Abstract
This article proposes a markedly new conceptual approach to group and social interaction analysis, grounded in transformative advances in dynamic network theory. The framework first theoretically identifies the small set of behavioral elements that can be influential across social contexts. Example behavioral elements include goal striving to advance a communication or viewpoint, system supporting of others, goal preventing, and affect-based system negating. The framework distinctively allows researchers to simultaneously combine the elements to explain complex units, such as a person disagreeing with another (goal prevention), but in a warm and caring way (system support). The approach also provides new insight into how behavioral elements elapse over time, such as our distinct characterization of support cycles (e.g., reinforcement, comfort, and submissive) and conflict cycles (e.g., basic, heated, and acidic). The approach uniquely demonstrates how lower-level behavioral elements can be used to predict higher-level emergent states and climates, such as hostile climates.
Keywords
The analysis of group and social interactions is of critical interest to a wide range of social and organizational scientists. However, many of the analytic approaches tend to be highly tailored to specific domains, thereby preventing a universal theoretical explanation that generalizes across social contexts and settings. For example, separate analytic approaches have been developed for small groups (Bales, 1950; Futoran, Kelly, & McGrath, 1989; Lehmann-Willenbrock, Meyers, Kauffeld, Neininger, & Henschel, 2011), negotiations (Druckman, 2003; Olekalns, Brett, & Weingart, 2003), and couples’ interactions (Gottman, 1998), to mention a few. A potential limitation is that many of the analytic dimensions in one domain may not translate to the dynamics in another domain because of the targeted focus of each approach. Although the genesis for such tailored research is quite understandable, often based on research expertise or specific areas of interest, it may preclude a wider theoretical understanding of the critical dimensions that generalize across any social setting, be it groups, dyads, couples, social networks, etc. Thus, broader theory is needed that allows for such generalizability.
On the other end of the spectrum, some researchers may rely on just one or two very broad analytic dimensions of interest to portray events in group and social interactions, such as coding the observed interaction events as being negative, positive, or neutral (see Gottman, Swanson, & Swanson, 2002, for a review of such studies). However, such simplifications can miss many of the common and rich complexities that occur across settings (Ekman, 1993). It may also result in confounded analyses because of oversimplified coding schemes. For example, forcing the analysis of social interaction events into negative, positive, or even neutral classifications, which seems quite intuitive, can miss situations where individuals are simultaneously manifesting behaviors that contain both negative and positive elements, such as a person that is arguing against another’s viewpoint (negative), while simultaneously doing so in a warm, caring, and empathetic manner (positive). Thus, more creative theorizing is needed that allows researchers to understand such complexities, while also being grounded in a parsimonious set of parameters that generalize across contexts. To address these issues, the purpose of this paper is to extend recent advances in dynamic network theory (Westaby, 2012; Westaby, Pfaff, & Redding, 2014) to build a new integrative framework of group and social interaction analysis that captures important behavioral complexities across contexts, while being grounded in a parsimonious set of theoretically based parameters. Dynamic network theory is particularly helpful for this objective, because the approach integrates theory and research from the psychological and social network sciences, which provides considerable generalizability across settings and contexts. However, our proposed framework substantively extends previous work on the theory, which has not examined group and social interactions in real time, the modeling of behavioral cycles, and the specific way in which behaviors result in various types of emergent states and climates. Westaby et al. (2014) called for more research examining such extensions.
This paper further attempts to address some of the shortcomings leveled against group scholarship. Though groups are frequently conceptualized as complex and adaptive systems (Arrow, McGrath, & Berdahl, 2000) that are informed by a social context over time (Ilgen, Hollenbeck, Johnson, & Jundt, 2005; Lepine, Piccolo, Jackson, Mathieu, & Saul, 2008; Sonnentag, 2012), various scholars have noted that the group literature has insufficiently incorporated the emergence, development, and impact of dynamic behavioral forces seen in group settings over time (Cronin, Weingart, & Todorova, 2011; Ilgen et al., 2005; McGrath, Arrow, & Berdahl, 2000). Furthermore, past research has not provided a sufficiently deep understanding of the rudimentary behavioral elements that explain the rich dynamics involved in group and social interactions across contexts, based on contemporary theory. Our framework attempts to provide this explanation.
As for the boundaries of our framework, we focus on those problems of interest to social and organizational scientists related to social exchanges occurring among individuals in groups, dyads, social networks, social media, or other structures of human interaction. The approach does not examine those tasks that do not involve social exchanges, such as when teams work on goals alone or in silence. Other broader models of social and organizational behavior can provide insight into those processes, including original work on dynamic network theory (Westaby et al., 2014). Moreover, we do not tap extensively into presumed unconscious processes, which have been discussed in other approaches (Noumair, 2013), although we discuss the importance of hidden psychological elements in our section on distortion and deception processes.
In our proposed framework, we first extend and modify the key parameters identified in dynamic network theory to uniquely describe the behavioral elements underlying group and social interactions, including multiplex behaviors that simultaneously contain a combination of elements. We then demonstrate how the small set of behavioral elements and multiplex units can be joined to explain other complex patterns that commonly occur in group and social interactions over time that have heretofore not been sufficiently articulated, such as those related to several types of support and conflict cycles. We will further demonstrate how researchers can use the behavioral elements in the framework to predict emergent states and climates, which has also been lacking in past research. Lastly, we concretely illustrate how the framework is used to analyze a hypothetical group interaction in efforts to demonstrate its potential use in research.
The fundamental behavioral elements
Dynamic network theory states that a finite set of eight social network role behaviors are needed to explain key goal pursuit and behavioral processes in social networks: (a) goal striving, (b) system supporting, (c) goal preventing, (d) supportive resisting, (e) system negating, (f ) system reacting, (g) interacting, and (h) observing (Westaby, 2012; Westaby et al., 2014). 1 The theory postulates that individual-level judgment and decision-making processes influence the enactment of these social network roles, which in turn impact goal achievement, performance, learning, as well as other dynamic outcomes, such as the network rippling of emotions, in a wide variety of social and organizational contexts. In this paper, we use and extend the social network role concept to portray fundamental behavioral elements underlying group and social interactions. Behavioral elements are defined as the verbal and nonverbal expressions of individuals, which drive group and social interactions in real time. In this paper, we use these behavioral elements to indicate how individuals are oriented around a presumed general goal (explicit or implicit) of advancing a communication or viewpoint during group or social interactions. This application differs from traditional dynamic network theory, which examines how social network processes influence a wide variety of goals. Through defining the goal and behavioral elements around a more general goal concerning social interaction, we are able to quantify the frequency of the behavioral elements manifest during interaction events and calculate individual- and group-level metrics to describe the dynamics of the interactions, to be described more fully in a later section. Furthermore, we also uniquely include the analysis of system competency from dynamic network theory as a behavioral element in our framework as further delineated below. In the sections that follow, we first describe the two primarily functional behavioral elements followed by a discussion of the antagonistic, observational, and system competency elements.
