May 3, 2004
Class: CEP 917 Design of Media for Learning
Instructor: Punya Mishra, PhD
College of Education, Michigan State University, Lansing, Michigan, USA
Collaborative design is complex, as designers interact with other designers and with their ideas. Through these interactions, ideas are suggested, adopted, or rejected. This process is similar to an ecosystem, in which organisms are spawned, nurtured, and killed. Careful analysis of a design group's interactions can suggest a map of these phenomena. By mapping a design group's ecology of ideas, researchers may gain deeper insight into the unique combination of roles and processes underlying the moment-to-moment interactions. Furthermore, a group may benefit from studying a map of their own ecology.
In the novel Foundation, author Isaac Asimov suggests that psychology and mathematics may eventually lead to powerful tools for predicting human interaction. (1951) By quantifying and mapping human behavior, Asimov's "psychohistorians" effectively reduce prophecy to algorithms. While such science fiction may be implausible, it illustrates a possible direction for scientific study. By carefully observing and mapping human interaction, even in small groups, researchers may find clearer ways to describe and understand important roles and processes.
Understanding group dynamics is one of the most important, enduring challenges in society and education. How can we create and maintain effective classrooms, military units, corporate divisions, creative design groups, etc.? Many educators tout group work, in "brickspace" or "clickspace." Design studios and similar models suggest advantages over traditional direct instruction. But in or out of education, offline or on, collaboration and other group dynamics are complex and poorly understood. For example, in their excellent exploration of collaboration and problem solving, Hennessy and Murphy reiterate that collaboration requires more than "simply placing pupils in groups to work or asking them to talk." (1999)
Ecology may be a useful metaphor for studying groups, especially collaboration in groups. Within any group of two or more people, ideas are generally introduced, elaborated on, criticized, adopted, or rejected. This is only a general characterization, because groups and their interactions are richly-detailed, complex phenomena. However, the fundamental dynamic is organic: ideas are spawned, nurtured, and killed, just like organisms in an environment. By combining traditional methods with this unusual metaphor, researchers may make the familiar strange, seeing commonplace phenomena in new ways. For example, by extending the metaphor to include parenting and predation, researchers may be inspired to look for major idea producers and major idea rejecters. The most powerful individual may be the most creative, offering many suggestions, or power may reside with the most prolific idea-killer. Hence, it should be valuable to understand idea ecologies in general and for specific groups, to create and maintain desirable, effective groups.
All groups have unique, often-changing ecologies. After working with a particular group over time, an individual can usually speculate about how the group or specific individuals will respond to certain behaviors and ideas. In other words, as social creatures, humans excel at developing primitive, intuitive maps of their groups' ecologies. Like any primitive map, such intuitive understanding has limited utility. More sophisticated understanding is necessary to form or reform groups to meet certain goals. Formally mapping a group's ecology is one potentially powerful tool (among many), particularly if it fosters new ways of discussing the history and goals of a group and its members.
Barab et al. used a kind of mapping in their study of collaboration in a technology-rich summer camp. (2001) They focused on the diffusion of practices, while this study focuses on ideas. In either case, identifying "initiators" and mapping interactions offer clues to a group's roles and relationships. Similarly, Rowell is inspired by Vygotsky to view collaboration as social negotiation of meaning and solutions. She suggests that effective collaboration is fostered when participants assume "complementary roles" (e.g., manager, assistant). (2002) Vygotsky himself argued for the causal power of roles in learning and development. (1978, p. 95) By playing a role like "spawner" or "nurturer," an individual contributes to a collaborative task while growing and learning as an individual. Not surprisingly, collaboration is central to many potentially-powerful teaching strategies, including reciprocal teaching, and cognitive tools and intellectual roles. (Palincsar and Herrenkohl, 2002)
In the ecology metaphor, the figurative violence of "killing" is intended to convey the sometimes-brutal dynamics of creativity and collaboration. Game designer Chris Crawford suggests that, "Sometimes the most brilliant design stroke is to kill the idea." (2003, p. 242) Five great ideas may combine in disaster, while eliminating two may produce an enjoyable game. An individual can feel wrenching disappointment when a group kills his idea, but the crucible of collaboration also offers sweet satisfaction when an idea is accepted and nurtured. Feeling a sense of ownership of an idea and the larger process is key. Kafai studied how a class that used game design for learning, and he emphasized the value of personal investment in collaboration: "learning is most effective when children build personal, meaningful objects." (1995, p. 11) The strong connotations of "spawn, nurture, and kill" suggest that group members experience vivid interaction with each other and their ideas. This vividness is the authenticity at the heart of constructivist pedagogies. The aim is creating "an environment in which learners use their minds and bodies as they would if they were practitioners in a domain." (Honebein, 1996) The violent dynamics of collaboration should engage participants and let them learn by playing roles in an authentic fashion.
