study of learning with an ECAS in dentistry education. Within the research project Learning Radiology in Simulated Environments, we have studied collabora- tive learning with one type of ECAS, a screen-based virtual reality (VR) simulator 1 (Sderstrmet al. 2008). In a prior paper, we presented comparative data on pro- ciency development and collaborative patterns for stu- dents training with the simulator (the SIM group), as contrasted to students working with a more conven- tional PowerPoint-based exercise (the CON group). In short, we found that while collaboration in the CON groups seemed to be more in line with our ideals, e.g. with more inclusive peer discussions, more thorough interpretations of what is shown on screen and more extensive application of subject-specic terminology, it was the SIM groups that developed more with regard to prociency development (ibid.). Others have argued that students may work with software without much reection or change in their conceptual understanding, and argued for the benets of thorough dialogue (e.g. Pilkington &Parker-Jones 1996; Pilkington 1998). This discrepancy thus intrigued us and is part of our motiva- tion to keep analyzing this data. In this article, we will continue the analysis of col- laborative learning with the radiological VR simulator. It is written in the intersection between the empirical research eld of learning with ECAS within health care education and the general research eld of CSCL. Based on Dillenbourgs general questions, two themes will be investigated, described in the next section along with the empirical questions about observed collabora- tive patterns, their causes, and their impact on learning for dentistry students training collaboratively with the radiological VR simulator. The rst theme is about a specic collaborative pattern, the application of subject-specic terminology during the training session. In sociocultural perspec- tives on learning, action with cultural tools or medita- tional means is one central concept (e.g. Wertsch 1998; Slj 2010). These tools can be physical artefacts such as an X-ray machine, as well as mental systems such as parallax technique for interpreting X-ray photographs. Some argue that a key aspect of adult learning, such as professional learning of the practice of radiological examinations, lies in the mastery of the cultural tools required for competent action within this practice. This mastery entails not only developing competence in applying specic tools but also being able to choose appropriate tools and design solutions to specic tasks (Slj 1999, 2000, 2010). Verbal interaction is a recur- ring focus of sociocultural studies because it can reveal participants use of mental tools and collaborative foci (e.g. Enyedy 2003; Mercer 2005), as well as non-verbal actions and interactions have proven to be valuable ele- ments of observation when researching health care edu- cation (e.g. Hindmarsh 2010). In this article, the application of subject-specic terminology is regarded as a visible indicator of students trying to use cultural tools from specic subject areas. The extent to which groups in SIM and CON apply subject-specic termi- nology during training was analysed, and it was found that SIM groups tend to apply them to a signicantly lesser degree (Sderstrm et al. 2008). The empirical questions for this theme are (1) [W]hat causes the extent of application of subject-specic terminology to differ Collaboration and educational simulations 449 2011 Blackwell Publishing Ltd between groups working with the ECAS (SIM) and groups working conventionally (CON)? and (2) [W]hat, if any, implications does the extent of application of subject-specic terminology have for prociency development? The second theme is about one of the fundaments of ECAS, the physical interaction with the simulator itself, and how this impacts collaboration and learning. The physical studentsimulator interaction is essential for enabling students to inquire or discover (e.g. Lazonder et al. 2010) that which is simulated, or to learn to performit (e.g. Engumet al. 2003). With a screen-based simulator such as the one under scrutiny here, manoeu- vring the simulator became a central activity. While all participants are able to see the screen at any given moment, they cannot all manoeuvre it at the same time. This separates it fromsome full body computer assisted mannequins, or integrated simulators (Bradley 2006), where participants interact with different parts of the simulator. Manoeuvring the simulator means control- ling what is shown on screen and has thus a direct impact on what visual information is available to the group. It is thus justied to investigate how the physical student-simulator interaction affordances impact col- laboration and learning. The empirical questions for this theme are (3) [A]re there distinguishable patterns of how students share control over the simulator?; (4) [D]o the patterns of control over the simulator on the group level, or time spent manoeuvring the simulator on the individual level directly impact prociency develop- ment?; (5) [D]o specic patterns of control over the simulator produce specic patterns of verbal interac- tion and does individual control over the simulator also entail control over the verbal interaction?; (6) [A]re the verbal patterns or individual control over the verbal interaction directly related to prociency develop- ment?; and (7) [H]owis the group collaboration around the studentsimulator interaction perceived by the stu- dents themselves? This article contributes to the growing interest of techniques for using ECAS in health care education. Drawing upon the eld of CSCL we will contribute to the investigation of collaborative learning as such, a technique with focus on relationships between cultural tools, interactive patterns, and learning. How does the transition from one tool (MS PowerPoint) to another (screen-based VR simulation) impact the application of other tools (subject-specic terminology), peer collabo- ration