Too Much Headspace

[Note: This was a proposal for a research study written in December 2000. The study ultimately did not receive funding; I post the proposal here for its insight into SCA decision making processes]

Introduction:
It is obvious that groups make decisions, but much debate has been raised over the years as to how they go about this process. Researchers have proposed a number of models to describe this process, and huge variation exists even within models. Among the more recent developments have been attempts to emulate decision making models using artificial intelligences (AI). Brick and Swinth (1980) developed an AI which emulated decision making in a multidisciplinary health care agency with exceptional levels of accuracy, while Stasser (2000) has used the DISCUSS computer modeling program to identify new areas of research.

This intense focus on laboratory testing of theories, however, ignores the fact that most decisions are made by groups in naturalistic settings, operating under uncontrolled, and many times uncontrollable, conditions. Take, for example, a decision making team in a corporation. It is entirely possible at any point that a member of the team will leave due to sickness, transfers, firing, or promotions, just to name a few reasons, and be replaced by a new member. Or perhaps the team will find out a month into the project that their budget has been cut in half. This is an unexpected pressure which will have a serious impact on how they continue with their decision. Studying decision making theories in real-world settings, therefore, can help us determine whether or not they will hold true under such adverse, uncontrolled conditions.

Bearing this in mind, the primary research question this study will address is how Phasic, Critical Event, and Social Construction theories of group decision making apply to a pre-existing group and how they operate in the context of an actual decision. The group chosen was the barony of Würmwald, a local branch of the Society for Creative Anachronisms, a group of medieval re-enactors. They will be observed as they plan for and prepare an annual event.

Literature Review:
Phase Models
While there is a great deal of disagreement about the specifics, all phase models break the decision-making process into discrete stages, or phases. Poole and Baldwin (1996) define a phase as “a coherent activity that serves some decision-related function.” Most phase models divide the process into three or four steps and move in a linear fashion, that is movement only goes forward and earlier phases cannot be revisited. Fisher’s (1970) phase model is one of the most frequently cited. It divides decision making into four phases. First is orientation. This is the time when group members become acquainted with each other. The basic problem is identified and members begin to express their attitudes. The next stage is conflict. It is at this point that controversy begins to arise. Since most of the SCA members have a pre-existing history with one another, this will likely be the first phase of interest. Following conflict is the emergence phase, wherein groups start to coalesce and dissent decreases. Davis (1996) found that factions will begin to emerge at this phase. In the SCA, for example, different factions might place different amounts of importance on combat, dancing, or storytelling. While there may be a general agreement that all should be present, it can be expected during the conflict phase that there will be a great deal of argument as to which of these activities deserves the most time and space dedicated to it. In the emergence phase, members will identify people with similar beliefs and form factions based around these arguments, with the result that there will be more ambiguity due to a three-way split, but less disagreement overall. The last phase, according to Fisher, is reinforcement. At this point the group has generally agreed on one course of action and they start developing unity and agreement through positive reinforcement of the plan. Bell (1982) also agrees with the phase model, but her research provides more evidence to the phasic model proposed by Bales and Strodtbeck, which emphases movement from orientation to evaluation to control.

Poole (1983 a, b, & 1981) feels these models are not quite complete, however. His work found that groups will occasionally need to revisit earlier phases or go back and forth between steps. This led him to support the idea of multiple sequence models, which allow groups to revisit earlier phases and seem to account for more of the complexity found in group decisions. Burnstein and Berbaum (1983) mention three kinds of cycles which can cause groups to revisit earlier phases: exogenous, endogenous, and cognitive cycles. Exogenous cycles represent those forces which are external to the group. For example; if the site the Barony chose for their event suddenly became unavailable, they would need to move to earlier stages of the planning process and adjust everything to fit the new location. Endogenous cycles are caused by forces within the group. This would be seen when discussion of a late stage generates ideas that would improve upon earlier stages, requiring the group to revisit that stage. The final type of cycle, cognitive cycles, represent the group’s own foresight. If a member sees potential negative consequences of a proposed plan, this may be justification to revise the plan and revisit an earlier phase.
From this, the following hypotheses were derived:

H1: Planning for the event will occur in a multi-phasic model.
The members of the barony will be subject to a large number of exogenous, endogenous, and cognitive forces as they plan, which will force them to revisit previous phases many times. In addition to this, many aspects of the event will be handled by sub-committees. As a result, the group as a whole will have to go through the phasic process many times as each subgroup goes through it individually.

H2: Orientation will be very brief or non-existent due to the group’s prior history.
Since this is a pre-existing group, most of the members already know each other, have relationships with each other, and are familiar with the rules and structures of the SCA. Therefore, very little time will need to be spent introducing members and concepts.