Functional relations
The first behavioral element relevant to our framing of group and social interactions is communicative goal striving. In its pure form, this goal striving behavior represents individuals that are trying to verbally (e.g., oral or written) or nonverbally (e.g., hand, body, or face gestures) advance a communication or viewpoint without antagonism or questioning. This is in line with the presumed underlying goal in many social interactions to advance a communication or viewpoint. Our use of a general communicative goal allows us to apply dynamic network theory parameters to the analysis of complex group and social interactions in real time. Such broad goal striving behavior during interactions may be cooperative in tone (Deutsch, 1973; Pruitt & Kim, 2004), focused on integrative information (Olekalns et al., 2003), or focused on simply conveying information content. 2
The second functional behavioral element is system support behavior. This occurs when participants appear to show authentic support, empathy, or positive affective regard toward others, verbally or nonverbally. It could also show a concern for others (Pruitt & Kim, 2004), facial expressions of enjoyment (Ekman, 1993) generated from a person’s reaction to another’s behavior in the interaction, or empathetic behaviors directed at others (Ivey, Ivey, & Zalaquett, 2014). In line with dynamic network theory, we anticipate that the more goal striving and system supporting observed in social exchanges, the more positive the overall interaction or climate would be judged by external observers (e.g., rating the interaction as productive or cooperative) or by the participants themselves (e.g., self-reporting satisfaction with the interaction or low stress levels). 3
Resistance forces
The next two behavioral elements in the framework address antagonistic forces. First, communicative goal preventing behavior, generally speaking, represents situations where individuals are working against another’s attempt to advance their communication or viewpoint. They are not sharing in the communication or viewpoint being advanced. Uniquely, pure forms of goal preventing do not include expressions of affect, negative emotion, or questioning, which are accounted for by other roles or multiplex (combination) roles in our framework, which will be discussed in a subsequent section. For example, pure goal prevention can be seen when group members are disagreeing in a professional and nonhostile manner. Although tangentially related to the notion of disagreement codes in classic group models (Bales, 1950) and bargaining process analysis (Druckman, 1986), pure goal prevention in our theoretical portrayal uniquely contains no negative affectivity in the argumentation. This theoretical difference is critical, because it allows us to differentiate between heated versus nonheated conflicts in the context of other behavioral elements as we further illustrate next in the section on negative affect and multiplicity. It also allows us to identify higher-level patterns of interaction, such as basic and heated conflict cycles as discussed in what follows.
Second, supportive resisting behavior represents individuals that are agreeing with another person’s goal prevention that is targeting a third party in the group or network. This most commonly occurs in groups, such as if Pat disagrees with Dana’s initial communicative goal striving and two other people in the group simultaneously nod in agreement with Pat’s antagonism toward Dana (i.e., the two individuals are displaying supportive resistance). We are not familiar with previous social interaction frameworks that have differentiated this behavior that illustrates indirect support for another’s communicative goal prevention or negativity.
When group members are engaged entirely in goal prevention and supportive resistance without moving back toward communicative goal striving or system supporting, we anticipate that the overall social interaction or climate would likely be viewed as negative in various ways, such as being low in cohesiveness, given that some communicative goal striving or support is usually required to move parties forward during interactions. However, it is important to note that some forms of goal prevention can represent constructive resistance (Westaby et al., 2014), such as when an individual’s disagreement with another actually illustrates how the other person could engage in a different form of behavior that provides the person with better outcomes. 4 It can also alleviate the potential for groupthink by inserting goal prevention dialogue that attempts to fix a faulty group strategy or course of action. However, even in such cases, at some point in the future one would expect that the parties would converge back to some communicative goal striving or system supporting to illustrate a functional shift in the interaction.
Negative affect
The previous behavioral elements did not explicitly account for the role of negative affect and emotions in social exchanges, which can be important during interactions (Gottman et al., 2002). Therefore, in this section, we uniquely demonstrate the importance of two such types of affective units. First, system negation represents people’s negative affect assertively expressed toward others, either verbally or nonverbally. In pure cases of system negation, the behavior shows negativity with no attempts to change a situation or communicate disagreement with others. For example, pure system negation may be observed in groups when individuals take delight in other people’s failures, misfortunes, or poor performance (e.g., laughing at their expense). Past approaches have not sufficiently accounted for this form of negativity and have often generalized this under dimensions of disagreement or tension, which may limit precision in the analysis.
As another unique contribution over past work on dynamic network theory (Westaby, 2012; Westaby et al., 2014), we extend the conceptualization of system reactance behavior in the context of human interaction in real time as passive negative affect revealed in reaction to another’s behavior in the exchange, such as crying in response to another’s communicative goal prevention. The potentially universal expressions of sadness and fear (Ekman, 1993) could also be accounted for by system reactance as well, when such reactions are to another’s behavior in the interaction.
Observer questioning behavior
Many of the behavioral elements discussed thus far have explicit or implicit motives to advance or prevent a given communication or viewpoint. In contrast, we conceptualize observer questioning behavior as occurring when one person is motivated to observe another person’s communication or viewpoint and thus asks questions in service of this observational motive. For example, a team member asking the question “What should we do next?” would represent one behavioral element of observer questioning. This dimension is somewhat related to Bales’s (1950) questioning category, although our framework goes beyond this one-dimensional characterization by allowing researchers to combine this behavioral element with others to explain more complex dynamics in group and social interactions, as illustrated below. 5
Competency-related behavior
Past theory addressing group and social interactions has not sufficiently accounted for how the competency of participants’ behavior can be described and quantified through the use of a finite set of behavioral elements in the network and psychological sciences. Nor has dynamic network theory’s construct of system competency been sufficiently applied to social interaction analysis. Hence, in our extended framework, system competency represents the degree to which social exchanges are efficacious or nonefficacious in their verbal or nonverbal delivery. Practically, we anticipate that it is often easier to observe or score system competency elements in the nonefficacious form. For example, this can be observed when group members’ verbal or nonverbal behaviors are manifest with minor awkwardness, such as “aaaahhhs” or “ummmms” or more major phrases or statements that are confusing. For example, a teammate saying: “I’m thinking, ahhhh, about going home” would represent an element of goal striving, followed by an element of low system competency, followed by one more element of goal striving. In contrast, a person saying “New Zealand is an island in the Philippines” would represent a single multiplex goal striving and low system competency behavioral unit (i.e., a statement that is confusing). We further discuss the importance of multiplex behavioral characterizations next.
Multiplex behavioral units
We assume that human behavior is far more complex than behavioral elements being manifest in isolation, such as individuals solely displaying communicative goal striving. In our integrative framework, many observed behaviors are presumed to represent a simultaneous combination of behavioral elements, which we refer to as multiplex behavioral units, extending parameters in dynamic network theory (Westaby, 2012). Such an approach has not been advanced in past group and social interaction analysis based on parameters in this theory, to our knowledge, and represents a unique way to more fully understand the complexities of observed human interaction. For example, while one group member may be simply talking in goal striving mode, another group member may be talking in goal striving mode, while smiling in kind support toward the recipient of the communication at the exact same time. This simultaneously observed event represents a combined goal striving and system supporting multiplex behavioral unit. This type of multiplex behavior is so common during group and social interactions that we refer to it as striving-support behavior. This conceptualizing also goes beyond important work on code combinations, which have been helpfully shown to make codes mutually exclusive (Bakeman et al., 2005), by grounding the combinations upon the finite set of parameters in dynamic network theory. We assume that this aids in the generalizability of our integrative approach.