and learning outcomes? Can we discern charac- teristics of the tools central to these effects? What are the implications for designing for collaborative learning with screen-based simulations in health care education and for future research on the subject? Method This study was built around a quasi-experimental core. Participants were recruited from a population of under- graduate students in the dentistry program at Ume University, Sweden, attending a course in Oral and Maxillofacial Radiology. Thirty-six students 20 female and 16 male volunteered. The core of the design was that (1) [I]nitially, all participants performed a prociency test. (2) [A]n experimental group (SIM) and a control group (CON) were formed, with 18 stu- dents in each group, matched to be comparable on the prociency test. Within the respective group, six work groups of three students were randomly created, result- ing in six SIM groups and six CON groups. (3) [T]he workgroups received training on the subject of object localization using parallax principles, i.e. analysing multiple X-ray photographs of a jaw in order to identify and position potential anomalies in relation to generic anatomical objects such as specic teeth. During this training SIM used a screen-based simulator and CON used a set of MSPowerPoint slides. These sessions were recorded on video. (4) [A]ll students performed a second prociency test after training. Later on 18 participants, nine from each group, participated in follow-up interviews. This design is illustrated in Table 1. Each of these elements is explained further in the next section. The prociency tests The students prociency in interpreting radiographs was evaluated before and after the exercises. The instru- ment used, a prociency test, was developed by two dental scientists who teach at the dentistry program. It measured what right now counts as subject prociency in this course at the dentistry program. The tests consisted of three subtests: a principle test, a projection test, and a radiography test; each part graded from 08 giving a total of 24 points. The principle subtest aimed at assessing the participants understand- ing of the principles of motion parallax. The projection 450 L.O. Hll et al. 2011 Blackwell Publishing Ltd subtest aimed at assessing the participants ability to apply the principles of motion parallax. Based on basic sketches, this subtest requires basic understanding of anatomy. The radiography subtest aimed at assessing the participants ability to locate object details in authentic radiographic images utilizing motion paral- lax. The participants were asked to report the relative depth of specied object details in pairs of radiographs. The prociency analyses in this study are based on the total score from all three tests. Training sessions and simulator The training session settings were comparable for the SIM and CON groups, with the exception of software and computer peripherals. The training tasks focused on object localization using the principles of parallax. The work groups (with three students in each group) were located in front of a PC equipped with either the radio- logical VRsimulation software, illustrated in Fig 1, or a set of PowerPoint slides with images and written tasks and subsequent answers, illustrated in Fig 2. Please note that while Fig 1 illustrates single-user simulation train- ing, the training discussed in this article was performed in workgroups with three students in front of the screens. A passive teacher was available during the 60-min-long sessions. The simulator was basically a PC equipped with simulation software, two monitors, and some special peripherals. Special glasses were worn by all group members to enhance the simulators 3D effects. One of the monitors represented 3D models of a patient, camera, and lm. The other represented virtual X-ray photographs. The control peripherals used for interac- tion include a standard keyboard, a 3D mouse, and a pen-like controlling device. While the keyboard is used scarcely, the pen and the mouse are central to manoeu- vring the simulator. Using the simulator, the students were able to col- laboratively perform real-time radiographic examina- tions of patients teeth. It allowed the students to position the three-dimensional model of the patient, the X-ray tube, and the lm. X-ray images could then be Table 1. Design of the study. Input Process Output Variable Pre-training prociency Training SIM Post-training prociency Training CON Experience Evaluation of variable Prociency test Observation Prociency test Survey Interviews SIM, simulator; CON, control. Fig 1 Illustration of a single-user training with the radiological VR simulator. During actual training, workgroups of three students collaborated in-front of these screens. Ange andring i proojekion mellan bild (1), (2), och (3). Lagesbestam mesiodensen i forhallande till angransande tanders rotter. Fig 2 Example of an MS PowerPoint slide from the CON group training. The textual feedback given in this image read as follows. Line 1 states the alteration in projection between image (1), (2), and (3). Line 2 locates the mesiodens in relation to the roots of adjacent teeth. Collaboration and educational simulations 451 2011 Blackwell Publishing Ltd exposed at will by students, immediately presented by the simulator as geometrically correct radiographs ren- dered from the positions of the models. Four types of exercises centred on object localization were included. In the fourth exercise, it was possible to view the two- dimensional X-ray image change in real time as the model was manipulated. For further technical specica- tions and discussion of validity, see Nilsson (2007). Video-recorded observation Like many other researchers of health care education (e.g. Hindmarsh 2010; Koschmann et al. 2010) and educational health care simulations (e.g. Rystedt & Lindwall 2004), we have used video-based observations to capture learning in practice. To enable analysis and comparisons of group interaction during