Volatile Psychological Dynamics in Social Interactions

Abstract

Does agreement or disagreement alone strongly influence people’s thoughts and emotions, or does cognitive and affective reaction relate more to shifts in agreement and disagreement? The present study investigated how attitudes and emotions respond to the evolution of different patterns of agreement and disagreement in an interaction. Combining conventional methods with novel dynamic systems techniques, we found that people’s attitudes and emotional valence undergo significant change and become less stable when disagreement replaces agreement in an interaction. However, the same reaction does not occur when agreement replaces disagreement. Findings also reveal that the attitudes and emotions do not respond to disagreement alone, but react to the interaction transitioning from an exchange based on agreement to one defined by disagreement. These results provide insight into how attitudes and emotions evolve as social interactions undergo change and confirm key ideas from dynamic systems theories about social interaction.

Keywords

conflict dynamical systems model attitude change attitudes emotion social interaction interpersonal processes

Agreement is often perceived favorably, while disagreement is perceived unfavorably. For example, the folk wisdom “let’s settle our differences” suggests that disagreement is a source for ill feelings and dispute, while achieving agreement yields positive feelings and problem resolution. Conclusions from many studies and theories fit with the perspective that disagreement, especially with attitudes and beliefs, is experienced negatively and contributes to a sense of dispute (e.g., Bodtker & Jameson, 2001; Boulding, 1962; Marcus, 2006). Furthermore, agreement is commonly associated with positive emotions and favorable interaction outcomes (e.g., Deutsch, 1973, 1994; Heider, 1958; McNeel & Reid, 1975). The capacity for agreement to elicit positive emotions and beneficial social change is discussed in conflict research. For instance, confronting differences and agreeing with an adversary’s perspective is a suggested beginning for dispute resolution (Keating, Pruitt, Eberle, & Mikolic, 1994), especially if the unilateral agreement leads to reciprocity (e.g., Goldstein & Freeman, 1990; Hüffmeier, Freund, Zerres, Backhaus, & Hertel, 2011; Maio & Haddock, 2007; Osgood, 1962). However, other research, case studies, and theories conclude that confronting differences even by emphasizing agreement may fail to yield meaningful interaction change (e.g., Coleman, 2000; Keating et al., 1994; Praszkier, Nowak, & Coleman, 2010). The present research is motivated by these different conclusions about the power of agreement and disagreement in social interactions. Is agreement versus disagreement a central factor in how people psychologically react in an interaction? What happens to thoughts and emotions when disagreement is replaced with agreement and how does this compare when disagreement emerges in an interaction that was formerly based upon mutual agreement?

Addressing these questions, we use a dynamic systems approach to examine how people’s attitudes and emotions respond to agreement, disagreement, and transitions between agreement and disagreement. Foremost, we suggest that changes in an interaction contribute more to psychological response than agreement or disagreement alone. We predict that interaction change from agreement to disagreement (or from disagreement to agreement) has greater influence on attitudes and emotion states than consistent agreement or disagreement. Furthermore, we propose that a person’s attitudes and emotions exhibit disproportionate reaction when an individual encounters newly developed disagreement rather than agreement in an interaction. This disproportionate response, we expect, will relate to people experiencing more intense negative emotion in cases of new and sudden disagreement and less intense positive emotion in cases of sudden agreement. Before detailing these predictions, we first summarize the dynamic systems approach that guides our predictions.

Overview of the Dynamic Systems Approach

The dynamic systems perspective is useful to understand how patterns of thought, emotion, and behavior change over time. The fixed-point attractor, a stable state that limits change, is the dynamic systems concept most often used in psychology (e.g., Guastello, Koopmans, & Pincus, 2009; Vallacher, Coleman, Nowak, & Bui-Wrzosinska, 2010a, 2010b). Fixed-point attractors are evident when a phenomenon exhibits little change over time. For example, both personality in adulthood (Roberts, Walton, & Viechtbauer, 2006) and emotion in people with low affect intensity (Larsen, 2009) demonstrate little change over time and illustrate fixed-point attractor properties. Conversely, greater change reflects instability, and may indicate that a phenomenon is temporarily perturbed, transitioning into a new state, or lacks an attractor. People who have considerable variation in affect intensity (Larsen, 2009) exhibit emotion patterns suggesting lack of a fixed-point attractor. When a given psychological state suddenly exhibits change, the instability, called destabilization, reflects that the system is significantly perturbed and may be undergoing important transition. In psychology, destabilization is apparent when someone abandons habit and rigid thought and embraces new ways of thinking, feeling, and behaving (Michaels, Parkin, & Vallacher, 2013; Nowak & Vallacher, 1998).

