Responses to dissatisfaction in friendships and romantic relationships: An interpersonal script analysis

Abstract

According to interpersonal script models, people’s responses to relational events are shaped by the reaction they expect from a close other. We analyzed responses to dissatisfaction in close relationships from an interpersonal script perspective. Participants reported on how a close friend or romantic partner would react to their expressions of dissatisfaction (using the exit-voice-loyalty-neglect typology). They were also asked to forecast whether the issue would be resolved (i.e., anticipated outcomes). Our main hypothesis was that people’s expectations for how a close other would respond to dissatisfaction would be dependent on their own self response. Further, we predicted that passive responses would be more common and viewed as less deleterious to a friendship than a romantic relationship. Results indicated that the responses that were expected from close others were contingent on how self responded. Moreover, as predicted, these contingencies followed different tracks depending on the type of relationship. Friends were more likely to expect passive responses to self’s expression of dissatisfaction, especially if self responded with neglect, whereas romantic partners expected more active responses. Furthermore, people anticipated that the issue would be more likely to be resolved if their friend (vs. romantic partner) responded passively and less actively (especially for destructive responses). It was concluded that people hold complex, nuanced interpersonal scripts for dissatisfaction and that these scripts vary, depending on the relationship context.

Keywords

Exit-voice-loyalty-neglect typology friendships interpersonal scripts responses to dissatisfaction romantic relationships

Periods of dissatisfaction are inevitable in close relationships (e.g., hurtful comments, disagreement about plans, jealousy of time spent with others) and people must decide how they will respond (e.g., Canary & Lakey, 2013; Spitzberg & Cupach, 2013). Take, for instance, Pat and Chris who have known each other for 10 years are close and interdependent, but, unexpectedly, Pat forgets Chris’s birthday. Chris is upset but unsure about how to respond (i.e., ignore it, wait it out, discuss the issue, threaten to end the relationship). Before responding, Chris considers how Pat will react and how the situation will turn out if, for instance, Chris decides that it would be best to discuss the issue. If Pat was a friend, a different reaction might be expected than if Pat was a romantic partner.

Understanding how people respond to problems in close relationships has been identified as a major goal of relationship science (see Overall & McNulty, 2017, for a review). Rusbult and colleagues (see Rusbult, Yovetich, & Verette, 1996, for a review) delineated four major categories of responses to dissatisfaction in close relationships: exit (e.g., contemplating ending the relationship, threatening to do so), voice (e.g., discussing the problem, suggesting solutions, and compromising), loyalty (e.g., waiting and hoping that things will improve), and neglect (e.g., ignoring one’s partner, refusing to discuss the problem). These responses can be organized in terms of two dimensions: active/passive and constructive/destructive. Active/passive refers to whether a response directly or indirectly influences the immediate situation (i.e., exit and voice are active; loyalty and neglect are passive). The constructive/destructive dimension refers to the impact of a response on the relationship (i.e., voice and loyalty are constructive; exit and neglect are destructive). Support for this model has been obtained in many studies (e.g., Drigotas, Whitney, & Rusbult, 1995; Overall, Sibley, & Travaglia, 2010; Rusbult, Johnson, & Morrow, 1986a, 1986b; Rusbult & Zembrodt, 1983).

An underlying assumption of Rusbult and colleagues’ work is that responses to dissatisfaction are interdependent. That is, in a dissatisfaction episode, an individual’s response (e.g., exit, voice, loyalty, neglect) is influenced by his or her partner’s response to dissatisfaction (see Rusbult et al., 1996, for a review). Moreover, as discussed later, there is evidence that these patterns are linked to relationship well-being (e.g., Overall & Sibley, 2008; Rusbult, Bissonnette, Arriaga, & Cox, 1998). This work has been important in documenting how people respond to dissatisfaction and the outcomes of those responses. There is an equally important question that remains unanswered, however, and that is, Why do people respond the way they do? We propose that an answer may be found in interpersonal script models of relationship expectations.

What determines people’s responses to dissatisfaction? An interpersonal script perspective

The central idea underlying interpersonal script models is that based on repeated interactions with a close other, people develop expectations about how he or she is likely to respond to them; these expectations are a determinant of self responses. Interpersonal script models assume that patterns of relating become stored in cognitive representation in the form of scripts. In a landmark article, Baldwin (1992) proposed that individuals internalize expectations for interpersonal events in the form of relational schemas. Relational schemas are mental representations of patterns of relating that guide the interpretation of new information about a relationship and influence interpersonal behavior. They are composed of a self-schema that represents one’s conception of the self, an other schema that represents the attributes of the other person, and an interpersonal script which refers to knowledge about interaction patterns between self and other. Such expectations are represented cognitively as if-then contingencies (e.g., “If I try to discuss an upsetting situation, then my partner will withdraw.”). These contingencies are based on repeated similar interactions with a close other and therefore provide information about how she or he is likely to respond to self’s expression of dissatisfaction. Expected reactions are hypothesized to shape the way that self chooses to respond. These ideas have received some empirical attention in the conflict and anger literatures (e.g., Fehr, Baldwin, Collins, Patterson & Benditt, 1999; Miller, 1991).

In the present investigation, our goal was to examine the contingencies between self responses and expected close other responses (i.e., interpersonal scripts) pertaining to dissatisfaction experiences in relationships. We expected that people would have knowledge of dissatisfaction scripts. We made the further prediction that these scripts would be nuanced, depending on the context.