training, the training sessions were recorded using a DV camera. Twelve groups meant twelve one-hour recordings. The camera was placed so that the upper half part of the stu- dents was visible while neither the computer screen nor the teacher was visible. The analysis of group interaction during training was performed through two phases. In phase one, three questions were posed to a number of randomly chosen video recorded training sessions. These questions resulted in detailed descriptions from which thematic categories of group interaction were inferred. These questions were, [W]hat are the participants talking about? How are they talking about it? How do they relate to each other and to the learning environment as a whole? When no more categories were found, i.e. satu- ration had been reached, phase one ended. In phase two, all video data was split into one-minute time segments and coded with the previously abstracted thematic categories. In our descriptive presentations of the observations (such as Tables 2 and 4, or Figs 3 and 4), these time-segments are our empirical unit of observation. The categories were limited to whether the content of the conversation was action-oriented, interpretive, technology-focused, meta-reective, theo- retical or social; whether or not subject-specic termi- nology applied; which member was manoeuvring the simulator; which members were active in the conversa- tion; which, if any, member had command over the verbal space. This allowed us to conduct a highly struc- tured analysis based on an understanding that was inuenced by the current set of data. The coding of video recorded sessions was per- formed by one of the researchers. In order to produce a measure of the coding stability (Krippendorff 2004), one of the sessions was re-coded, by the same researcher, and compared with the original for each cat- egory described above. The per cent agreement between original coding and re-coding was 97% for content, 92% for terminology, and 98% for manoeuvre, verbal space, and verbal activity, respectively. Follow-up interviews After an initial analysis, follow-up interviews were per- formed with one half (18) of the participants. The aim was to better understand the training experience in SIM, how it differs from CON and to better understand the observed differences between the two. With open-ended questions, we talked about the respondent as a learner, training impact on learning, the respondent as a group member, the tasks, realism and functionality. These interviews included 18 participants in total, nine from the respective group. They were performed individually with an effective time usage of 3050 min, during a period of 2 weeks. Video recordings of the respondents training session were played on a laptop computer to support recall. All interviews were recorded on tape. A qualitative approach was adopted in the analysis focus- ing on inferring categories of responses, sometimes called meaning concentration (Kvale & Brinkmann 2009), rather than a quantitative count of howmany said what, i.e. we posed specic questions to each transcript, extracted the responses related to it, and inferred catego- ries of ideas from these responses. The end product is illustrated in Table 2, where categories of respondent ideas about reasons for using subject-specic terminol- ogy is presented along with quotes that fall into each cat- egory. A selection of results will be presented as the different categories of responses identied within the SIM group or in the SIM and CON groups. Quotes of student responses have been translated from Swedish into English. Results and analysis Application of subject-specic terminology As previously mentioned, there was a signicant difference between SIM and CON with regard to the 452 L.O. Hll et al. 2011 Blackwell Publishing Ltd frequency with which they applied a subject-specic terminology as opposed to subject-non-specic termi- nology (Sderstrm et al. 2008). The question we pose here is, what causes the extent of application of subject- specic terminology to differ between groups working with the ECAS (SIM) and groups working convention- ally (CON)? The rst step was asking the students themselves, through interviews, why they were using the subject- specic terms during training. As a group, the students presented a few thematic reasons, not all of them com- patible with one another, for using these terms. These are presented in Table 2 and include (1) The terms enable communication; they are the only words they know (for projection and anatomy). (2) Students all share the same understanding of the terms. (3) The terms make communication easier and more efcient. The alternative, using everyday words, would be more complicated. (4) Teachers have previously encou- raged students to use them. (5) The solution to the tasks is supposed to be presented in these terms (CON only). When trying to understand what caused differences between SIM and CON, there are two particularly interesting themes in these responses. The rst is the idea put forward in CON that the solution was sup- posed to be presented in subject-specic terms (4). As one of the CON students puts it, [example (5) in Table 2], Because in the answer you are supposed to respond, this is more distal or more mesial. That makes it easier if you communicate in that way from the beginning. It was, however, not the case that the students were required to produce a solution in the subject-specic terms. The students in the CON groups were, however, required to describe changes in projection between images and to describe the relation among certain objects. As indicated by the responses, the students found it much easier to do so by using terms for projection and anatomy than by using everyday lan- guage. However, the feedback after each task was in CON, the correct solution if you will, given as one- sentence descriptions in subject-specic terms. This suggests that for the students, the correct answer may tend to be identical to the one written by the