The dynamic systems concepts of the attractor and destabilization have been used to understand social interaction. For instance, if two people who get along fall into a period of disagreement, they may experience destabilization in their attitudes and emotions as they search for a new way to understand their relationship. Eventually, the people may perceive their interaction more negatively, leading to hostility and an emphasis on disagreement. This frame of thought may become increasingly stable, with those involved expressing rigid disagreement and consistently negative feelings. This emergent, rigid exchange is thought of as an attractor and illustrates a core idea from dynamic systems models of conflict development and persistence (e.g., Coleman, Vallacher, Nowak, & Bui-Wrzosinska, 2007; Kriesberg, 1980; Vallacher et al., 2010a, 2010b). When rigid thought and disagreement come to define an interaction, the pathway to compromise and agreement requires that adversaries abandon rigid ways of thinking and feeling and instead engage one another with greater flexibility (e.g., Bartoli, Bui-Wrzosinska, & Nowak, 2010).

Dynamic systems theories about how processes like attitudes and emotions evolve during interactions are formalized in a model developed by Liebovitch et al. (2008). The key model finding is that attitudes and emotions remain stable when an interaction involves consistent agreement or disagreement, suggesting that agreement or disagreement alone is not a source for significant psychological perturbation. Instead, from model simulations, attitude and emotion perturbation takes place after unilateral shifts from agreement to disagreement or vice versa. These model results follow dynamic systems theory and broader literature, suggesting that attitude and emotion destabilization can lead to a cognitive search for new, orderly ways to frame the interaction (cf. Michaels et al., in press; Tullett, Teper, & Inzlicht, 2011; Vallacher & Wegner, 1985, 2012). In a case where two people are mired in disagreement, a cognitive shift may facilitate new, or renewed, agreement.

While the dynamic systems theories and Liebovitch’s et al. (2008) model offer insight about attitude and emotion evolution during social interactions, there remains a need for further research. Coleman (2000) emphasizes a need for empirical studies examining how psychological processes respond (if at all) to interaction transitions. Furthermore, there is a need for research investigating emotion evolution in cases of disagreement versus agreement (cf. Kozan, 1997; Nair, 2007). Guided by the dynamic systems theory and a mathematical model, the present research addresses these needs.

The Present Research

The present study is based on two core predictions. First, during a dyadic interaction, shifts in expressed agreement and disagreement more strongly influence attitudes and emotions than consistent agreement or disagreement alone. This prediction follows from dynamic systems theory and simulation results which suggest that interaction changes destabilize psychological processes. Second, we predict that attitudes and emotions become more destabilized when an interaction shifts from expressed agreement to expressed disagreement rather than from disagreement to agreement. We further predict that the transition from agreement to disagreement is experienced as strongly negative, whereas a shift from agreement to disagreement elicits only mildly positive emotion. Both previous predictions are based on dynamic systems theory and broader research that demonstrate people have stronger reaction to negative than to positive stimuli. For example, people are more sensitive to negative than positive information in relationships (e.g., Gottman, Murray, Swanson, Tyson, & Swanson, 2005). From game theory, people react stronger to hurtful than to helpful actions (Offerman, 2002). Evidence suggests that such asymmetric reaction may relate to physiological response, as there is faster neural response to negative than to positive stimuli (Smith, Cacioppo, Larsen, & Chartrand, 2003), and numerous studies have demonstrated people exhibit stronger physiological reaction to disagreement versus agreement (for review, see Taylor, 1991). Finding such asymmetric response to disagreement versus agreement would provide evidence supporting dynamic systems theory about how destabilization is a primary factor in the development of persistent disagreements and disputes (e.g., Bartoli et al., 2010; Liebovitch, Vallacher, & Michaels, 2010; Vallacher et al., 2010a, 2010b). Since stability and destabilization are fundamental to the dynamic systems approach, we used a novel method for measuring and analyzing participants’ emotions. Our goal was to understand whether participants’ emotions became more or less stable and how this stabilization or destabilization relates to the experience of positive versus negative emotions (valence).

Inspired by Vallacher, Nowak, Froehlich, and Rockoff (2002), our method had participants move a computer cursor to indicate positive versus negative valence. Presented with a computer screen featuring a black background with seven equidistant white lines, participants were instructed to move the cursor toward the right edge of the monitor to indicate more positive valence and toward the left edge to indicate more negative valence. As participants moved the cursor, the cursor’s pixel location was recorded every 10 ms. This recording provided a time series of the pixel coordinates, with x-axis coordinates corresponding to participants’ self-report of positive versus negative feelings. We used two novel techniques to analyze the valence time-series data to discern whether participants’ valence stabilized or destabilized over time.