Critical Event Models
Critical event models work on the idea that a group’s workflow is punctuated by certain key events. The most frequently mentioned critical event model was developed by Gersick (1988, 1989) and stated that the chronological midpoint of a project serves as a natural alarm clock for the group. Her model has three points of transition, consisting of the first meeting, the midpoint, and completion of the task. There are two phases, representing the time before the midpoint and the time after. According to her model, a group gets very little work done during phase one, reaching full productivity only after half their time has past. The Barony’s event can be examined using this model as there is a definite starting point (when the autocrat, or person in charge of the event, is chosen) and a definite deadline (the date when the event is held).
Gersick predicts that planning is the primary activity in phase one, raising the question of how valuable pure planning can be. Stout et al (1999) did a study in which the most successful groups were those which used planning to build a shared mental model of what the project’s outcome should be. Planning allows the group to “set goals, create an open environment, share information related to task requirements…and clarify each team member’s roles and responsibilities.” Beyond that, increased levels of planning at the early stages of development corresponded to improved levels of group performance during high workload situations down the line. An added benefit of planning is an increased awareness of which members possess what knowledge. Propp (1997) found that information acceptance was greatly affected by three variables: pre-discussion preference, information redundancy among members, and the valence of information items. Of these, redundancy of information was the strongest predictor, but almost no information was accepted if it failed to meet a minimum valence level.

VanLear and Mabry (1999) confirmed this. Their study also found that redundant information-sharing interactions increased the likelihood a group would be able to reach consensus. Spending time in phase one building shared mental models increases the amount of redundant information which can be shared, increasing the group’s ability to make a well-informed decision.
From these studies, it would seem planning has value in and of itself. If Gersick’s model holds true for the Baronial event, it is quite likely this planning behavior will be all that is observed during phase one, but the work of VanLear, Mabry, and Stout et al leads one to believe this will be mostly beneficial.

In addition, VanLear and Mabry’s previously mentioned study focused specifically on discriminating between consensual and dissentient groups. They found in this study that just the presence of a strongly disagreeing faction could prevent the group from reaching consensus. Since the Würmwald barony makes all decisions by consensus, this could be a strong argument that the critical events model is a better fit. Most of the phasic models have at least one phase characterized by high levels of argument and disagreement. If this prevents the group from reaching consensus, the decision making process will be severely hampered. It will be of particular interest to see if a conflict phase does, in fact, emerge, and if so, how it will impact the group’s attempts to reach consensus.

The following hypotheses were created regarding critical events theories:

H3: Planning behaviors will be the primary activity during phase one.
Earlier studies look just at planning and execution behaviors. This study, however, proposes a third class of activities: preparation. Many essential aspects of the event, such as cooking, setting up booths, and so on, will be impossible to complete (execute) until immediately before the event. Preparation, therefore, exists as a distinctly separate stage between planning these delayed activities and executing them. To give an example: planning the meal for the event would involve selecting a menu; preparation would consist of buying the ingredients and supplies; finally, execution would be actually cooking the meal. Combining this idea with Gersick’s work leads to the third hypothesis as well as the next two.

H4: The transition point for the project will occur substantially after the project’s chronological midpoint.
H5: Sub-groups will show a marked increase in execution immediately before the event.
The first reason for this anticipated delay is the necessary delay in execution for most of the project. A large number of tasks will be completed at the last minute, and the temporal and social pressures of school (a large number of group members are students at the University of Illinois, Urbana-Champaign) will also serve to delay the completion date. Therefore, while phase two activities will occur after the temporal midpoint, it will be logical to extend the endpoint of phase one until a later date. A delay of the transition point does not necessarily negate Gersick’s model; it just modifies it. If, however, the transition point does not occur, it could represent a serious problem with the model.

Social Construction Models
Social construction theories “account for how a decision is constituted in and through interaction, how the contingency factors are mediated by member reactions, and how interaction advances the group toward a decision” (Poole & Baldwin 1996). Social construction models draw heavily from the theories of both structuration and symbolic convergence, which, in turn, relates to parts of both multiphasic and critical event models. Structuration theory is based on the idea that a group’s process determines its output, and a group’s output influences the process (Poole, Seibold, & McPhee 1996). This relates back to the multiphase model’s tendency to revisit earlier phases because of work done in later phases. More specifically, it ties into Burnstein and Berbaum’s exogenous, endogenous, and cognitive cycles. Symbolic convergence, on the other hand, represents the group’s efforts to create a shared “fantasy” (Bormann 1996) of what the final output should be. This leads to the creation of shared mental models. While social construction models seem to be advantageous in that they are more inclusive, the theories supporting these models suffer from a lack of parsimony when compared to the other models.
The major interest in social construction theories for this project stems from the pre-existing relationships between many members of the group. They socialize with each other outside of formal group meetings, therefore it will be of interest to see how “informal” discussion of tasks impacts the “formal” meetings. Since there are more opportunities to meet outside of the formal meetings, the final hypothesis is proposed:

H6: Most of the planning will occur outside of formal meetings.