Another common multiplex behavior that communicates antagonism occurs when system negation is expressed with goal prevention, such as a person arguing in a hostile or negative tone against another’s viewpoint, thereby impeding the other’s progress towards the implicit or explicit goal of advancing a communication or viewpoint. This behavior is so common during heated conflicts that we refer to it as negation-prevention behavior. Nonverbally, negation-prevention is also likely seen in presumed universal facial expressions of anger and disgust (Ekman, 1993) when directed at another. 6 A more nuanced example of multiplex behavior is seen when system reactance and system support are joined in multiplex form as well, such as a person crying because of what someone else said, while nodding in agreement with the comment. In Table 1, we propose a number of possible multiplex examples, which future research needs to test, modify, or extend based on empirical findings, particularly with regard to cultural differences, given their potential effect on social interactions (Druckman, 2003). Our aim in this article is to establish a general theoretical approach that can help ground and guide future empirical research.
Exploratory examples of behavioral elements and multiplex units.
Note. g = communicative goal striving; s = system supporting; p = communicative goal preventing; v = supportive resisting; n = system negating; r = system reacting; o = observer questioning; c = low system competency utterance or phrase. The order of multiplex letters are alphabetically arranged for consistency. Numerous other combinations are possible. Behavioral elements and units are analyzed in relation to the motivational orientations between the exchange partners themselves in this initial framework (e.g., a person talking to a friend about a third party not in the conversation that is making the person really angry would be scored as communicative goal striving between the exchange partners, not negation-prevention).
The emergence of behavioral cycles and patterns
Behavioral elements rarely occur randomly. As a fascinating human tendency, they often emerge into common and stable patterns or subpatterns within group and social interactions, somewhat akin to how fixed-point attractors in dynamical systems theorizing help stabilize systems around common or frequent behaviors (Gottman et al., 2002). In this section, we introduce a spectrum of behavioral cycles that often occur dynamically in interactions and explain how the parameters in our framework uniquely capture and characterize these dynamics.
Goal striving cycles
One of the simplest and most common cycles of behavioral elements is the goal striving cycle. This occurs when individuals engage in continuous communicative goal striving without any antagonism, negative affect, or questioning from anyone in the interaction. This frequently happens when individuals are sharing stories or information. A highly cooperative professional conversation where knowledge is being disseminated and integrated is another common example of a goal striving cycle. These cycles continue as long as the individuals in the interaction are only showing communicative goal striving.
Support cycles
Individuals commonly show repeated support to one another during interactions. Grounded in the behavioral elements in our framework, we discuss three types of support cycles in this section: reinforcement, comfort, and submissive.
Reinforcement support cycles
First, reinforcement support cycles occur when at least one party in a group or social interaction is continuously showing system support in response to another’s communicative goal striving or system supporting, without any party manifesting other antagonistic or questioning behaviors. For example, teammates may go back and forth with goal striving statements with corresponding nods of agreement. In the couples literature, these cycles are similar to the notion of “positive reciprocity sequences” (Gottman, 1998), although our approach is generalizable to group and team settings and we make explicit assumptions about how to code these cycles in the context of other behavioral elements as illustrated in the Appendix. Such cycles are also likely influenced by common reciprocity norms, where people are often motivated to reciprocate other people’s behaviors (Weingart, Prietula, Hyder, & Genovese, 1999). This is also akin to Deutsch’s (2006) “crude law” where a previous behavior is likely to motivate a similar behavioral response. One may also see relatively longer sequences of one-sided reinforcement support cycles in leadership -follower settings when followers are responding affirmatively to most, if not all of the phrases emanating from a powerful, respected, or charismatic leader. Individuals may be motivated to show such support in efforts to be viewed positively by those speaking or to possibly garner benefits in the future, for example. Or, it may be associated with agreeable personalities (Cuperman & Ickes, 2009). 7
Comfort support cycles
In contrast to reinforcement support cycles, comfort support cycles are reparative or nurturing in nature, generated from a situation where a party is providing comfort to another who is feeling badly or remorsefully during the interaction. Specifically, these cycles occur when at least one party shows continuous system support to another’s system reactance. To illustrate, one group member may provide comfort to another who is upset because of what another person just said. This support pattern is often meant to provide a reparative function by providing social relief to group members that are experiencing stress, anxiety, or emotional upheaval. In group theory, this is consistent with Marks, Mathieu, and Zaccaro’s (2001) notion concerning affect management: “Techniques involved in regulating emotions may involve attempts to calm members down, control frustration levels, boost team morale and cohesiveness among members, and provide empathy” (p. 369). However, our framework extends such theorizing through the delineation of the exact behavioral elements that constitute these acts during group and social interactions in real time. In some domains of group dynamics conceptualizing, individuals frequently engaged in providing support during these cycles may identify with broader roles taken-up by members embedded in the group, such as the “harmonizer” (Benne & Sheats, 1948).
Submissive support cycles
As a sharp contrast to the aforementioned support cycles, submissive support cycles illustrate situations where individuals are submissively showing support to others that display resistance or negativity toward their behavior. As for behavioral elements, these cycles occur when at least one person is showing system support (e.g., agreement) with another’s goal prevention, supportive resistance, and/or system negation toward them. This often happens when a person recognizes that he or she has behaved incorrectly or inappropriately and is agreeing with another individual’s disapproval or criticism, such as goal prevention. This pattern can promote a reparative function after human errors and mistakes, because the system support signals or implies that the individual will not repeat the behavior in the future, thereby showing a commitment to change. This also relates to apology and forgiveness mechanisms in social relations (Brett et al., 2007). Moreover, we anticipate that submissive support cycles will occur more frequently during conflicts between group members that have differential or asymmetrical power (Druckman, 2003). Lower power members may be more likely to exhibit this pattern, sometimes undergoing considerable emotional labor (Hothschild, 1983), to ensure that higher power members continue to believe that the lower power individuals will fall in line with group expectations and leadership. 8
Conflict cycles
Conflict cycles represent situations in which individuals are systematically engaged in ongoing conflict behaviors with one another. We assume in our framework that conflict behaviors represent goal prevention, supportive resistance, system negation and/or system reactance. We delineate three types of conflict cycles in this introductory paper: basic, heated, and acidic. Each is considered in turn.
Basic conflict cycles
First, basic conflict cycles (or nonheated conflict cycles) represent exchanges where at least one person is continuously engaged in pure goal prevention or supportive resistance, where no negative affect is manifest. Employees engaged in a prolonged professional argument or a disagreement without any hostilities or negative emotion is a good example. Interactions entirely characterized by such cycles without any movement to communicative goal striving or system supporting are predicted to be perceived as nonproductive, nonconstructive, and potentially competitive. Ironically, however, we assume that such conflict can represent constructive resistance (Westaby et al., 2014), if new learning emerges and cooperative goal striving or system supporting becomes manifest by the end of the interaction. Such system support at the conclusion of the interaction may signify that agreement has been reached. Negotiation scholars have also noted how crises, which we presume often trigger a strong form of goal prevention, may functionally alert participants about problems that need solving before moving on to more constructive stages (Druckman, 1986).