Dynamic Analysis Method 1: Convex Hull

The convex hull is the largest polygon surrounding a set of data points (see Online Supplemental Appendix found at http://spps.sagepub.com/supplemental). In the present study, the set of points are cursor locations reflecting a participant’s moment-to-moment valence. If a participant expressed a broader and more changing range of valence, the convex hull area would be larger. Comparatively, expression of a more limited and less changing range of valence would present a smaller convex hull area. Accordingly, a convex hull area that increases over time suggests destabilization, while a shrinking convex hull area suggests stabilization and potential attractor formation.

We anticipate that the asymmetric psychological response to disagreement versus agreement will influence participants’ valence dynamics. Specifically, we expect more negative valence will be associated with greater valence destabilization, and that this destabilization will correspond to a larger convex hull area as well as a more complex correlogram pattern (discussed below).

Dynamic Analysis Method 2: Correlograms

We also assessed valence dynamics with correlograms, which plot how time-series data are correlated across different time scales (see Online Supplemental Appendix found at http://spps.sagepub.com/supplemental). Simple correlogram patterns are associated with highly stable, attractor dynamics, whereas complex patterns suggest less stability (cf. Cryer & Chan, 2008; Vallacher & Nowak, 1994). Examples of different correlograms are provided in the Online Supplemental Appendix found at http://spps.sagepub.com/supplemental. Following our prior predictions, we expect to find that more negative emotional states will be associated with greater destabilization and that destabilization will be revealed through larger convex hull area and more complex correlogram patterns.

Summary of Predictions

Having discussed the dynamic systems approach and methods, we briefly recap our predictions. First, we anticipate that people’s attitudes and valence exhibit the greatest change in interactions involving a shift from agreement to disagreement, or vice versa, but that little attitude or valence change will take place in response to agreement or disagreement alone. This prediction aligns well dynamic systems theories of human interaction, especially conflict. Second, we expect that people exhibit the greatest attitude and valence destabilization in scenarios where agreement is replaced by disagreement. This prediction follows evidence of how negative information, such as disagreement, is more impactful than positive information, such as agreement. This differential psychological response to disagreement versus agreement is expected to be reflected in participants’ exhibiting significantly more negative valence when agreement is replaced by disagreement, but little valence response when disagreement is replaced by agreement.

Method

Participants

Eighty-four undergraduates (33 males, 51 females) from a large public university completed all parts of this study in exchange for course credit.

Procedure

Participants completed a two-part study. The first session was used to identify participants’ baseline attitudes about 40 statements using a survey with a 6-point scale ranging from 1 = strongly disagree to 6 = strongly agree. Statements ranged from benign (e.g., “Coke is better than Pepsi”) to controversial (e.g., “Nuclear weapons are necessary to protect our nation”). We used the 6-point scale to avoid an ambiguous response that would hinder identification of whether a participant agreed or disagreed with a given item, as identifying whether a participant agreed/disagreed with each item was important for our interaction procedure, detailed below. After finishing the survey, participants selected a time to return to the lab (approximately 1 week later) when they were randomly assigned to one of the five conditions, as detailed next.

Interaction Session Procedure

Participants’ baseline survey responses were used to prepare a script that informed confederates how to verbally respond to each participant during a controlled interaction. This controlled response was used to set up different patterns of agreement or disagreement. A sense of agreement was evoked by having confederates express attitudes congruent with the participant’s (agreeing with what the participant agreed with and disagreeing with what he or she disagreed with). In contrast, a sense of disagreement was promoted by a confederate expressing attitude incongruence (e.g., disagreeing to anything a participant agreed with).

Participants were randomly assigned to one of the five conditions that differed only in the pattern of confederate’s responses during the exchange. These conditions included constant agreement, constant disagreement, agreement switching to disagreement halfway through the exchange, disagreement switching to agreement halfway through the exchange, or a control case involving no expressed agreement or disagreement (participants were told to simply take turns reading the statements aloud with the confederate).

For the interaction, participants were told that they would take turns reading a set of statements with another person (the confederate). All statements were taken from the first survey. Participants were instructed to verbally indicate whether they agreed or disagreed with each statement and to listen to whether the confederate agreed or disagreed. Meanwhile, participants completed a survey using a 6-point Likert-type scale to indicate their level of agreement with each statement. Participants and confederates only expressed agreement/disagreement; no information about the Likert-type scale response was ever exchanged. Importantly, participants and confederates interacted remotely using only an audio signal to eliminate the possibility of influence from body language, facial expressions, or physical appearance.

Finally, participants were recorded during their interaction so that they could view the recording while completing our dynamic cursor task to indicate retrospective valence (described below). After completing this task, participants were debriefed, revealed the study’s purpose, and awarded course credit.