teacher. The SIM students were also required to interpret projection and locate objects but, and this is a signi- cant difference, they were tasked to manoeuvre the model based on this interpretation. And the feedback they were given is in part visual, a correct positioning of the model was shown, and numerical, the distance or angle between the students position and the correct position was shown in numbers. If the medium of feedback impacts the process, then this would explain part of the difference in the application of subject- specic terminology. It would also indicate that if Table 2. Categorization of respondents answers to the question of why they use the academic terminology during training. Inferred category Example quote Present in group (1) Terms enable communication (1) We cant just say that this tooth is . . . we have to say that it is distal on this tooth. That is the term we have learned for that trajectory. Both (2) These are the only words we have. It would not work without them. Both (2) Students share understanding of the terms Because we have learned them all three of us know what we are talking about, because it is correct. If we say buccal everyone in the group understands. Both (3) Terms make communication efcient It is easier to work with my classmates if we can use these terms. There is a lot of anatomy we have to learn. Both (4) Prior teachers orders We wouldnt have if the teachers hadnt forced us to start using them. It is the language you are supposed to use when working and writing. Both (5) The solution is to be presented in these terms Because in the answer you are supposed to respond: this is more distal or more mesial. That makes it easier if you communicate that way from the beginning. CON n = 18. CON, control. Collaboration and educational simulations 453 2011 Blackwell Publishing Ltd SIM students were given textual feedback similar to that in CON, then there would be an increase in their use of these terms. The second thing to note in the responses is that the subject-specic terms were associated with anatomy and projection, as illustrated here by the example quotes for (1) and (3) in Table 2. In other words, these terms were especially relevant when trying to interpret changes in projection and to locate objects. They were not necessarily central for manoeu- vring the simulator. This suggested that there should be a positive relationship in the observational data between what students talked about and the terminol- ogy they used for talking about it. As illustrated by Table 3, such relations do exist. When the content of the verbal interaction was interpretations, SIM students applied subject-specic terms to a much greater extent, as compared to when the content was action-oriented (64.5% of cases compared to 11.4%, chi square 93.8; P < 0.001). This means that the differences in tasks between SIM and CON, which encourage more focus on action for SIM, impact the application of subject-specic termi- nology. However, when comparing the interpretive content of SIM and CON, it is apparent that CON students still apply subject-specic terms much more often (90.5% of all interpretive instances compared to 64.5% for SIM, chi square 46.4; P < 0.001). We con- clude that while differences in tasks impact the extent of application of subject-specic terminology, this does not explain all of the difference between the SIM and CON. We suggest that the medium of feedback is also important. The second question for this theme is, what, if any, implications does the extent of application of subject- specic terminology have for prociency development? Table 4 illustrates how often SIM groups applied subject-specic terminology (as the percentage of the observational time-segments in which they occur), in relation to group prociency development. It reveals that the group that most often applies subject-specic terms, SIM 2, also develops most (14 points); it also reveals that the group that most rarely applies these terms, SIM 4, develops the least (-2 points). These two cases do, however, seem to be exceptions to a pattern of randomness, because there is no relation between the two factors in the remaining four groups (Spearmans r 0.6; P > 0.05). We conclude that the extent of appli- cation of subject-specic terminology have little direct impact on prociency development. Table 3. Comparison of how verbal content and applied terminology co-appear for SIM and CON groups. Training Verbal content Subject-specic terminology Not subject-specic terminology CON Interpretation 90.5% 9.5% Action-oriented 10% 90% SIM Interpretation 64.5% 35.5% Action-oriented 11.4% 88.6% Percentage based on the total number of observational time segments where the category is identied. CON, control; SIM, simulator. Table 4. Extent of application of subject- specic terminology in relation to group prociency development, divided by SIM group. Group Subject-specic terminology Group prociency development SIM 1 41% 2 SIM 2 46% 14 SIM 3 37% 10 SIM 4 10% -2 SIM 5 28% 9 SIM 6 30% 3 Percentage based on the total number of time segments for the respective group. SIM, simulator. 454 L.O. Hll et al. 2011 Blackwell Publishing Ltd Student-simulator interaction affordances manoeuvring the simulator Are there distinguishable patterns for how students share control over the simulator?As illustrated in Fig 3, the observations revealed two types of collaboration around the student-simulator interaction. The rst one, illustrated by SIM 1, SIM 2, SIM 4, and SIM 5, is char- acterized by manoeuvral turn-taking. Each group member got a chance to manoeuvre the simulator, albeit not for an equal amount of time. The second one, illus- trated by SIM3 and SIM6, is characterized by manoeu- vral dominance where one of the members operated the simulator almost all the time. Acommon characteristic of the manoeuvral dominance groups was that the member sitting in the middle, right in front of the com- puter, was the one manoeuvring the simulator. When asked, through interviews, if the choice of seat