Method:
Participants:
The Society for Creative Anachronisms was originally founded on May 1, 1966 as a science fiction fan group; the Society was seen mainly at science fiction conventions in the Berkley area for years. However, it quickly gained a national following and has since become an international not-for-profit organization dedicated to researching and recreating pre-17th century European history and culture. However, if you ask Lord Folo, Baron of Würmwald (the Champaign-Urbana chapter), he will tell you that, “the society was really started so we can dress up, play knight, and have fun.”

Group activities include, but are hardly limited to, arts and crafts, costume making, cooking, brewing, wine making, archery, dance, fencing, and hawking. In all of these, participants research and study in order to best emulate the methods, techniques, and styles used by our ancestors. The period of study extends from 1000 AD to 1600 AD, and anything which can be documented (or in some cases imagined) is fair game. Above and beyond all these activities, however, is one event: combat. “Combat is the glue which holds the Society together,” Folo Watkins (the Baron’s mundane name) said; “You can get rid of dance, you can get rid of fencing, you can get rid of everything else; combat holds it together. We’ve created a martial art.”
Despite the central focus on combat, there is very little residual conflict resulting from fights between members; what happens during a fight ends there. While most members create a persona for use at SCA events, this usually just represents a name, style of dress, and focus on a specific historical period. Few people walk around completely “in character” and almost none carry inter-persona conflicts into the mundane world. Any conflict that exists tends to be social or political in nature and is not much different than would be seen in any social group. Still, combat is so important that many of the highest positions in the society are awarded directly or indirectly through combat; to explain these positions, however, the hierarchy of the SCA should first be discussed.

The basic unit of the SCA is the local group. This can be a Shire, a Barony, a College, or a Canton – there is little significant difference except that a Barony is lead by a Baron. The heads of these groups are elected, and policy-making procedures vary from group to group; Würmwald works by consensus – everyone must agree on a proposal – but many groups vote or rely on simple dictatorship. For purposes of this consensus, membership is defined by attendance at the various meetings and activities. The barony of Würmwald boasts almost 60 members; of these, however, maybe twenty attend meetings regularly and help plan the events. These twenty, then, will be the focus of the study.

Baron Folo stated that many times the consensus acts as a formality; when the treasurer needs to send a report to the regional board, for example, the entire group is not called together to approve it. Similarly, during the formal meetings, consensus frequently seems to exist more as a lack of opposition to an idea, rather than unanimous agreement. This means the leadership usually does what they think best, but any group member is allowed to object or suggest new alternatives.

Above the local is the regional level. If a region has a prince, it is called a principality, but is otherwise identical to a region. A prince is chosen in combat, serves a term, then steps down and the heir is chosen by a special tournament. A retired prince receives the title of viscount, which carries no actual power, but lends a great deal of informal authority.

At the top of this neo-feudalist system is the kingdom which is lead, obviously, by a King. As with the Prince, the position of King is won in combat, and after retirement the King receives an honorary title, such as count or duke. There are currently thirteen kingdoms in the United States, and a number more around the world. Above the kingdom is the national board of directors, which runs the administrative, bureaucratic aspect of the Society. This board is self-selecting; when an existing member steps down, he or she names a replacement. Frequently this person has similar views on many key issues which frequently results in the national board having a very insular, disconnected view of the SCA.

Local officials send regular reports to their superiors at the regional and national levels, who, in turn, distribute monthly newsletter containing information about events nationwide. While the kings and board members can influence events greatly and create new guidelines, almost all of the activity and change in the SCA happens at the local level.

Gaining access to the local groups is very easy; the only requirement is attendance at meetings and events. In Würmwald all official activities occur publicly. Members are self-selected, and anyone with an interest in the historical period (or just in combat) is eligible to join. The barony of Würmwald has an event in the middle of January every year; this event, called the Festival of Maidens, and the planning which lead up to it will be the focus of this study. The local official of interest for this event will be the Autocrat. The Autocrat is the overseer of an event; as such, this person is ultimately responsible for everything that happens at the event and is in charge of assigning tasks and monitoring their completion. A new Autocrat is assigned for every event, and so it will be with the assigning of the autocrat, usually around the end of August, that the event’s timeline will begin.

It should be noted at this point that I am currently [as of the time of this writing] a member of the Würmwald barony. While this may create a potential for bias, this risk is minimal. Coding of the meetings will be done via video tapes, allowing me to participate in meetings unfettered. Research Assistants will be responsible for distributing and collecting surveys, thereby minimizing my potential for contamination there. And finally, while I will be interviewing the Autocrat myself (as explained below), the only qualitative questions asked will address the percentage of work remaining to be done, an area where any bias I may bring with should be fairly negligible.