What follows is a primitive ecological analysis of a creative design group. It illustrates the utility of the ecology metaphor and suggests ways of mapping a specific group's ecology. The primary purpose of the analysis is this illustration. Researchers performing a case study or other more directed investigations would benefit from combining this analysis with more traditional methods (e.g., correlational analysis of coded transcripts). By itself, the mapping analysis has several expected results:
The participants were five 8th grade boys and an adult teacher-facilitator (TF). The group named themselves the Neptune Jihadis, and selected pseudonyms: Hydrus, Telescopium, Lynx, MoNOSseros, and Sagitta (H, T, L, M, and S). They met as a creative design group for ten weekdays, half a day each day. They were one of eight similar gender- and grade-level-homogenous groups created from a pool of 5th and 8th grade girls and boys. The Neptune Jihadis interacted with the other boys frequently, but never with the girls.
All groups were focused on designing educational computer games about space exploration, as part of a two-week, space-themed summer camp at a major midwestern university. The majority of the first week was spent on educational/enrichment activities (e.g., field trips), while the second week was mostly dedicated to group-based collaborative design work.
The Neptune Jihadis were selected for this analysis from the two 8th grade boys' groups. It was hoped that older boys would be more familiar with computer games and hence more cogent in their discussions (compared to girls or younger boys).
On the last day of the camp, each group presented their final game design to the other same-gender groups. The Jihadis' PowerPoint slideshow is used in the analysis as a unique indicator of which ideas were most successful at surviving.
Throughout the camp, each group was studied by two researcher-observers, who took notes about the participants' verbal and nonverbal interactions. The Jihadis' researcher-observers were ET and TM. Their notes were later entered and coded using NVivo software, by two other researchers. The coders achieved about 80% reliability. The notes on the Jihadis were coded by a single researcher (who was not ET or TM).
The Jihadis' ecology was mapped based on a node the coders identified: Design process: How. This node was linked to any instance in the notes of verbal or nonverbal interaction that seemed to include the introduction, consideration, elaboration, synthesis, or rejection of design ideas. Of the many nodes used in the larger space camp project, this node seemed most likely to track the spawning, nurturing, or killing of ideas.
Each interaction was classified as (S)pawn, (N)urture, or (K)ill, after careful study. The coded interactions were studied individually, in the context of the larger transcript. This was detective work: studying both observers’ notes, working backwards and forwards in time, judging idea similarity, etc. The resulting data is a far-from-perfect set of points illustrating the group's idea ecology. Each point includes the observer, day, participant, the participant's action, and the idea acted upon. 5 of the 82 points were found in both observers' notes. 58 of the 82 points were found in ET's notes. Here are some examples of how actions were [categorized]:
MoNOSeros: You should start off on earth I think, and then stop off at the moon for refueling and maybe then play sports for a little while; [Spawn: stages, Spawn: sports]
MoNOSeros: And I think food fights should be a part of the trip as well; [Spawn: food fights] (TM, day 4, line 72-73)Telescopium shouts, “No man, No sports! Get that into your head!” [Kill: sports] (TM, day 10, line 76)
The data are imperfect because the notes are a poor record of the richness of the interactions. For example, sarcasm or rhetorical questions may not have been noted as such by the observer. Does talking about an idea count as nurturing it? (This particular analysis assumes so.) The categories of S, N, and K only roughly capture the nuances of specific interactions. Some details were indiscernible, e.g., "Telescopium begins to provide a detailed explanation of the game-idea he has; Others listen quietly" (TM, day 6, line 18), or "HYDRUS: I have a question. Do you all like this design?" (ET, day 1, line 13). If a coded interaction wasn't a clear S/N/K action about a specific idea, it wasn't included in this analysis.
The greatest challenge was identifying and tracking specific design ideas. Greater granularity would honor the richness of each interaction. Yet if the ideas are categorized as too granular, it's difficult to find patterns of nurturing or killing. This analysis tries to strike a balance. But seemingly pervasive ideas (e.g., mars tycoon) could have been broken down into disparate elements.