Dependent Variables

There were two dependent variables. Attitude stability was measured by computing how much a participant changed responses between the baseline attitude survey and the interaction session survey. We computed the absolute value of the difference in Likert-type scale responses for identical items on the two surveys. The absolute value was used as we were interested only in the magnitude of attitude change. After computing the absolute attitude change, we summed the amount of change for items completed during the first half of the interaction and for items completed during the second half of the interaction. Finally, we took the difference of these sums (sums of the change). This difference would reveal whether people’s attitudes became more stable (negative values) or more unstable (positive values). Computational details are provided in Online Supplemental Appendix, Section 1 found at http://spps.sagepub.com/supplemental.

A measure similar to the attitude stability was used for assessing participants’ valence stability. Using a method similar to one by Vallacher et al. (2002), each horizontal cursor position coordinate was transformed into a dichotomous variable. Any single cursor position within the left half of the monitor was coded as “−1” to indicate negative valence, whereas any position within the right half of the monitor was coded as “+1” to indicate positive valence. Summing these recoded cursor positions indicated whether a participant expressed more positive valence (sum > 0) or more negative valence (sum < 0) during the first half and during the second half of the interaction. Like the attitude stability measure, we took the difference of these valence sums. The difference allowed us to identify whether a participant moved the cursor to express more frequent positive valence (difference > 0) or more frequent negative valence (difference < 0) during the second half of the interaction. Finally, each participant’s data were normalized by dividing the difference by his or her total time of interaction. This was necessary as each session’s duration varied due to individual differences in reading and response speed.

The previously described transform was used for several reasons. First, each participant may have a different understanding of how positive versus how negative each area of the computer screen may be since the display is intentionally ambiguous to facilitate more fluid responding. Second, participants’ cursor movements vary in magnitude making it difficult to make comparisons using summary statistics. Transforming the data helps limit data variability allowing for more meaningful statistical comparisons. Details on the valence transform are provided in Online Supplemental Appendix, Section 2 found at http://spps.sagepub.com/supplemental.

Results

Attitude Stability

In line with predictions, participants’ attitudes exhibited significant change when the interaction began with the confederate agreeing before switching to disagreeing, F(4, 79) = 3.993, p = .005 (η _p ² = .168; retrospective power = .892). Multiple pairwise comparisons using Tukey’s honestly significant difference revealed that participants who encountered an interaction transitioning from agreement to disagreement had significantly greater attitude change (M = 6.8, SD = 9.5) than the control (M = −5.8, SD = 13.1, p = .008), constant disagreement (M = −3.6, SD = 8.0, p =.041), and disagreement switching to agreement group (M = −4.5, SD = 12.8, p = .017). There was, however, no significant difference between participants who experienced initial agreement that transitioned to disagreement versus those who experienced constant agreement (M = −1.7, SD = 9.1, p = .129). No other significant differences were found to exist between any of the groups (Table 1).

Table 1.

Attitude Stability Descriptive Statistics and Mean Differences for All Groups.

Condition		M	SD	Mean Differences
Condition		M	SD	M ₁	M ₂	M ₃	M ₄	M ₅
M ₁	Control	−5.8	13.1	0	4.1	2.2	12.6**	1.3
M ₂	Agreement	−1.7	9.1		0	−1.9	8.5	−2.8
M ₃	Disagreement	−3.6	8.0			0	10.4*	−0.9
M ₄	Agree → Disagree	6.8	9.5				0	−11.3*
M ₅	Disagree → Agree	−4.5	12.8					0

Note. *p < .05. **p < .01.

These results (Figure 1) are supportive of our general prediction that people’s attitudes exhibit the greatest change when an interaction transitions from agreement to disagreement. While there was no significant difference found between the group experiencing this transition and the group encountering a confederate who interacted with continual agreement, the mean attitude change for this latter group was still below 0. From our computational method, an average below 0 indicates greater attitude stability over the duration of the interaction. The only group to exhibit less stable attitudes (M > 0) was the group experiencing a switch from agreement to disagreement. Finally, it is worth noting that there was no significant difference in attitude reaction to constant disagreement versus constant agreement. This null finding is supportive of our suggestion that interaction change is a driving factor in cognitive reaction.

Figure 1.

Mean attitude change based on interaction type with ±2 SEM indicated. Positive numbers indicate greater attitude change over time.