Heated conflict cycles
Heated conflict cycles occur when at least one person is continually responding with system negation, system reactance, or relevant multiplex expressions, such as negation-prevention. For example, if at least one of the professionals in the previous example were stating their arguments with angry tones or hostile words, it would represent a heated conflict cycle. We assume that heated conflict cycles are somewhat related to various interaction sequences observed among couples in the counseling literature, such as “negative reactivity” (Margolin & Wampold, 1981) and longer chains of “negative continuance sequences” (Gottman, 1998). Gottman et al. (2002) have also summarized that “It has been consistently found, in marital interaction research, that the reciprocation of negativity is more characteristic of unhappy than of happy couples” (p. 332). As a more generalized approach across contexts, our framework importantly distinguishes between those negative cycles that are heated or not heated in terms of displayed negative affect—creating a new method for the differentiation between types of negativity. Theoretically, heated cycles may also generate more intense “negative absorbing states,” applying Gottman’s (1998) theorizing, where nondistressed dyads or groups would escape from these harmful sequences while distressed dyads or groups could not escape once they entered. One may anticipate that heated conflict cycles exhibit greater absorbing states than basic conflict cycles, which future research should examine in different contexts. We also anticipate that basic or heated conflict cycles will be more likely to appear during Marks et al.’s (2001) transition phases than action phases.
Acidic conflict cycles
Unfortunately, some behavioral elements are combined into patterns that intentionally display torment to others. For example, we conceive of acidic conflict cycles as representing situations where one individual repeatedly enacts pure negation or negation-prevention toward another who is passively responding with system reactance, such as tears or emotional angst. Some individuals may believe that these cycles are well deserved, such as a mother screaming at a person convicted of killing her son through vehicular manslaughter, where the perpetrator is crying in response to confronting the mother and the guilt of the act. In other cases, a bully may continually show pure system negation by making fun of a victim who is continuously crying because of the bully’s acidic behavior, which to most people would constitute deplorable behavior. Acidic conflict cycles in groups or social interactions highlights a unique type of conflict behavior that is highly saturated with affectivity, which has not been sufficiently differentiated in past research.
Future research should investigate other cycles as well, such as observer questioning cycles, which may appear in the context of the general roles that group members implement. To illustrate, these cycles may occur among journalists, interviewers, students, interrogators, and investigators. Such cycles would likely cease or transform into other roles when adequate responses are provided or the general requirements of the role are satisfied (e.g., the questions have been answered).
Conflict escalation, spirals, and de-escalation
In our new framework, behavioral elements, multiplex units, and goal striving, support, or conflict cycles can be uniquely joined to explain additional conflict escalation and de-escalation processes during interactions. As a first basic mechanism, we anticipate that an individual’s initiation of conflict behaviors, which we define as goal prevention, supportive resistance, system negation and/or system reactance, may increase the likelihood that another person will introduce reactionary conflict behaviors, thereby resulting in the initial escalation of a conflict. This is consistent with conceptions of reciprocity (Weingart et al., 1999), the crude law of social relations (Deutsch, 2006), and tit-for-tat responses in conflict (Pruitt & Kim, 2004). Or, underlying decision-making factors or reasons may trigger the conflict response (Westaby, 2005; Westaby et al., 2014).
The behavioral elements in our framework can further distinguish important types of behavioral escalation. For example, basic conflict escalation represents situations wherein goal prevention and/or supportive resistance behavioral elements become more extreme or frequent over time, but do not transform into heated conflict. For instance, a supervisor may professionally tell an employee to not have any more unexcused absences from work in a quick conversation. When the employee repeats the behavior, the supervisor increases the frequency of communicative goal prevention elements during a longer conversation, warning the employee that his or her employment is in jeopardy. In contrast, heated conflict escalation represents situations where nonconflict or basic conflict behaviors consecutively transform into heated conflicts, or the heated conflicts themselves increase in intensity over time. To illustrate the former, a basic conflict behavior may be followed by heated exchanges of negation-prevention behavior. We also anticipate that basic conflict cycles are more likely to escalate into heated conflict cycles when critical underlying goals remain unaccomplished for too long and the basic conflict dialogue remains unproductive, given that the parties may be sensing a need for closure (DeGrada, Kruglanski, Mannetti, & Pierro, 1999). For example, in the workplace, Sofia may calmly try to change the mind of her colleague, Janet, through pure goal prevention, but start to include system negation, such as anger in tone or hostility in wording when she realizes that Janet is continuing to persist in an action that she perceives to be inappropriate or immoral.
Research needs to examine if the various conflict-oriented elements and units have differential power in the flow of interactions as well. For example, Gottman et al. (2002) generally observed that “during conflict, negativity has a bigger impact on one’s partner’s immediately subsequent behavior than positivity” (p. 335). However, it is unclear if this generalization holds across our differentiation of goal prevention, supportive resistance, system negation, system reactance, and their various combinations, although we anticipate that such effects for negativity, especially nonlinear ones, are most relevant to highly emotionally charged elements or units, such as negation-prevention and pure system negation, and less so for highly self-regulated pure goal prevention.
Conflicts can transform into conflict spirals as well (De Dreu, 1997; Olekalns et al., 2003; Pruitt & Kim, 2004). To differentiate spirals from incremental conflict escalation, we conceive of a conflict spiral as a nonlinear amplification of conflict behaviors over time, often where resources, desires for cooperative goal striving, and/or system supporting behavior become highly restricted between relevant parties. For example, two group members may go from calmly disagreeing to extreme shouting where both parties eventually feel little or no desire for further contact, discussion, or willingness to understand one another’s point of view. In dynamical systems thinking, a catastrophe process has occurred where a positive stable steady state no longer exists (Gottman et al., 2002). In the conflict literature, the seeds for an intractable conflict may have also emerged (Vallacher, Coleman, Nowak, & Bui-Wrzosinska, 2010).
Different behavioral cycles can also be combined to reflect even higher order stage processes, which have been popular in the group dynamics literature for some time, especially to portray longer time horizons of group development (Kozlowski & Bell, 2003) or to imply gradual shifts in strategy (Olekalns et al., 2003). As a micro demonstration in the context of group and social interactions, some conflicts may go through an extensive escalation–de-escalation stage model, such as the following: basic conflict cycles → basic conflict escalation → heated conflict escalation → heated conflict cycles → heated conflict spiral → heated conflict de-escalation → basic conflict de-escalation → goal striving cycles → support cycles. The embedded process of conflict-to-support in this staging is consistent with the flow from distribution processes to integration processes in negotiation theories (Olekalns et al., 2003), although our framework provides new insight into the behavioral elements underlying these mechanisms. It is also important to note that stages often represent benchmarks or turning points to participants, which can be generated by important identifiable events according to Druckman (1986). Applied to our framework, if negotiations are unable to move beyond basic or heated conflict cycles, it would likely represent a serious impasse in the system.
In nonlinear cases, some stages may be skipped, consistent with empirically supported episodic and phase models (Baxter, 1982; Holmes, 1992, 1997). For instance, a heated conflict cycle could go directly to widespread cooperative goal striving, such as when two competing groups stop goal preventing verbal behavior and cooperatively merge their efforts against a suddenly more threatening third party, thereby serving a superordinate goal. Of course, relapses can occur, especially in historically intractable conflicts, such as a period of de-escalation regressing back to heated conflict cycles. Goal striving cycles can also suddenly transform into conflict escalation when there are threats to talks, which past theorizing may regard as a crisis to the negotiation process (Druckman, 1986). 9 Future research could also use mathematical modeling to estimate set-point attractor dynamics (and influence functions) of the various parameters in the theory, building from past work on mathematical modeling during social interactions (Gottman et al., 2002).