The data were collated in a computer spreadsheet, and a visual map was created in a computer drawing program, by hand. Several descriptive statistics were also tabulated. No correlations were calculated because of the imperfections of the data and the purely descriptive nature of this analysis. Finally, the Jihadis' PowerPoint slideshow was examined. The slideshow was the group's culminating product, so its ideas represent "the survival of the fittest."
Due its size, the map is a separate file, available here. Recall that the data points are imperfect representations of rich, sometimes-ambiguous interactions. Don't read too much into the proximity of actions. Some proximal data points are separated by dozens of lines (or minutes) in the research-observers' notes. Researcher-observer TM didn't describe the "(unspecified)" idea, but TM's notes show a clear spawn-kill sequence so (unspecified) is included in the map and subsequent analysis.
These summary statistics offer a variety of different ways to cleave the Jihadis' ecology map. Tables 1a, 1b, and 1c all analyze the same data in slightly different ways: the roles and power of each individual.
Table 1a: Analysis by Person, Counts.
Table 1a shows a breakdown of each person's actions in the group. H appears to be the most active member (22), followed closely by T (19) and M (18). L and S appear to be the least active (6 and 4). This fits with a casual reading of the transcripts, in which H seems to quickly dominate and lead the group. Note that L and TF (the teacher-facilitator) never killed an idea. H, M, and T are major producers (9, 9, and 7), while TF, H, M, and T are major nurturers (10, 8, 7, and 7).
Table 1b: Analysis by Person, Percents.
Table 1b shows the same data, by percent (e.g., H's %Spawns = H's Spawns / H's TotalActions). Given the small number of data points, it would be premature to read much into the balance of %Spawns, %Nurtures, and %Kills. However, for a larger data set, this balance could be more indicative of a specific group's fundamental dynamics. Appropriate questions would include: How many of a group's actions are constructive versus destructive? How does an individual's balance compare to the groups? Does the group's productivity depend on a few hyperactive individuals?
Table 1c: Analysis of Contributions to the Group, Percents.
Table 1c continues this line of analysis. It reveals what percentage of the group's spawns, nurtures, and kills are attributed to a specific individual. This reinforces a sense of the relative lack of influence L and S had in the group.
Table 2 suggests a different direction of analysis, looking at the ideas themselves. Each idea is further labeled by whether it appears in the PowerPoint slideshow. Finally, since H seemed to have a particularly strong influence on the group, his relevant actions are included.
Table 2: Analysis by Ideas, Counts.
Table 2 illustrates the wide variety of ideas the group explored. Note the large number of orphaned ideas, those with only a single action (i.e., S) that don't appear in PowerPoint. Orphaned ideas include: food fights, pee machine, matter to food, mission: entertainment, mission: escape earth, SimCity Mars, pop music, and futuristic ship. When using this kind of table to reform a group, an appropriate question might be: Given the amount of time spent by the group, is the number of orphaned ideas too large, too small, or just right? Considering this question may be more valuable than the answer; there isn't necessarily a "correct" number of orphaned ideas to strive for.
Note that H spawned and had a strong nurturing influence on the success of the most-acted-on idea, mars tycoon. With a larger or less ambiguous set of data points, it might be appropriate to break down the other individuals' influences, and test possible correlations between a specific individual's involvement and the success of the idea. The observers' notes describe how H almost exclusively controlled the creation of the PowerPoint slideshow, so his actions are particularly relevant when reviewing the presence or absence of ideas therein. Note that H originally killed racing (on day 7), then nurtured it during the creation of the PowerPoint slideshow (on day 10), at which point T killed it, but racing survived as a bullet point in the slideshow.
Conflict can be a healthy part of collaboration. Table 3 tries to explicate specific relationships between individuals: who killed who's ideas?
Table 3. Spawners and Their Killers.
Note the relatively strong interaction between H, M, and T. They killed each other's ideas almost equally. Table 1a shows that S spawned only 1 idea, and Table 3 shows that M killed it. Note that the total kills for H and S are less here than in Table 1a. This is because on day 10, H and S killed an idea spontaneously; the idea didn't appear to have been spawned by anyone, but the kills were clear.
The map clearly shows more actions on some days than others. Table 4a breaks this down numerically.
Table 4a: Analysis by Day.