Methods:
The primary instruments used in this study will be interviews, surveys, and videotapes of the meetings. The survey will be distributed to everyone attending the biweekly meetings (a final survey will be distributed the day of the event). Interviews with the Autocrat will be used, among other things, to keep track of new sub-groups that are formed. If these group leaders miss meetings, efforts will be taken to get a survey to them outside of the meeting. This survey will ask participants to approximate how much time they spent working on their assigned tasks in the past two weeks, what percent of their time was spent in planning, what percent in preparation, and what percent in execution. Next they will be asked to estimate how much of the time they spent on the project was in meetings and how much happened informally. Finally, they will be asked how much of the work on their part of the project is complete. A copy of the survey can be found in Appendix B.

Interviews will only be used on one person: the event’s Autocrat. Since the survey is being used to keep track of the group as a whole and all information regarding task completion will make its way to the Autocrat, any information that could be gained from interviews with other members can also be gained by interviewing the Autocrat. To save time, therefore, this will be the only person interviewed. Once Autocrat selection has occurred, an interview will be used to perform a task analysis of the event. This will be used to determine what needs to be accomplished, what work will be delegated, what sub-groups will be created, and what tasks each of these groups needs to perform. Using this information, the research team can keep track of how much of the project is completed and compare this to an “ideal” schedule. After that, interviews will occur once a month just to touch base and get the autocrat’s perspective on how much work remains, as well as to receive formal updates on any new sub-groups that may have been formed.

Videotaping will occur at each of the biweekly meetings the SCA holds. A camera will be place in the corner of the room, hopefully catching the entire group or at least as much as possible. These tapes will be used by the research team to later code the meetings according to the previously established guidelines.

After all the information has been collected and the event is complete, the results will be considered. Following is a brief explanation of each hypothesis explaining how the data will be used.

H1: Planning for the event will occur in a multi-phasic model.
If planning occurs in a plain phasic model, then introduction, argument, side-taking, and assistance will each dominate the group discussion for a period. However, if this is the case they will each dominate once and only once, and will be seen in the order they are listed here. If a behavior dominates more than once, it will be a sign that the group is following a multiphasic decision-making scheme.

H2: Orientation will be very brief or non-existent due to the group’s prior history.
Very little time will be spent on introduction during the meetings. This will be difficult to measure, as one orientation event (such as answering the question, “How do I sign up for combat?”) can last through several of the five minute segments. Regardless, this phase will most likely be very brief overall.

H3: Planning behaviors will be the primary activity during phase one.
This result will be based on the collected survey data and coding of video tapes; before the chronological midpoint, planning should be the most frequently reported behavior.

H4: The transition point for the project will occur substantially after the project’s chronological midpoint.
To determine this, the research team will review the surveys and interviews to see what time period most of the work occurs in. The beginning of the period when preparation and execution outweigh planning will signify that the transition point has passed.

H5: Sub-groups will show a marked increase in execution immediately before the event.
This result will also be based on the surveys; most of the reported execution should occur in the last month of the project, and especially in the last two weeks.

H6: Most of the planning will occur outside of formal meetings.
One of the survey questions asks how much planning occurred in meetings and how much occurred outside. The results of this item will pertain to H6, showing the research team when most planning occurs.




Conclusions:
This study is important for a number of reasons. Chiefly, it provides an opportunity to test communication theories in a setting far different from the controlled, sterile laboratory. If results from this study find support for these theories, their predictive power and validity will increase greatly. If the results contradict these theories or find flaws, however, it will open a doorway to new areas of research as future studies will need to address the reasons the theories did not fit.
In addition, the primary focus of this study is observing how this group works. Once we know more, future studies will be able to use this group and others like it with some reasonable confidence that the group is understood. Better understanding of the group will allow it to function as a natural laboratory, providing a space to test future theories or modifications to existing ones. Once we know, for example, how Gersick’s theory applies to the Barony, a future study could go in and explore ways to manipulate this; can steps be taken to extend or shorten the length of phase one? Will the transition point occur without prompting from the Autocrat? Before these studies can be conducted, however, the group must be understood.
The third major benefit of this study is it provides more experience working with naturalistic groups. The SCA is prone to as many sources of variance as almost any group imaginable; by learning how to conduct studies in such a chaotic setting, we can conduct future naturalistic studies with more confidence that our methods and techniques will succeed.
The study does leave a few openings for future research, as well. First and foremost, this study only uses one group. Since local groups in the SCA can be completely different from one another, doing the same study with a different group could provide different results. Further research would extend the validity of this study.
In addition, the question remains of how these results would apply to other settings. Would they fit in a corporate environment, for example. Or perhaps the results would be different in larger groups, or groups with more limited focus. This same study could be done with many types of groups in many different contexts to help fully develop these communication theories.