was con- scious, the dominant female operating the simulator in SIM 6 claimed, I decided to sit in the middle . . . I want to have control and take command. Most students claim, however, that it was not conscious. We conclude that SIM groups differ in their sharing of control over the simulator and that there are two distinguishable col- laborative patterns, turn-taking and dominance. Do the patterns of control over the simulator on the group level, or time spent manoeuvring the simulator on the individual level, directly impact prociency devel- opment? As illustrated by Table 5, there are better and worse developments for groups characterized by turn- taking (2, 14, -2, 9 points of development, respectively) as well as dominance (10, 3 points of development, respectively). On the individual level, there seems to be no clear-cut relationship between time spent manoeu- vring the simulator and prociency development (Spearmans r 0.1; P > 0.05). The operators in the Fig 3 Illustration of the time spent by each group member manoeuvring the simulator. Members of each group are labelled after their position in front of the computer: left, middle, or right. Collaboration and educational simulations 455 2011 Blackwell Publishing Ltd dominance groups, SIM 3 and SIM 6, developed a little more than their next-best group member (2 and 1 points better, respectively). And the primary operators in the turn-taking groups developed equally (SIM 2) or less (SIM 1, 4, 5) in comparison to their group members. In one of the groups, SIM 5, the member with the least amount of manoeuvring-time developed much more than her peers (7 points in comparison to 2 and 0). We conclude that neither the patterns of control over the simulator nor the time spent manoeuvring the simulator directly impact prociency development. Do specic patterns of control over the simulator produce specic patterns of verbal interaction, and does individual control over the simulator also entail control over the verbal interaction? Figure 4 illustrates howoften a given member has command over the verbal space in the respective SIM group. Beginning with the turn-taking SIM 1, the verbal pattern was signicantly different from the manoeuvral pattern with one person dominating the verbal space throughout the training session. In the similar turn-taking group SIM5, a differ- ent pattern appears, a distinct dyadic pattern with two members in tight communication. Moving to SIM 4, a manoeuvral dominance group, we nd that the member that operated the simulator throughout the training session rarely took command of the verbal space. But in the similar SIM6, the dominant operator also took clear command of the verbal space. We conclude that specic patterns of control over the simulator do not produce specic patterns of verbal interaction. Nor does indi- vidual control over the simulator also entail control over the verbal interaction. We nd two patterns of verbal interaction, one dyad-ish and one based on dominance. Are the verbal patterns or individual control over the verbal interaction related to prociency development? Regarding SIM 2, 3, 4, 5 as variations of a dyadic pattern with one more or less dominant member, and SIM 1 and 6 as variations of a dominance pattern, there is no clear-cut relation between verbal pattern and pro- ciency development. Developments in the dominance groups were weak (2, 3 points, respectively), and some- what varied in the dyad-ish groups (14, 10, -2, 9, respectively). The fact that SIM 1 and SIM 6 were char- acterized by having one particularly strong member who developed very little with regard to prociency test scores (1 and 2 points, respectively) would also suggest that dominance does not ensure the individuals development, and also seems to do little to help the other members. However, because the verbally shy member Table 5. Manoeuvral patterns and prociency test scores and development for SIM groups and individual SIM group members. Group Man. pattern Member* Pre test score Post test score Development Group dev.** SIM 1 Turn-taking Left 19 19 0 Middle 15 16 1 Right 15 16 1 2 SIM 2 Turn-taking Left 7 14 7 Middle 5 12 7 Right 16 16 0 14 SIM 3 Dominance Left 12 15 3 Middle 13 18 5 Right 11 13 2 10 SIM 4 Turn-taking Left 7 7 0 Middle 18 15 -3 Right 14 15 1 -2 SIM 5 Turn-taking Left 8 15 7 Middle 14 16 2 Right 20 20 0 9 SIM 6 Dominance Left 16 17 1 Middle 10 12 2 Right 13 13 0 3 *Group members are described by their physical position in front of the simulator. **Group dev. is the combined development score for members of a given group. SIM, simulator. 456 L.O. Hll et al. 2011 Blackwell Publishing Ltd in SIM 5 gained so much (7 points), the data gives no clear-cut support for the opposite either, that mutual regulation of the verbal space is essential for collabora- tion to be effective with regard to prociency develop- ment. No signicant correlation between individual control over the verbal space and prociency develop- ment was found (Spearmans r -0.05; P > 0.05). We conclude that neither the verbal pattern nor individual control over the verbal interaction is directly related to prociency development. How is the group collaboration around the student- simulator interaction perceived by the students them- selves?The interviews showed that the SIMstudents are aware that some members spend more time operating the simulator and that some members have more inu- ence on the verbal interaction. Table 6 illustrates six cat- egories of factors put forward by students to explain these differences. (1) Personality: some people want to lead and others do not, and this determines a persons behaviour in the specic collaboration. (2) Social rela- tions: two members are more familiar with each other encouraging a dyad-ish interaction. (3) Competence: members subject competence varies and this decides how central they are to the collaboration. (4) Techno- logical: position in front of the computer determines how central members are to the collaboration. (5) Chance: members behaviour varies