Valence Stability

We had mixed support for our predictions about valence. Results from a single factor Welch’s analysis of variance reveal that participants’ valence changed in significantly different ways depending on interaction type, F _Welch(4, 38.565 = 2.839, p = .037). Pairwise comparisons were made with the Games-Howell test. The Welch’s and Games-Howell procedures were used as the data exhibited heterogeneous variance (see Todman & Dugard, 2007). Participants experiencing a shift from agreement to disagreement expressed more negative valence (M = −0.139, SD = 0.299) than did participants who encountered a control case (M = 0.109, SD = 0.148, p = .029) or an exchange transitioning from dissimilarity to similarity (M = 0.099, SD = 0.120, p = .029), as evident in Figure 2. However, no significant differences were found between groups experiencing a switch from agreement to disagreement and groups involving constant agreement (M = 0.055, SD = 0.167, p = .136) or constant disagreement (M = 0.024, SD = 0.262, p = .437). There were no other significant valence differences (Table 2).

Figure 2.

Mean valence change based on interaction type with ±2 standard error of mean indicated. Positive numbers indicate more positive valence while negative numbers indicate more negative valence over the course of the interaction.

Table 2.

Valence Stability Descriptive Statistics and Mean Differences for All Groups.

Condition		M	SD	Mean Differences
Condition		M	SD	M ₁	M ₂	M ₃	M ₄	M ₅
M ₁	Control	.109	.148	0	−.054	−.085	−.248*	−.010
M ₂	Agreement	.055	.167		0	−.031	−.194	.044
M ₃	Disagreement	.024	.262			0	−.163	.075
M ₄	Agree → Disagree	−.139	.299				0	.238*
M ₅	Disagree → Agree	.099	.120					0

Note.*p < .05.

These results partially confirm our hypothesis that an interaction shift from agreement to disagreement evokes more negative emotions. Compared with a scenario where an exchange transitions from disagreement to agreement, the shift from agreement to disagreement is experienced with markedly negative emotions. However, this same transition does not evoke emotions that are especially more negative than an encounter involving disagreement alone or, interestingly, agreement alone. Before exploring possible explanations for these findings, we first examined the valence data using our dynamic systems methods.

Dynamical Analysis and Results

We analyzed participants’ valence dynamics by creating the data phase space, which is used to identify the presence or absence of attractors (see Liebovitch, 1998; Strogatz, 1994). To create the phase space, we paired each cursor location (reflecting positive vs. negative valence) with the associated amount of cursor movement (reflecting rate of change) and finally plotted the data pairs in a standard two-dimensional space (details and further explanation in Online Supplemental Appendix, Section 3 found at http://spps.sagepub.com/supplemental). After generating the plots, we used MATLAB’s convex hull algorithm to determine the phase space area during the first and second half of the interaction (details in Online Supplemental Appendix, Section 4 found at http://spps.sagepub.com/supplemental). An area that expanded from the first to the second half of the interaction suggests valence destabilization, whereas an area that contracted suggests stabilization since data convergence is an attractor property.

Next, we generated correlograms for each participant’s valence data. Five trained raters ranked each correlogram in order of pattern complexity, from 0 = simple to 3 = highly complex and erratic, Inter-rater reliability with: Cronbach’s α = .89 (see Online Supplemental Appendix, Section 5 found at http://spps.sagepub.com/supplemental for correlogram examples). Before continuing analysis in light of our predictions, we first verified the relationship between the correlogram ratings and convex hull area, as a larger area should correlate with a more complex correlogram if these methods reveal destabilization. A Spearman correlation verified this expected correlation between the convex hull area and correlogram complexity during both the first (rs = .348, p = .001) and second parts of the interaction (rs = .458, p < .001). Participants with more complex valence patterns tended to express a greater valence range, verifying that correlogram complexity identifies destabilization versus stabilization.

We next wanted to understand whether participants’ valence dynamics were associated with a change to more positive or more negative emotions during the interaction. To assess this, we grouped participants into categories based on their correlogram complexity: low complexity suggestive of stabilization or high complexity suggesting destabilization (examples in Online Supplemental Appendix, Section 5 found at http://spps.sagepub.com/supplemental). An independent t-test revealed a significant difference in valence change, t(82) = 2.243, p = .028 based on these groups. Participants whose valence exhibited a complex correlogram and implied destabilization tended to become more negative (M = −0.0894, SD = 0.0413), whereas those whose valence correlogram pattern suggesting stabilization tended to become more positive (M = 0.0501, SD = 0.0285) as shown in Figure 3. This finding from our dynamic systems based methods provides support for our prediction that valence destabilization is experienced negatively.

Figure 3.

Mean valence change based on whether participants’ valence patterns exhibited stability or instability, with ±2 standard error of mean indicated.