Propositions could also be advanced about the probability of behavioral elements occurring at different stages of interpersonal conflict, based on past research. For example, in line with Gottman’s (1998) dyadic counseling framework, one may expect that negativity (e.g., negation or negation-prevention) and repair attempts (e.g., goal striving and support cycles) are more likely to occur during middle arguing phases. Olekalns et al.’s (2003) integrative work that synthesized past group, negotiation, and communication models also demonstrated how negotiations often start with distributive phases (e.g., conflict behavior in our approach) and end with integrative phases (e.g., system support in our framework). Hence, one may expect to find greater conflict cycles early in negotiations, ending with greater system support cycles. Starting in a distributive manner may also help ensure that negotiators are able to express their needs (or concerns) to find satisfactory outcomes as found by Olekalns et al. (2003). 10 Our framework substantively adds to this work by providing more detailed analyses of the behavioral elements, multiplex units, and behavioral cycles that underlie the interactions over time.
Emergent states and climates
The concepts of emergence, emergent states, and climates have received considerable attention in the group and organizational sciences. Marks et al. (2001) defined emergent states as “constructs that characterize properties of the team that are typically dynamic in nature and vary as a function of team context, inputs, processes, and outcomes” (p. 357). Unfortunately, while considerable research has examined emergent states and climates using survey methods (Schneider, Ehrhart, & Macy, 2013), relatively less attention has been paid to how emergent states and climates occur in real time group and social interactions. For example, how do researchers or the participants themselves judge various states, climates, or phenomena that perceptually emerge about a given interaction sequence? To this end, we address how the perception of general emergent states and climates during interactions can be based on the underlying behavioral elements observed in those interactions.
To calibrate broader emergent states and climates to behavioral elements, units, or cycles, we first make new distinctions about emergent fragment and summation states. That is, we define emergent fragment states (or climate fragments) as broad constructs that portray overall characteristics or qualities related to specific time intervals within a group or social interaction. For example, a person may say “the first 10 minutes of the meeting went great, while the last 10 minutes were just OK.” This could illustrate two emergent fragment perceptions about “meeting performance” that characterize an important quality of the 10-minute interaction. As a methodological example, observers could code a group’s cooperative climate fragment in preset intervals, such as every 2 minutes of the interaction. 11 This notion about fragments is somewhat related to the idea of episodes in Marks et al.’s (2001) model, although we initially focus these global fragment qualities or characteristics on equal time intervals occurring within social interactions in order to calibrate them to underlying behavioral elements and units occurring within those same corresponding intervals. Episode fragments may be more variable.
Second, emergent summation states (or climate summations) are broad constructs that portray overall characteristics or qualities related to an entire group or social interaction under analysis. To continue our previous example, as a perceived emergent summation state, a person may describe the overall meeting as going “fairly well.” Future research will need to examine which psychological dimensions or factors represent the most relevant emergent states or climates for different group and social contexts, such as the social interactions being rated as cohesive, productive, cooperative, hostile, or the participants rating their satisfaction with the interaction (Brett et al., 2007; De Dreu & Weingart, 2003; Deutsch, 1973). Although emergent state concepts and climates can be assessed on group behavior where little or no interaction occurs with survey methods, our focus in this paper is on emergent states and climates that are observable during group and social exchanges.
In our framework, we hypothesize that the frequency of the observed behavioral elements occurring during group and social interaction periods will be related to relevant emergent fragments and/or summation ratings for those same periods and potentially future periods. For example, the greater the frequency of system support observed during a specific fragment should be related to higher ratings of positive climate fragments for that same time period. In contrast, greater frequencies of goal prevention, supportive resistance, and/or system negation during specific fragments should be related to higher ratings of negative climate fragments during those fragments, such as a hostile work climate. Future research should also examine the degree to which lower-level behavioral elements contribute to predicting overall ratings of emergent summations across full interactions, using multilevel analyses.
The theorizing here is further grounded in Cronin et al.’s (2011) observation that emergence “describes a process where a higher-level phenomenon comes into existence based on interaction among lower-level elements” (p. 574). Grounded upon Morgeson and Hofmann’s (1999) framework, Cronin et al. (2011) further surmised that “elemental interaction is what creates emergence” (p. 576), which is highly consistent with our approach that focuses on observable social interaction in real time. That is, lower-level behavioral elements are assumed to generate an array of higher-level emergent states or climates for the interactions. The selected emergent state or climate concepts could be derived from the participants themselves, such as in pilot elicitation research or verbal protocol analysis, or selected by researchers, based on previous research or the research question.
Likewise, our framework can explain how lower-level processes result in metaphorical bubbling-up effects (Druckman, 2003). Specifically, the manifestation of behavioral elements at a micro level of granularity (Bakeman & Quera, 2012) would theoretically trigger the bubbling-up of emergent states or perceived climates at a more macro level. This conceptualizing further relates to Klein and Kozlowski’s (2000) notion of configural team properties that “originate in, or emerge from, individual team members’ experiences, attitudes, perceptions, values, cognitions, or behaviors” (p. 217). Our approach adds to this literature by detailing the specific types of behavioral elements that can serve as key antecedent variables.
Emergent states and climates can also provide feedback to team members and set expectations for future behavior, extending notions by Chen and Kanfer (2006). Cronin (2011) also recognized that “it is not just that global dynamics emerge from local ones; it is that local and global dynamics can feed back upon each other within a single construct to create nonlinear growth” (p. 583). In a related vein, research needs to examine how past emergent states and climates systematically influence the observed frequencies of future behavioral elements in group and social interactions. Theoretically, the emergence and stabilization of a highly cooperative climate summation, to illustrate, may serve as an attractor (Gottman et al., 2002; Vallacher et al., 2010) that sets expectations for how all group members should implement behavioral elements and units in future interactions. In this illustration, group members might be expected to show more goal striving and system supporting behavioral elements without conflict behavior. This may also reflect a type of dominant strategy orientation, as discussed in the negotiation literature (De Dreu, Weingart, & Kwon, 2000). In contrast, an interaction that was previously considered to be an uncooperative climate summation may generate expectations that goal prevention and perhaps system negation behavioral elements will appear in the future and hence prepare participants for battle (or defensiveness) instead of engaging in a more cooperative mindset (Brett et al., 2007). These propositions also illustrate how higher-level constructs can influence lower-level behaviors through cross-level mechanisms, consistent with Chen, Kanfer, DeShon, Mathieu, and Kozlowski’s (2009) model. This conceptualizing builds on the reciprocal and cyclical effects observed in the negotiation literature (Druckman, 2003) and Marks et al.’s (2001) proposition that emergent states act as inputs that can impact teamwork processes. Our framework adds much needed specificity about the underlying behavioral elements and multiplex behavioral units at the base of such group interactions, which have received insufficient attention in past multilevel theory.