from day to day due to factors such as fatigue. Similar topics were raised when the students were asked to make suggestions for how to support group collaboration. We conclude Fig 4 Illustration of how often group members dominate the verbal space. Members of each group are labelled after their position in front of the computer: left, middle, or right. Collaboration and educational simulations 457 2011 Blackwell Publishing Ltd that the students are aware of the unequal distri- bution of control over the simulator and the verbal space. They posit four categories of explanations as to why this occurs, individual, social, technological, and contextual. Summary and discussion This article has contributed to the investigation of col- laboration as a technique for using ECAS in health care education. We have focused on relations between tools, collaboration and learning, and specically on two themes. First, causes and effects of the extent of applica- tion of subject-specic terminology. Second, patterns of collaboration around the student-simulator interac- tion and the effects of these patterns on prociency development. Observing SIM and CON groups side by side origi- nally made it obvious that students training conven- tionally made use of subject-specic terminology to a much greater extent than students working with the ECAS. In this article, we have argued that these terms are part of cultural tools associated with academic domains of anatomy and projection, and that these are more tightly associated with interpretive activity as con- trasted with the action-oriented activity that is required by simulation tasks. This suggests that the differences are, in part, explained by differences in what the stu- dents are tasked to achieve. We argue, however, that the differences in tasks do not explain all the difference in applications of subject-specic terminology and propose that the mediumof feedback may also be part of the explanation. While SIM students were given visual and numeric feedback, CON students were given textual feedback in subject-specic terms. If this hypothesis is true, that a certain characteristic of the tools, the medium of feedback, contribute to these effects, then SIM students would increase their applica- tion of subject-specic terminology if given textual feedback similar to that given in CON. Further research may yield more information about how different media of feedback after subtask completion impacts the col- laborative process and learning outcomes. We have, however, found no clear-cut relation between the extent of application of subject-specic terminology and prociency development within the SIM groups. It is possible that a more qualitative analysis, following the students problem-solving processes, is necessary to make useful associations between collaborative pat- terns of tool usage and prociency development. It is also possible that collaborating extensively with subject-specic terms has other benets than what is evaluated by prociency tests such as the one used in this study. When training with an ECAS, interaction with the simulator is essential, and depending on the type of simulator, the hands-on physical interaction may be more or less central to the collaboration. We studied how the physical student simulator-interaction affor- dances impact collaboration and prociency develop- ment. Observations of the SIM groups made it obvious howcontrol over the simulator is shared and not shared, Table 6. Categorization of respondent (SIM only) statements about differences in time spent manoeuvring the simulator and verbal space. Inferred category Example quote (a) Personality 1) I am one of those that are not afraid to speak my mind, and I can be completely wrong. I am very active in classes too, ask a lot, Im not someone to sit quiet and think. Me and X talk much more, but I see that as natural, thats just the way it is, she is much less talkative than me and X. 2) I dont want to be the one leading the discussion, or lead anything at all. Maybe I rather listen and then Ill speak if something gets strange. Better to let those lead who want to lead. (b) Social relations But also that I and X are pretty familiar with each other. We already knew each other well. (c) Competence 1) X is very good at these things, so one tends to trust him. 2) Im the one that in most cases takes the initiative in presentations, transcriptions etc. It is probably because if I do it, it is done correctly. (d) Technological I also think that you get to hold it more if you sit in the middle. (e) Contextual It also depends on what day it is, if you are tired etc. Some days you just want to listen. n = 9. SIM, simulator. 458 L.O. Hll et al. 2011 Blackwell Publishing Ltd but it is harder to knowif and howthis impacts learning. We have shown that there are two distinct collaborative patterns for sharing control over the simulator. One was characterized by turn-taking, albeit with different degrees of equality, and the other by one-member domi- nance. Both turn-taking and dominance patterns produce both better and worse group prociency devel- opment, suggesting that neither pattern is in itself better. In the two dominance groups, being the sole operator may have been slightly benecial, but in the turn-taking groups the impact on prociency development was varied. We also studied whether these two patterns produced specic patterns of verbal interaction. Observations revealed differences in how the verbal space is distrib- uted among group members. We found a dominance pattern, where one of the members has command over the verbal space almost all the time. We also found variations of a dyad-ish pattern where a somewhat dominant member shares more or less space with one of the other members while tending to exclude the third. There was, however, no clear-cut relationship between the patterns of control over the simulator and the pat- terns of control over the verbal space. Nor was there a clear-cut relationship, on the individual level, between controlling