Finally, we used regression to understand the relationships between valence, dynamic patterns, and attitude change. The final model from stepwise linear regressions reveals that valence during the second half of the interaction is predicted by a combination of dynamic patterns and the amount of attitude change during the first half of the interaction. The model, F(3, 83) = 7.836, p = .001, R ² = .227, included whether participants’ valence time series demonstrated a simple correlogram pattern, suggesting stabilization (β = 856.829, p = .024), lack of correlogram pattern, suggesting destabilization (β = −992.259, p = .037), and the amount of attitude change during the first half of the interaction (β = −33.336, p = .048). Two variables were used for simple versus complex correlogram pattern as raters used a 4-point scale to categorize the correlograms. To include the correlogram patterns in the regression, we dummy-coded each of the four levels of complexity and included only the significant categories in the final model.

This regression model unifies our main findings and confirms our hypotheses. A person’s valence relates to emotional and cognitive dynamics. Negative emotions arise in conjunction with destabilization in both attitudes and valence, and greater destabilization emerges from interaction transition from agreement to disagreement.

Discussion

The present study combined conventional experimental methods and statistics with novel dynamic systems inspired techniques to investigate how people’s attitudes and emotions evolve during the course of an interaction. Consistent with the hypotheses, people’s attitudes and valence exhibited the greatest change and destabilization when an interaction transitioned from a period of agreement to disagreement. In contrast, little change emerged in attitudes or valence when agreement followed a period of disagreement. Likewise, our study found no evidence that agreement or disagreement alone was particularly perturbing. Instead, dynamic shifts in agreement or disagreement between two people more strongly influenced the evolution of attitudes and emotions, especially when an interaction progressed from agreement to sudden unilateral disagreement. Our results demonstrate that people respond asymmetrically to disagreement versus agreement and provide insight into how attitudes and emotions evolve during interaction transitions.

Using new methods, the present research found that attitude and emotional destabilization are associated with negative valence. Furthermore, we identified that attitude and valence stabilization predicts more positive valence, whereas destabilization predicts more negative valence. These results fit with dynamic systems accounts that characterize stabilization as facilitating coherence while destabilization reflects diminished coherence. Coherence is a key component of theories pertaining to how people form attitudes, generate meaning, and maintain beliefs (e.g., Festinger, 1957; Heider, 1958; Tullett et al., 2011). When people encounter information that is inconsistent with their own attitudes, they experience negative emotions related to uncertainty and anxiety (Harmon-Jones, 2000; Nowak & Vallacher, 1998), and this negativity is manifest in a breakdown of cognitive coherence (Vallacher, Nowak, Markus, & Strauss, 1998).

Construal (action identification) may explain the relation between instability and negative valence. From action identification theory, people seek to reframe the meaning of their actions as they integrate actions into a coherent whole (Vallacher & Wegner, 1985, 2012). Participants exposed to a partner who shifted from agreeing to disagreeing experienced a disruption in their attitude coherence, which may have facilitated a cognitive shift. This shift would reflect participants’ search for new attitude coherence. Valence instability would be expected to accompany participants’ search for new ways of integrating their thoughts and emotions (cf. Nowak & Vallacher, 1998, pp. 174–175). This interpretation, while speculative, warrants investigation.

In contrast to the attitude and valence reaction found when agreement is interrupted by disagreement, the introduction of agreement after disagreement yielded little response. While some research suggests a response should emerge whenever an interaction shift takes place (e.g., Liebovitch et al., 2008), we predicted that agreement replacing disagreement would facilitate limited attitude and valence change based on the literature identifying that people have stronger psychological reaction to negative than to positive stimuli (for review, see Taylor, 1991). Therefore, our results provide another example of how people exhibit asymmetric response to disagreement and negative information versus agreement and positive information.

Results provided additional insight into psychological dynamics during interactions progressing with different patterns of agreement and disagreement. As predicted, unchanging agreement or unchanging disagreement did not evoke significant attitude or valence change. People who experienced a constantly disagreeing partner expressed valence that was no more negative than those who experienced a constantly agreeing partner, suggesting that disagreement alone is not a primary source for negative emotions in an interaction. However, this finding may reflect that our method’s controlled interaction without a meaningful purpose may have led participants to exhibit minimal emotional reaction. This possibility would explain why we found little difference between groups experiencing a shift from agreement to disagreement during the interaction versus groups involving constant agreement or constant disagreement. In the latter case, participants may have dismissed the confederate once disagreement was quickly established, perhaps characterizing the confederate as an out-group member (e.g., Andersen, Moskowitz, Blair, & Nosek, 2007). Alternatively, the lacking valence response to disagreement alone may reflect that people are simply insensitive to irrelevant negative stimuli (Sweeney, Shepperd, & Carroll, 2009). With these limitations, it is unclear whether people would respond differently to more consequential interactions. Future work may address this by replicating the experiment with a meaningful social interaction.