Distortion, deception, and impression management
A challenging yet fascinating issue related to the observation of human dynamics is the fact that some people, if not many, publicly display behaviors that may not map onto internal psychological representations of the private self (Triandis, 1989). For example, many individuals will display external behaviors that disguise their true feelings (Goldberg & Grandey, 2007), such as a subordinate nodding in agreement with a boss (system support), when in reality the person is in disagreement with the boss (goal prevention). Such behavior may be motivated by (a) impression management, such as to secure support from the boss, (b) intentions to deceive, such as sins of commission (e.g., giving false information, showing false expressions; Ekman, 1993) and sins of omission (e.g., concealing information), or (c) managing emotional displays in service of organizational goals (Grandey, 2000) through emotional labor (Hothschild, 1983), such as by engaging in surface acting, which represents the management of observable emotions, or deep acting, which involves deliberate management of internal feelings, thereby resulting in changes in observable emotions (Hülsheger et al., 2010).
Some may argue that disguised behaviors may be functional or dysfunctional in groups, depending on the context. For example, if a supervisor is able to control his or her true internal desire for hostile negation-prevention toward a sensitive subordinate that made a mistake, it may reduce the likelihood that the subordinate would overreact and become distracted and hence increase functional learning about the incident. This also illustrates the old adage that “it’s not what you say, but the way you say it,” often implying that a negative tone from system negation or negation-prevention can be dysfunctional and can work against goals of advancing a shared communication. In other cases, some individuals are capable of translating their true intentions of goal prevention (i.e., working against another’s intent) into perceived communicative goal striving among their exchange partners. To illustrate a translation, a diplomat may try to advance a highly controlled, nonthreatening viewpoint discussing “areas of opportunity” instead of displaying explicit negation-prevention toward another country’s action that may be infringing on his or her nation’s interests. Although this may reduce system reactance or tit-for-tat goal prevention or system negation in the dialogue, it may also introduce greater ambiguities about one’s intent and underlying desires, thereby impeding communication. Empirically, Brett et al. (2007) also demonstrated that showing negative emotions and commands had maladaptive effects in negotiations, implying that self-control of such expressions may be constructive. Although such a controlled strategy may work when others perceive the behavior as a sincere attempt at cooperation where a win–win situation might result, it could also cause conflict, such as negation-prevention reactions, when the other party interprets the controlled behavior as insincere, deceptive, or untrustworthy. Research needs to examine these nuances and the potential difficulties in coding a person’s ambiguous communication where intentions of underlying resistance or negativity are unclear. For now, until other guidelines are established, we recommend that such unclear or ambiguous behaviors be coded as communicative goal striving, because the communication is still literally moving forward and because clear resistance or negative affect is not saliently observed.
Further, when individuals ineffectively attempt to disguise their motives during interactions, it may generate distrust among other group members who become uncertain about the individual’s true intent. This could breed conflict responses, because individuals may feel defensive and self-protective instead of focusing on open group dialogue (Weingart et al., 1999). Various negotiation models also establish the importance of trust in interpersonal outcomes during negotiations (Olekalns & Smith, 2005). In dynamic network theory (Westaby et al., 2014), trust is typically presumed to be based on situations where individuals are (or will be) engaged in partner goal striving and system supporting as needed without unwelcomed or inordinate levels of goal prevention, supportive resistance, or system negation.
Ironically, from an intergroup perspective, some displays of negation-prevention behavior may be highly motivational when displayed in appropriate contexts, such as a leader displaying his or her anger with a rival during an intense interaction because of the rival’s unethical behavior that is hurting his or her group’s reputation. Such goal prevention may also garner strong supportive resistance from the boss’s ingroup and hence build group commitment. These examples illustrate the potential motivational effects of negation-prevention in promoting ingroup bonding, but it would likely come at the cost of intergroup conflict and prejudice with outgroup members (Tajfel & Turner, 1986). Thus, conflict would remain in the broader system resulting from the interaction.
Another advantage of our integrative framework is that it can provide new opportunities to examine the dynamics associated with impression management, acting, and deception in real time group interactions. Past work on surface and deep acting, for example, has often relied on self-reported survey techniques. To illustrate a new approach to retrospective assessment, which we refer to as covert analysis in the context of our framework, immediately after a group, negotiation, or dyadic interaction is videotaped, researchers could ask each participant to privately and confidentially review the video and respond to various instructions to access and self-report one’s assumed true and nondistorted psychological orientation, to the extent that this is possible. This builds from past methods that have also used or discussed retrospective or concurrent assessments of psychological processes (e.g., Barber & Roehling, 1993; Isenberg, 1986; Olekalns & Smith, 2005; Vallacher, Nowak, & Kaufman, 1994), but within a covert domain. It also builds on Kagan and Kagan’s (1991) interpersonal process recall approach by providing detailed metrics about the discrepancy between demonstrated behavior and truly felt sentiments during group or social interactions. Such data may also be beneficial in applied feedback sessions to groups or individuals. However, one must be mindful about the debate concerning individuals’ abilities to accurately account for their mental processes (Nisbett & Ross, 1980), when interpreting the data.
Various new methods could be developed in this vein. As one possibility, participants could indicate the exact places where they perceived that they were, in any manner, (a) distorting, acting, faking, or trying to deceive another (e.g., sins of commission), or (b) inwardly suppressing their true feelings, desired behaviors, or pertinent information (e.g., sins of omission). 12 Researchers could then document the types of behavioral elements or units that are being distorted and examine their effect on various dependent measures. For example, are conversations with high levels of false system support perceived to be more constructive than conversations with no false system support? Alternatively, do individuals that overly show system support more likely to be exploited by others, given their agreeable nature?
As another method, in those exact places where self-distortions are reported, researchers could ask participants to stop the video and indicate what they were really thinking or what they would have liked to have said or done instead, if there were no negative consequences. By video recording these reactions, researchers could then code this additional verbal and nonverbal behavior for the presumed underlying true behavioral elements and examine the consequence of those distortions. For instance, do individuals that chronically suppress goal prevention or negation-prevention more likely to have negative intrapsychological effects over time, given the potential for emotional labor (Hülsheger et al., 2010) and/or could these self-regulated behaviors help ensure that groups function without conflict, which in turn promotes well-being for the group and its members in the long run? And, to what extent do group members accurately detect others’ distortion or manipulation of behavioral elements in their interactions? If detected, how does this knowledge impact the group process? Such questions address important issues related to an individual’s dynamic network intelligence in dynamic network theory (Westaby et al., 2014), which represents how accurately people know one another’s network role behaviors. Future research should also examine how the detection of lying behavior through micro expressions of facial behavior (Ekman, 1993) or other physiological or neurological indicators further inform our understanding of distortion and deception processes. To illustrate, researchers may be able to find situations where individuals are reporting no distortion, but physiological, neurological, or micro expression data may indicate that a distortion has occurred. This would imply that individuals are unaware of (or unwilling to report) their distortion behavior. 13
Example
Many researchers have analyzed social interactions in terms of the salient events manifest in dyadic and group settings. Scholars have used terms such as sense units, meaning units, and momentary events (Altman, 1974; Bakeman, 2000; Bakeman et al., 2005; Sacket, 1978) to denote the salient acts coded during group and social interactions. Our behavioral unit concept generally subsumes these terms, but has the added advantage of portraying multiplex behavioral units that contain a combination of behavioral elements, grounded upon a finite set of underlying parameters in dynamic network theory. For example, while some behavioral units can be quite simple, such as a head nod to show system support after another’s behavior, other behavioral units may be multiplex in nature, such as a person stating a phrase while smiling at another at the same time (i.e., a goal striving and system support multiplex unit). In terms of goal striving, the words and, because, if, which, when, and who often represent key separation cues in sentence structure to denote that new behavioral units are being conveyed in the interaction. 14
In the following example, a group of four individuals (i.e., Persons 1, 2, 3, and 4) are interacting and talking about their day and the possibility of one group member moving out of the city. In the following syntax, which uses our extensive frequency method discussed further next, the first number indicates which person is behaving and the subsequent letters indicate his or her behavior. Of course, the duration of each behavioral element could also be included and examined in code-unit grids (Bakeman & Quera, 2012), if that type of temporal specificity is desired. In this example, we identify each of the behavioral elements occurring during the group interaction. Italicized text illustrates the relevant nonverbal behaviors.