the simulator and controlling the verbal space. We also found no relationship between pro- ciency development and verbal pattern, or individual control over verbal space. Further research may give us deeper insights into if and how these collaborative pat- terns, in other respects, impact collaborative learning with ECAS in health care education. This is, however, an illustration of Sljs suggestion that technologies have implications for social activities, some of which may be foreseen and others which materializes as the technology becomes integrated into specic activities (2010, p. 54). The students themselves were aware of the differ- ences in the distribution of verbal space and control over the simulator. When trying to explain it and when sug- gesting how collaboration can be optimized, they name factors such as personality, social relationships, compe- tence, position in front of the computer and chance. The fact that students have ideas about whomthey think they would collaborate well with lead to ideas about student- regulated matchmaking and group creation mecha- nisms. According to Wessner and Pster (2001), [G]roup formation in CSCL environments is either done manually with little support fromthe system, or the system needs an elaborated model of the learning domain in order to select potential peer learners and to form learning groups in a pedagogically sound way. Because our students had studied together as a class for about 2 years, it would have been interesting to know how the collaborative training with this VR simulator would have been affected by such a simple thing as having students create their own groups. Would they in practice have teamed up with individuals that enhanced collaboration, or just made the training more social? Then again, maybe there is a professional value in stu- dents continuously learning to collaborate with different individuals irrespective of their personal differences? Either way, the social composition along of groups is one important aspect toconsider, alongwiththe collabo- rative activities and the technological support, when designing CSCLenvironments (Stahl 2004, p. 87) As with any scientic method, we set limitations for the generalizations possible to make from the empirical data, and some of them need to be mentioned. One of these limitations is that the comparisons made on group level are made with a relatively small number of units, six. This means that we need to be cautious in making statistical generalizations from these comparisons. They do, however, serve a purpose as indicators of the state of our simulation training. This is why we have given descriptive statistics a prominent role when dealing with relationships between group level patterns. The limitations in space have also required us to focus on some issues and leave others for future papers, such as further analysis of the interrelationships between factors considered in this text. It is possible for instance that there is a positive correlation between application of subject-specic terminology and prociency devel- opment but that this effect is countered by factors such as group interaction patterns. Introducing simulator logs of student activity, such as time on task and perfor- mance on tasks could also enrich our understanding of the matter. One nal thing must be noted about the value of col- laboration. The mean prociency development for the collaborative SIM groups in this study, a signicant two-point development, is identical to that found by Nilsson et al. (2007) for comparable students training individually with the same radiological VR simulator. That is, having students train individually and having them work collaboratively in groups of three, in the Collaboration and educational simulations 459 2011 Blackwell Publishing Ltd manner they did in this study, seem to make no differ- ence with regard to the groups mean prociency devel- opment. We may be able to enhance our support for collaboration by identifying benecial collaborative patterns and encouraging these during training as Dil- lenbourg et al. (2009) suggests, thereby supporting stu- dents prociency development. But, are we not equally likely to be able to enhance support for individual train- ing to make that more efcient? Is this a type of screen- based simulation training, and training with ECAS in general, activities where the answer to the how do we optimize training efciency question should generally be answered by collaborative activity? These questions are related to the questions of what we want to achieve with education and what part of this we can, and do, evaluate with student examinations. Some would say that collaboration has benets other than that of learn- ing to solve the task at hand, like developing critical thinking (e.g. Gokhale 1995), and that it is necessary to achieve certain learning goals (Kirschner et al. 2004). As long as ECAS continues to spread in health care edu- cation, be it because of a paradigm shift in educational models or because of other factors, research needs to help us better understand how to support students when training with them, irrespective of whether the training is done collaboratively or individually. Note 1 For topological discussions of ECAS in health care see examples for Gaba (2004) and Nehring and Lashley (2009). References Abrahamson S., Denson J. & Wolf R. (1969) Effectiveness of a simulator in training anesthesiology residents. Journal of Medical Education 44, 515519. Bradley P. (2006) The history of simulation in medical educa- tion and possible future directions. Medical Education 40, 254262. Chopra V., Gesink B.J., De Jong J., Bovill J.G., Spieerdijk J. & Brand R. (1994) Does training on an anaesthesia simulator lead to improvement in performance? British Journal of Anaesthesia 73, 293297. Crook C. (1994) Computers and the Collaborative Experi- ence of Learning. Routledge, London. Cuban L. (2001) Oversold and Underused: Computers in the Classroom. Harvard University Press, Cambridge, MA. Dillenbourg P., Jrvel S. &Fischer F. (2009) The evolution of research on computer-supported collaborative learning. In Technology Enhanced Learning (eds N. Balacheff, S. Ludvigsen, T. de Jong, A. Lazonder &S. Barnes), pp. 319. Springer, Berlin. Engum S.A., Jeffries P. & Fisher L. (2003) Intravenous cath- eter training system: computer-based education versus traditional learning methods. The American Journal of Surgery 186, 6774. Enyedy N. (2003) Knowledge construction and collective practice: at the intersection of learning, talk, and social con- gurations in a computer-mediated mathematics class- room. Journal of the Learning Sciences 12, 361407. Gaba D.M. (2004) The future vision of simulation in health care. Quality and Safety in Health Care 13 (Suppl. 