No S/N/K actions were observed on day 1, 2, 3, or 5, at least as related to the game design task. The Jihadis did collaborate on some minor tasks, including their group name and logo. However, these tasks had much shorter durations than the overall game design, and were excluded from this analysis. No doubt the Jihadis also discussed their design ideas outside the researcher-observers' perceptions (e.g., in the hallways), so all this only reiterates the crudeness and ambiguity of the data points and subsequent analysis. Nevertheless, in reviewing Table 4a, the camp organizers might choose to revisit the agendas for days 7 and 8, which were ostensibly collaboration-focused days, and consider why other days may have fostered more observable S/N/K actions.
Table 4b analyzes daily activity as well, broken down by individual.
Table 4b: Analysis by Day and by Individual.
Note that H and M became active early, perhaps foreshadowing their eventual strong influence on the group. The teacher-facilitator (TF) was particularly active on day 6, the first day of the second week of camp, when the group began to focus almost exclusively on the design work. The TF remained active but less influential on the remaining days.
As previously mentioned, most of the data points came from one of the two researcher-observers (RO's), ET. 5 data points were found in both RO's notes. Table 5 shows a further breakdown of who observed what.
Table 5: Analysis by Research Observer.
Given the imbalance of uncorroborated observations (58, 19), it's difficult to make a distinction between the RO's observation patterns. However, in general, this kind of analysis could reveal systematic bias (e.g., perhaps one RO is more likely to observe spawns than kills). This is a critical issue for this kind of analysis. The map of a group's ecology of ideas should illustrate their unique collaborative processes, not the filters of the observer(s).
This analysis focuses exclusively on ideas related to the game design. This simplified the map of the ecology, but may exclude many critical details. Relatively inactive individuals may be contributing the group in other ways, or their inactivity may be due to non-design tensions or animosity. Any map will only be a representation of its group, and hence an imperfect picture.
The low number of data points makes it difficult to draw sweeping conclusions about the Neptune Jihadis. In some cases just a few data points are attributed to an individual (e.g., apparently, S only spawned 1 idea). This is probably indicative of the necessarily incomplete records the researcher-observers created. It's impossible to record all the details and nuances of a group, especially over long periods of time. Video recording would provide a much richer record, but it would still be incomplete. However, while the data are ambiguous, they often may indicate group dysfunction. Perhaps S's single idea was killed so violently that S was reluctant to contribute again. Interviews with the individuals would help corroborate such interpretations, and suggest possible mitigations if the goal is reforming the group.
The teacher-facilitator (TF) had an unexpectedly significant influence on the ecology of ideas. As Table 1a shows, the TF was more active than two of the students. Many of the TF's actions were nurturing by asking questions. This is the traditional role of coach in a design group, and such nurturing is not necessarily a bad thing. However, the content and phrasing of questions can nurture otherwise-dying ideas by reminding the group of what was said earlier. To authentically capture the ecology of just 8th grade boys, the TF would have to be far less active. In Barab et al.'s similar study, the authors express similar concerns about the influence of TF's. (2001) Further, Hennessy and Murphy suggest students must have "a controlled degree of autonomy in decision making" as "an essential prerequisite for supporting learning through collaboration." (1999)
This analysis was started with a naive expectation of blatant actions that would be easy to categorize. Instead, the complexity in situ of the group made categorization a compromise between completeness and accuracy. For example, a kill action doesn't always kill an idea. Hence, all kills could be separated into permanent or ineffectual, but the efficacy of a kill is already apparent in the map. Another example: H kills then nurtures racing. Should the first kill be counted as a kill? (It was.) H and S spontaneously killed techno, which apparently hadn't been spawned. Should these be counted as kills? (They were.) Should ideas spawned during a brainstorming activity be categorized just like those spawned during enrichment or other activities? (They were.) Synthesis is particularly challenging to categorize: when does an action build on an idea, and when does it replace it? There's nothing wrong with these phenomena; they are a natural part of the ecology of a creative design group. Rather, when mapping a group's ecology, researchers should be clear about their assumptions and compromises, and weigh the significance of the map's patterns against these issues.