“
“
“
“
First long pause. Group members are looking around the room. First time interval from the start of the social interaction to the end of this pause equals 2 minutes. (See Appendix for duration indicators.)
“
Second long pause and group members separate. All in-person interaction roles have ended at this time. Second elapsing time interval since the first time interval equals 2 minutes. This is the end of the group’s interaction (4 minutes total).
To illustrate climate fragments and summations, two independent judges could each perceptually rate the cohesion in the group for each of the 2-minute intervals and the overall interaction summation, respectively. In this example, the judges may both rate the first 2-minute cohesive climate fragment as strong (e.g., 4 on a 5-point scale ranging from “Very weak” to “Very strong”). In contrast, they may both rate the second cohesive climate fragment as very weak (e.g., 1 on the 5-point scale). In this example, one can see that when communicative goal striving and system supporting were elevated for a given time interval, so too was the rating of the cohesive climate fragment. When goal prevention and system negation became prominent, the rating of the cohesive climate fragment was low. Hence, the higher level climate is presumed to be based on the underlying behavioral elements. Lastly, the new judges may rate the overall cohesive climate summation of the entire interaction as rather poor (e.g., 2 on a 5-point scale). This, too, may imply that the low cohesive climate results, in part, from the manifestation of the goal prevention and system negation, especially as they appeared toward the end of the interaction. Future research will need to identify such effects and statistical interactions over time using multilevel and time-series analyses. 17
Capturing the full interaction
Various methods can capture the complexity of group and social interactions over time. We propose an extensive frequency method as discussed earlier, followed by a salient frequency method. Advantages and disadvantages of the two approaches are discussed.
Extensive frequency method
Consistent with research utilizing frequencies to represent behavior (Olekalns & Smith, 2005), the extensive frequency method shows all of the behavioral elements (and units) manifest during the interaction for all individuals over time. The syntax for this method differentiates all group members and their behaviors, including multiplex ones. In our framework, underscore symbol separators (or spaces) are used to separate each behavioral unit for the members, indicating a shift in time between behaviors. 18 The initial number indicates the individual in the group and the following letter (or letters) denotes the behavioral unit (or multiplex behavioral unit) of the member(s). To illustrate in the hypothetical previous example, the interaction would be portrayed as follows:
1g_2o_2o_1g_1gs_1gs_2s3s_1g_1g_1g_1g_1g_1g_1g
(First 2-minute interval)
2p_2p_2c_2p_1np_1np_4r
(Second 2-minute interval)
There are several advantages of this method. First, behavioral element frequencies can be calculated for each time interval and used to predict emergent fragment states or climate fragments for those periods (or other variables of interest). Second, because individuals are differentiated, researchers can examine how each individual may be influencing important outcomes in the system or what more general roles individuals may be taking up in the group (Noumair, 2013), which also accounts for the importance of individual variation in Kenny’s (1994) model of interpersonal relations. In the vein of dynamical systems theorizing, differential equations could also be derived for the individuals (Gottman et al., 2002). Further, individual-level feedback could be provided to individuals in applied work, such as in coaching and counseling. Third, it can examine transitional frequencies and probabilities between individuals. For example, a researcher’s analysis may show how Person 1’s communicative goal striving in a group elicits Person 2’s system support in 34% of the instances of Person 1’s goal striving. See Bakeman and Quera (2012) for other statistical representations of sequential data and behavioral contingencies, such as using the odds ratio and Yule’s Q. Fourth, researchers can account for simultaneous responses from multiple group members. For example, Team Member 2 and Team Member 3 simultaneously showed system support to Person 1 in the previous example. This was represented as “2s3s” with no underscore (or space) between the indicators to show that the event occurred simultaneously in time. A limitation of the extensive frequency method is that it can become extremely cumbersome when analyzing large group behavior. The salient frequency method provides an alternative.
Salient frequency method
The salient frequency method assesses the frequency of behavioral elements occurring during group and social interactions while not differentiating individuals in the analysis for data efficiency reasons. Such a representation may be particularly helpful in research primarily focused on predicting emergent states and climates in groups. The following illustrates the syntax used to describe the previous example:
g_o_o_g_gs_gs_s2_g_g_g_g_g_g_g
(First 2-minute interval) 19
p_p_c_p_np_np_r
(Second 2-minute interval)
In contrast to the previous method, the behavioral units of group members in this approach are shown first and the identifiers for each individual are eliminated, but numbers are used to denote the number of individuals manifesting the behavior at the same time. This approach allows group researchers to easily account for data in large groups that are simultaneously responding in their behavior. To illustrate, a stadium of 1,000 people clapping in support of a leader’s comment can be very easily coded as “s1000.” In comparison, the extensive frequency method would require researchers to indicate that 1,000 different individuals demonstrated system support (e.g., 1s2s3s4s5s6s7s8s9s10s11s12s…1000s). The salient frequency method can further account for group members simultaneously engaged in communicative goal striving in large groups, such as a group of 80 members simultaneously chanting a group pledge of four phrases (i.e., g80_g80_ g80_g80). This syntax can also easily address mixed behaviors in large groups. To illustrate, “g_s12np4” could indicate a situation where one group member had one element of communicative goal striving followed by 12 members of the group simultaneously showing approving head nods, while four other group member are shaking their heads in sharp disagreement with hostile expression on their faces at the same time. As for traditional data representations, Appendix A illustrates how these behavioral element data (i.e., independent variables) across these time intervals could be transposed into a data matrix for predicting emergent fragments and summation states or climates (i.e., dependent variables). 20 Such data could also be aggregated or averaged to present feedback to groups about their observed behavior during interactions, including the presentation of various ratios, such as a positive focus ratio (i.e., [∑g + ∑s] / [∑g + ∑s + ∑p + ∑v + ∑n + ∑r]), which represents the degree to which a group or social interaction was void of any conflict or negativity. See Westaby (2012) for additional ratio conceptualizations. As for disadvantages of the salient frequency method, the simplicity of the approach comes at the cost of not differentiating entities. Hence, a scholar’s research focus should dictate the appropriate approach.