1), 210. Gokhale A.A. (1995) Collaborative learning enhances critical thinking. Journal of Technology Education 7, 2230. Hindmarsh J. (2010) Peripherality, participation and commu- nities of practice: examining the patient in dental training. In Organisation, Interaction and Practice (eds N. Llewel- lyn & J. Hindmarsh), pp. 218240. Cambridge University Press, Cambridge. Issenberg S.B., McGaghie W.C., Petrusa E.R., Lee Gordon D. & Scalese R.J. (2005) Features and uses of high-delity medical simulations that lead to effective learning: a BEME systematic review. Medical Teacher 27, 1028. Kirschner P.A., Martens R.L. & Strijbos J.W. (2004) CSCLin higher education? A framework for designing multiple collaborative environments. In What We Knowabout CSCL and Implementing It in Higher Education (eds J.W. Strijbos, P.A. Kirschner &R.L. Martens), pp. 330. Kluwer Academic Publishers, Boston. Koschmann T., ed. (1996) CSCL: Theory and Practice of An Emerging Paradigm. Lawrence Erlbaum, Hillsdale. Koschmann T., LeBaron C., Goodwin C. & Feltovich P. (2010) Can you see the cystic artery yet? Asimple matter of trust. Journal of Pragmatics 43, 521541. Krippendorff K. (2004) Content Analysis. An Introduction to Its Methodology. Sage, Thousand Oaks, CA. Kvale S. & Brinkmann S. (2009) Den kvalitativa forskn- ingsintervjun. [The Qualitative Research Interview]. Stu- dentlitteratur, Lund, Sweden. Lazonder A.W., Hagemans M.G. &de Jong T. (2010) Offering and discovering domain information in simulation-based inquiry learning. Learning and Instruction 20, 511520. Luengo V., Aboulaa A., Blavier A., Shorten G., Vadcard L. & Zottmann J. (2009) Novel technology for learning in medi- cine. In Technology Enhanced Learning (eds N. Balacheff, S. Ludvigsen, T. de Jong, A. Lazonder & S. Barnes), pp. 105120. Springer, Berlin. McGaghie W.C., Issenberg S.B., Petrusa E.R. & Scalese R.J. (2010) A critical review of simulation-based medical edu- cation research: 20032009. Medical Education 44, 5063. 460 L.O. Hll et al. 2011 Blackwell Publishing Ltd Mercer N. (2005) Sociocultural discourse analysis: analysing classroom talk as a social mode of thinking. Journal of Applied Linguistics 1, 137168. Nehring W.M. & Lashley F.R. (2009) Nursing simulation: a review of the past 40 years. Simulation & Gaming 40, 528 551. Nilsson T.A. (2007) Simulation supported training in oral radiology. Methods and impact in interpretative skill. (Doc- toral dissertation, Ume University). Nilsson T.A., Hedman L.R. &Ahlqvist J.B. (2007) Arandom- ized trial of simulation-based versus conventional training of dental student skill at interpreting spatial information in radiographs. Simulation in Healthcare 2, 164169. OMalley C. (1995) Computer Supported Collaborative Learning. Springer, Berlin. Ovasapian A., Yelich S.J., Dykes M.H. & Golman M.E. (1988) Learning beroptic intubation: use of simulators v. traditional teaching. British Journal of Anaesthesia 61, 217220. Peugnet F., Dubois P. & Rouland J.F. (1998) Virtual reality versus conventional training in retinal photocoagulation: a rst clinical assessment. Computer Aided Surgery 3, 2026. Pilkington R.M. (1998) Dialogue games in support of qualita- tive reasoning. Journal of Computer Assisted Learning 14, 308320. Pilkington R.M. & Parker-Jones C.H. (1996) Interacting with computer-based simulation: the role of dialogue. Comput- ers and Education 27, 114. Quinn F., Keogh P., McDonald A. &Hussey D. (2003) Astudy comparing the effectiveness of conventional training and virtual reality simulation in the skills acquisition of junior dental students. European Journal of Dental Education 7, 164169. Rystedt H. & Lindwall O. (2004) The interactive construction of learning foci in simulation-based learning environments: a case study of an anaesthesia course. PsychNology Journal 2, 165188. Slj R. (1999) Learning as the use of tools: a sociocultural perspective on the human-technology link. In Learning with Computers (eds K. Littleton & P. Light), pp. 144116. Routledge, London. Slj R. (2000) Lrande i praktiken: ett sociokulturellt pers- pektiv [Learning in Practice: ASociocultural Perspective]. Prisma, Stockholm. Slj R. (2010) Digital tools and challenges to institutional traditions of learning: technologies, social memory and the performative nature of learning. Journal of Computer Assisted Learning 26, 5364. Sderstrm T., Hll L.-O., Nilsson T. & Ahlqvist J. (2008) How does computer based simulator training impact on group interaction and prociency development. Proceed- ings of the International Conference of Information Com- munication Technologies in Education Corfu, Greece, 10-12 July, 2008. Stahl G. (2004) Group cognition in computer-assisted col- laborative learning. Journal of Computer Assisted Learning 21, 7990. Wertsch J.V. (1998) Mind As Action. Oxford University Press, NewYork. Wessner M. &Pster H. (2001) Group formation in computer- supported collaborative learning. In Proceedings of the 2001 International ACM SIGGROUP Conference on Sup- porting Group Work (eds S. Ellis, T. Rodden & I. Zigurs), Sep 30Oct 3, 2001, Boulder CO, pp. 2431. ACM, New York. Collaboration and educational simulations 461 2011 Blackwell Publishing Ltd Copyright of Journal of Computer Assisted Learning is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.