This analysis was started with an expectation that 8th grade boys would be better at talking through game design. However, a casual reading of other groups' transcripts counters this argument. While the older boys may have more experience with games, they may be less fluent in design discussion in general. The Jihadis seldom seemed to offer justifications for their nurture or kill actions, so it's difficult to get a sense of their thinking. Casual reading of the notes about other groups reveals contrasting behavior. For example, compared to the boys, many of the girls seemed more expressive about their reasons for nurturing or killing ideas, thus offering a potentially richer picture of their ecology. As another contrast, some of the individuals in the other 8th grade boys group were obsessed with a single commercial game, and that model clearly drove their group's design process. Any other design was effectively over-shadowed by the model. A logical next step would be mapping some of these other groups and comparing the maps.
Most of the kills happened in the last few days. The individuals may have been more comfortable with each other by that point, or perhaps bolder since the camp was about to end and they wouldn't see each other as much (if at all). The pending presentation and need to finalize the design may also have fostered more vibrant interactions.
A group's ecology is partially determined by its individuals. It's also determined by the activity or product. A single game design is a convergent product, so some killing of ideas is expected, perhaps even more than observed in the Jihadis. A convergent activity would seem to benefit from the unifying vision of a leader. In other words, the nature of the task at hand may foster a certain group ecology. (cf. Rowell, 2002) So it's not surprising that H consciously assumes a leadership role, as seen in this interaction with the teacher-facilitator, Sam:
Hydrus arrives nearly a half hour early to camp after talking to Sam, yesterday after the day ended.
Apparently, the chat consisted of him asking Sam whether he was “in charge” of the group. This is consistent with my observations yesterday. For the latter half of the day, I sensed that he knew we were picking-up on his leadership and role in the group. Today, from the very beginning his demeanor implied a leadership/dominant role (TM, day 10, lines 18-20)
The map was created by hand. This kind of analysis would be easier if the mapping process was automated and streamlined. This analysis was performed by an expert on computer games, which helped greatly in distinguishing design ideas. Domain-specific expertise would invaluable in mapping a group's ecology of ideas in a different context (e.g., architecture). The Jihadis were amateur game designers and at times may not have understood what they were talking about. For example, the mars tycoon idea was popular and successful, but it's not clear they knew what a tycoon is. (To wit: their game design seems more focused on saving humanity than making money, admirably.)
The ecology metaphor evokes one further, critical idea. Just as ecosystems change over time, groups change. For example, individuals may gain more confidence or inspiration in and outside the group, and become more active producers. At best, the Jihadis' map is an imperfect snapshot of two weeks. Were the camp repeated the following summer, a wholly different ecology and map might emerge. A long-term group might benefit from periodically reviewing their ecology with a map, and a collection of maps would be the equivalent time-lapse photography of the group's growth and development.
Mapping a group's ecology is one potentially powerful tool (among many) for understanding and reforming the group, as well as understanding groups in general. Such maps would be particularly useful if they foster new ways of discussing the history and goals of a group and its members. Perhaps such an analysis early in a group process would help the group mature over time, in the direction they and their clients or supervisors desired.
Asimov, I. (1951). Foundation. New York: Gnome Press.
Barab, S. A., K. E. Hay, et al. (2001). "Constructing Virtual Worlds: Tracing the Historical Development of Learner Practices." Cognition and Instruction 19(1): 47-94.
Crawford, C. (2003). Chris Crawford on Game Design. Indianapolis, New Riders.
Hennessy, S. and P. Murphy (1999). "The Potential for Collaborative Problem Solving in Design and Technology." International Journal of Technology and Design Education 9: 1-36.
Honebein, P. C. (1996). Seven Goals for the Design of Constructivist Learning Environments. Constructivist Learning Environments: Case Studies in Instructional Design. B. G. Wilson (Ed.). Edgewood Cliffs, Educational Technology Publications: 11-24.
Kafai, Y. B. (1995). Learning Through Design: Review of Theory and Research Issues. Minds in Play. Hillsdale, Lawrence Erlbaum Associates: 1-29.
Palincsar, A. S. and L. R. Herrenkohl (2002). "Designing Collaborative Learning Contexts." Theory into Practice 41(1): 26-32.
Rowell, P. M. (2002). "Peer Interactions in Shared Technological Activity: A Study of Participation." International Journal of Technology and Design Education 12: 1-22.
Vygotsky, L. S. and M. Cole, John-Steiner, Vera, Scribner, Sylvia, and Souberman, Ellen (Eds.) (1978). Mind in society : the development of higher psychological processes. Cambridge, Harvard University Press.
Created by Kym Buchanan | http://KymBuchanan.org | This work is licensed under a Creative Commons License.