Focusing on cycles alone
Another advantage of our integrative framework is that researchers can decide to exclusively focus on the duration of select behavioral cycles of interest instead of analyzing all behavioral elements, when helpful to the research focus. Although this strategy offers less precision, it can simplify analyses and allow researchers interested in behavioral cycles an alternative and relatively easy approach. For example, researchers may be interested in examining how an experimental manipulation impacts the observed support cycles occurring in a group interaction. In contrast, another researcher may want to compare how the duration of the following cycles predict relevant climate fragments and summations in a group interaction: goal striving cycles (gc), support cycles (sc), basic conflict cycles (bcc), heated conflict cycles (hcc), and questioning cycles (qc). The Appendix illustrates how data can be structured for a salient duration method of cycles for the example provided earlier with hypothetical time durations as follows: 21
gc (3 sec)_qc (2 sec)_gc (2 sec)_sc (8 sec)_gc (16 sec)
(Behaviors in first 2-minute interval)
bcc(6 sec)_hcc(7 sec)
(Behaviors in second 2-minute interval)
Theoretically, one could hypothesize that communicative goal striving and support cycles will have positive effects on emergent states and climates, while basic conflict cycles will have negative effects, and heated conflict cycles will have the most negative effects. Researchers could examine further deconstructions as well, if relevant to the research and setting, such as distinguishing comfort and submissive support cycles that can be embedded within the conflict cycles. 22 Applied researchers could also aggregate or average such data to provide feedback during coaching, counseling, or feedback sessions.
Conclusion
The observation of group and social interactions is often dynamic and complex. Surprisingly, many observational approaches have been either overly simplistic in portraying such interactions or highly idiosyncratic in nature. Hence, this study introduced an integrative framework to capture the complex patterns of behavioral elements and units involved in dynamic group and social interactions across contexts and over time. Given the importance of integrative research across boundaries (Carlile, 2004), our framework is not only grounded in the psychological and network sciences, it also integrates important concepts from group, negotiation, and communication theorizing. This should also allow the framework to examine the significant behavioral elements influencing important emergent concepts and climates in a wide range of contexts, such as examining the influential elements that promote cohesive climates, emergent cultural dimensions, and productive interactions among participants in various contexts (e.g., among supervisors/subordinates, coaches/coachees, counselors/clients, and couples).
Although we advanced a host of new integrative propositions in this paper, more theory building that further integrates higher-level group and organizational theories (Mathieu & Chen, 2011) with microlevel social interaction models would be fruitful. When doing so, it is important for researchers to be mindful of some key theoretical distinctions. For example, the current framework focuses on micro social exchanges where people often engage in free-ranging communications, verbal and nonverbal, oral and written, on social media or in person. We view this as the interaction system. However, at a wider scope, many observed group interactions are embedded within broader dynamic network systems where the individuals are working on goals in groups associated with other focal tasks or goals, such as working with equipment or striving for task objectives (Westaby, 2012; Westaby et al., 2014). 23 Theoretical work interfacing these systems is needed in addition to the covert system that regulates hidden information about individuals’ true psychological orientations.
The current framework initially focused on examining the behavioral elements generally observed between individuals in an immediate interaction. Future research is needed to extend the framework to model the behavioral elements directed at external parties not directly participating in the interaction. For example, a person may be engaged in communicative goal striving with another, but the person is talking very negatively about another person or group who is not present. In this example, the person’s communicative goal striving in the immediate conversation could be further subcoded as negation-prevention directed toward a third party not present in the current discussion. Additional coding of the direction of these elements would be required, which would need to be assessed for reliability as well. Likewise, research needs to examine the utility and reliability of coding intensity levels for the behavioral elements, such as at two levels (e.g., normal vs. intense or extreme levels). Future theory and research also needs to examine the implicit meaning of nondisplayed behavioral elements, such as when a person makes a provocative claim during goal striving and no one disagrees with the claim (i.e., no goal prevention), which some may believe implies support in the group, despite no explicit system support behavior (e.g., no head nods, no statements of agreement). The use of covert analysis of hidden process, as described earlier, may reveal that disagreement did exist, but members controlled their true sentiments for various reasons, which should also be examined empirically.
As for other areas of applied research, there are numerous opportunities beyond those discussed earlier. For example, future research could use parameters in the framework in conjunction with computer analytic programs (e.g., word/phrase counting systems) to make assumptions about the social dynamics occurring in large digital platforms, email systems, or social media websites (Pennebaker et al., 2003). 24 To illustrate, such work could analyze the salient behavioral elements, units, and patterns occurring in different systems, such as identifying high instances of negation-prevention wording in various organizations or social media websites and their impact. Another promising area of application in organizational psychology, negotiation, and communication research is to use the parameters in the approach to assess group and negotiation dynamics, and, in turn, provide confidential and appropriate feedback about observed processes to individuals or to the group overall. Moreover, interventions could be leveraged that attempt to motivate individuals to self-regulate, control, or functionally suppress behavioral elements, if possible, that are predictive of unhealthy interaction states or climates (and promote ones that are predictive of healthy ones). Such interventions would be aimed at improving relationships and group dynamics. This work would build upon notions by Gottman et al. (2002), but using behavioral elements and units in our framework.
In terms of formal evaluations, the framework could be used in assessment centers to evaluate the interaction styles of participants in interviews, role plays, group interaction events, or from previously available recordings (or transcripts) as well as to examine which model parameters best predict emergent fragment and summation states concerning candidate performance. In such cases, the analyses could also be simplified into an egocentric format where assessments would be conducted on only the target individual for feedback reporting purposes. Furthermore, triangulating the observational data with multirater feedback measures would also be an important line of inquiry to provide additional insight into human behavior.
In conclusion, social and organizational scientists have made great strides in the analysis of group and social interactions. However, past frameworks have missed opportunities to explain the fundamental behavioral elements involved in more complex and dynamic group processes occurring over time, grounded in contemporary psychological and network sciences. We presented an integrative framework in this paper as one alternative to help fill this large gap.
Footnotes
Acknowledgement
The authors thank Debra Noumair for her helpful comments on a previous version of this manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Notes
Appendix. Potential representation of behavioral element and climate data
| Part 2: Salient duration method for select behavioral cycles of interest | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Two-minute intervals | Duration of observed behavioral cycles (seconds) | Emergent state or climate ratings on a 5-point cooperation scale | ||||||||
| Goal striving cycle | Support cycle | Basic conflict cycle | Heated conflict cycle | Questioning cycle | Fragment | Summation | ||||
| 0–2 min. | 21 sec. | 8 sec. | 0 sec. | 0 sec. | 2 sec. | 4 | -- | |||
| 2–4 min. | 0 sec. | 0 sec. | 6 sec. | 7 sec. | 0 sec. | 1 | -- | |||
| 2 | ||||||||||
Note. g = goal striving; s = system supporting; p = goal preventing; v = supportive resisting; n = system negating; r = system reacting; o = observational questioning; c = low system competency. Various ratios, such as the positive focus ratio could also be calculated. Although not demonstrated in this Appendix, to represent the extensive frequency method, variables for each behavioral element in Part 1 would be distinguished for each individual. The following assumptions delineate the behavioral elements underlying the configuration of the mutually exclusive cycles in Part 2, which future research should further examine, debate, or refine as needed.
In the Part 2 illustration of mutually exclusive categories, low system competency elements are not differentiated in any of the cycles and would be simply subsumed in the relevant cycles that they are embedded within (e.g., “I was, umm, working from home” would be represented with the duration of the goal striving cycle).