Conclusion

Settling differences by emphasizing attitude agreement may not be as helpful to dispute resolution as folk wisdom suggests. People experience little attitude or emotion change when disagreement ceases and agreement emerges. Conversely, people experience swift reaction with attitude destabilization and negative, unstable emotions when agreement stops and disagreement begins in an interaction. These findings provide insight into the evolution of attitudes and emotions as agreement and disagreement evolve in a social exchange. It is easy to evoke negativity and rapid cognitive change when disagreement arises in an interaction, but it is difficult to evoke positivity or change once disagreement, however brief, develops in a social situation.

Footnotes

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was partially supported by the James S. McDonnell Foundation, grant number 220020112.

References

Andersen

S. M.

Moskowitz

G. B.

Blair

I. V.

Nosek

B. A.

(2007). Automatic thought. In Kruglanski

A. W.

Higgins

E. T.

(Eds.), Social psychology: Handbook of basic principles (2nd ed., pp. 138–175). New York, NY: Guilford.

Bartoli

Bui-Wrzosinska

Nowak

(2010). Peace is in movement: A dynamical systems perspective on the emergence of peace in Mozambique. Peace and Conflict: The Journal of Peace Psychology, 16, 211–230.

Bodtker

A. M.

Jameson

J. K.

(2001). Emotion in conflict formation and its transformation: Application to organizational conflict management. The International Journal for Conflict Management, 12, 259–275.

Boulding

K. B.

(1962). Conflict and defense. New York, NY: Harper and Row.

Coleman

P. T.

(2000). Intractable conflict. In Deutsch

Coleman

P. T.

(Eds.), The handbook of conflict resolution: Theory and practice (pp. 428–450). San Francisco, CA: Jossey-Bass.

Coleman

P. T.

Vallacher

R. R.

Nowak

Bui-Wrzosinska

(2007). Intractable conflict as an attractor: A dynamical systems approach to conflict escalation and intractability. American Behavioral Scientist, 50, 1454–1475.

Cryer

J. D.

Chan

(2008). Time series analysis: With applications in R. (2nd ed.). New York, NY: Springer.

Deutsch

(1973). The resolution of conflict: Constructive and destructive processes. New Haven, CT: Yale University Press.

Deutsch

(1994). Constructive conflict management for the world today. The International Journal of Conflict Management, 5, 111–129.

10.

Festinger

(1957). A theory of cognitive dissonance. Evanston, IL: Row and Peterson.

11.

Goldstein

J. S.

Freeman

J. R.

(1990). Three-way street: Strategic reciprocity in world politics. Chicago, IL: University of Chicago Press.

12.

Gottman

J. M.

Murray

J. D.

Swanson

C. C.

Tyson

Swanson

K. R.

(2005). The mathematics of marriage: Dynamic nonlinear models. Cambridge, MA: MIT Press.

13.

Guastello

S. J.

Koopmans

Pincus

(Eds.). (2009). Chaos and complexity in psychology: Theory of nonlinear dynamics. New York, NY: Cambridge University Press.

14.

Harmon-Jones

(2000). Cognitive dissonance and experienced negative affect: Evidence that dissonance increases experienced negative affect even in the absence of aversive consequences. Personality and Social Psychology Bulletin, 26, 1490–1501.

15.

Heider

(1958). The psychology of interpersonal relations. New York, NY: John Wiley.

16.

Hüffmeier

Freund

P. A.

Zerres

Backhaus

Hertel

(2011). Being tough or being nice? A meta-analysis on the impact of hard- and softline strategies in distributive negotiations. Journal of Management. doi:10.1177/0149206311423788

17.

Keating

M. E.

Pruitt

D. G.

Eberle

R. A.

Mikolic

J. M.

(1994). Strategic choice in everyday disputes. International Journal of Conflict Management, 5, 143–157.

18.

Kozan

M. K.

(1997). Culture and conflict management: A theoretical framework. The International Journal of Conflict Management, 8, 338–360.

19.

Kriesberg

(1980). Interlocking conflicts in the middle east. Research in Social Movements, Conflicts and Change, 3, 99–119.

20.

Larsen

R. J.

(2009) Affect intensity. In Leary

M. R.

Hoyle

R. H.

(Eds.), Handbook of individual differences in social behavior (pp. 241–254). New York, NY: Guilford.

21.

Liebovitch

L. S.

(1998). Fractals and chaos: Simplified for the life sciences. New York, NY: Oxford University Press.

22.

Liebovitch

L. S.

Naudot

Vallacher

R. R.

Nowak

Bui-Wrosinska

Coleman

P. T.

(2008). The dynamics of two-actor cooperation-competition conflict models. Physica A: Statistical Mechanics and Its Applications, 387, 6360–6378.

23.

Liebovitch

L. S.

Vallacher

R. R.

Michaels

(2010). Dynamics of cooperation-competition interaction models. Peace and Conflict: Journal of Peace Psychology, 16, 175–188.

24.

Maio

G. R.

Haddock

(2007). Attitude change. In Kruglanski

A. W.

Higgins

E. T.

(Eds.), Social psychology: Handbook of basic principles (pp. 565–586). New York, NY: Guilford.

25.

Marcus

E. C.

(2006). Change and conflict: Motivation, resistance, and commitment. In Deutsch

Coleman

P. T.

Marcus

E. C.

(Eds.), The handbook of conflict resolution: Theory and practice (pp. 436–454). Hoboken, NJ: John Wiley.

26.

McNeel

S. P.

Reid

E. C.

(1975). Attitude similarity, social goals, and cooperation. Journal of Conflict Resolution, 19, 665–681.

27.

Michaels

J. L.

Parkin

S. S.

Vallacher

R. R.

(2013). Destiny is in the details: Action identification in the construction and deconstruction of meaning. In Hicks

Routledge

(Eds.), The experience of meaning in life: perspectives from the psychological sciences. New York: Springer.

28.

Nair

(2007). Towards understanding the role of emotions in conflict: A review and future directions. International Journal of Conflict Management, 19, 359–381.

29.

Nowak

Vallacher

R. R.

(1998). Dynamical social psychology. New York, NY: Guilford.

30.

Offerman

(2002). Hurting hurts more than helping helps. European Economic Review, 46, 1423–1437.

31.

Osgood

C. E.

(1962). An alternative to war or surrender. Urbana: University of Illinois Press.

32.

Praszkier

Nowak

Coleman

P. T.

(2010). Social entrepreneurs and constructive change: The wisdom of circumventing conflict. Peace and Conflict: Journal of Peace Psychology, 16, 153–174.

33.

Roberts

B. W.

Walton

K. R.

Viechtbauer

(2006). Patterns of mean-level change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132, 1–25.

34.

Smith

N. K.

Cacioppo

J. T.

Larsen

J. T.

Chartrand

T. L.

(2003). May I have your attention please: Electrocortical responses to positive and negative stimuli. Neuropsychologia, 41, 171–183.

35.

Strogatz

S. H.

(1994). Nonlinear dynamics and chaos. Cambridge, MA: Westview Press.

36.

Sweeney

Shepperd

J. A.

Carroll

P. J.

(2009). Expectations for other’s outcomes: Do people display compassionate bracing? Personality and Social Psychology Bulletin, 35, 160–171.

37.

Taylor

S. E.

(1991). Asymmetric effects of positive and negative events: The mobilization-minimization hypothesis. Psychological Bulletin, 110, 67–85.

38.

Todman

Dugard

(2007). Approaching multivariate analysis: An introduction for psychology. New York, NY: Psychology Press.

39.

Tullett

A. M.

Teper

Inzlicht

(2011). Confronting threats to meaning: A new framework for understanding responses to unsettling events. Perspectives on Psychological Science, 6, 447–453.

40.

Vallacher

R. R.

Coleman

P. T.

Nowak

Bui-Wrzosinska

(2010a). Rethinking intractable conflict: The perspective of dynamical systems. American Psychologist, 65, 262–278.

41.

Vallacher

R. R.

Coleman

P. T.

Nowak

Bui-Wrzosinska

(2010b). Dynamical foundations of intractable conflict: Introduction to the special issue. Peace and Conflict: Journal of Peace Psychology, 16, 113–125.

42.

Vallacher

R. R.

Nowak

(1994). The stream of social judgement. In Vallacher

R. R.

Nowak

(Eds.), Dynamical systems in social psychology (pp. 251–278). San Diego, CA: Academic Press.

43.

Vallacher

R. R.

Nowak

Froehlich

Rockloff

(2002). The dynamics of self-evaluation. Personality and Social Psychology Review, 6, 370–379.

44.

Vallacher

R. R.

Nowak

Markus

Strauss

(1998). Dynamics in the coordination of mind and action. In Kofta

Weary

Sedlek

(Eds.), Personal control in action: Cognitive and motivational mechanisms (pp. 27–59). New York, NY: Plenum.

45.

Vallacher

R. R.

Wegner

D. M.

(1985). A theory of action identification. Hillsdale, NJ: Lawrence Erlbaum.

46.

Vallacher

R. R.

Wegner

D. M.

(2012). Action identification theory. In van Lange

Kruglanski

Higgins

E. T.

(Eds.), Handbook of theories in social psychology (pp. 327–348). London, England: Sage.