Variations in interpersonal scripts as a function of relationship type

An important, but generally overlooked, point is that expectations are not fixed, but rather, follow different paths (Abelson, 1981). Abelson referred to these variations in scripts as tracks, which he defined as “different script variants, each embodying characteristic paths, scene selections, and props not shared by other tracks” (p. 723). Furthermore, Planalp (1987) stated that “people approach others with some knowledge of what behavior is appropriate for the type of relationship they have” (p. 176). In the present research, we examined whether people hold different scripts for dissatisfaction in romantic relationships compared to close, same-sex friendships.

Romantic relationships generally are characterized by explicit expressions of dissatisfaction. For example, in Rusbult and colleagues’ program of research on the exit-voice-loyalty-neglect model, the active response of voice (talking it over) is more commonly reported in dating and marital relationships than the passive responses of loyalty and neglect (Drigotas et al., 1995; Rusbult et al., 1986a, 1986b; Rusbult, Morrow, & Johnson, 1987). In contrast, research on conflict interactions in friendships (i.e., reports of people’s actual experiences, rather than their cognitive representation of them in the form of scripts) has shown that friends typically respond to anger and conflict with avoidance (Baxter, Wilmot, Simmons, & Swartz, 1993; Canary & Spitzberg, 1987; Fehr, 1996; Gergen & Gergen, 1988; Miller, 1991; Sillars, 1980). This finding is confirmed in studies that have directly compared conflict resolution strategies in friendships and romantic relationships—friends are more likely opt for indirect strategies, avoidance, and passivity than are romantic partners (e.g., Baxter et al., 1993; Canary, Stafford, Hause, & Wallace, 1993; Davis & Todd, 1985; Wright, 1985). Based on these differences in relationship experiences, it seemed likely that interpersonal scripts for dissatisfaction would be different for friendships and romantic relationships. More specifically, we predicted that in the context of friendships, people would expect more passive responses (and less active responses) than in the context of romantic relationships. Conversely, we expected that active responses would be viewed as more common in the scripts for romantic relationships than friendships.¹

To summarize, there is evidence that people respond differently to conflict, anger, and dissatisfaction, depending on the context (e.g., relationship type). We propose that based on these experiences, people develop knowledge of how interactions about dissatisfying situations are likely to unfold, and that these scripts differ, depending on whether their interaction partner is a romantic partner or a close friend.

Extending interpersonal script analyses to include anticipated outcomes

Our main hypothesis was that people have knowledge of how a close other is likely to react to their expressions of dissatisfaction. We made the further prediction that people also have knowledge of how these reactions are likely to shape the outcome (i.e., knowledge of whether the issue will be resolved). For example, it is plausible that based on repeated similar experiences, a person might learn: “If my partner responds to my expression of dissatisfaction by ignoring the issue, the issue will fester.” This person also might learn that “If my partner responds to my expression of dissatisfaction by calmly engaging in a discussion of the issue, the issue will get resolved.” Thus, we are proposing that people do not only have knowledge of how a close other is likely to respond to their expressions of dissatisfaction, but also have knowledge of how the other person’s response will influence the eventual outcome.

Rusbult and colleagues’ research on people’s experiences of dissatisfaction (rather than the cognitive representation of these experiences, which is the focus of our research) is consistent with this idea. For instance, Rusbult et al. (1986b) asked participants to report on their tendency to respond with exit, voice, loyalty, or neglect when they were unhappy with their partner as well as perceived partner responses. To examine interdependence, participants were also asked to provide open-ended descriptions of how they would react if their partner responded with exit, voice, loyalty, or neglect. Generally, constructive self and perceived partner responses were negatively associated with relationship distress, whereas destructive responses were positively associated with distress (the correlations were stronger for destructive responses). With respect to interdependence, it was found that a willingness to inhibit retaliatory destructive impulses and respond constructively to a partner’s destructive behaviors (i.e., use of voice or loyalty in the face of a partner’s exit or neglect response) predicted lower distress (Rusbult et al., 1986b). This finding has been replicated in other studies and is referred to as accommodation or the “golden rule” of responses to dissatisfaction (e.g., Arriaga & Rusbult, 1998; Rusbult, Verette, Whitney, Slovik, & Lipkus, 1991).

In a subsequent diary study, participants reported on their own and their partner’s responses to daily incidents of dissatisfaction (using the exit-voice-loyalty-neglect typology) as well as the perceived impact of these responses on the problem and the relationship (Drigotas et al., 1995). In addition to replicating the general finding in this literature that constructive responses have a more positive impact than destructive responses, the researchers found that constructive responses did not have equally beneficial effects. That is, the active, constructive response of voice had a more positive impact on the problem than the passive, constructive response of loyalty. Overall et al. (2010) also found that voice had a more beneficial effect than loyalty. In fact, in their research, the loyalty response was just as detrimental as the destructive responses. Thus, it appears that loyalty responses, although constructive in theory, do not necessarily have salutary effects, presumably because they are not as visible to the partner (Drigotas et al., 1995). Loyalty responses may also be misinterpreted as neglect (Overall et al., 2010).

The studies we have discussed so far were focused on romantic relationships. To our knowledge, similar studies have not been conducted on the link between responses to dissatisfaction and outcomes in the context of friendships. As discussed earlier, there is evidence that friends tend to respond to dissatisfaction, conflict, and anger more passively than do romantic partners. Thus, we conjectured that friends might be more likely to anticipate that issues would be resolved if they “let things be” (i.e., standing by and hoping that things will improve [loyalty] or by ignoring or avoiding the issue [neglect]).

To conclude, when reviewing research on the exit-voice-loyalty-neglect typology, we observed the following: First, responding constructively versus destructively has been the focal point of norm violations in relationships; the consequences of norm violations related to the active–passive dimension (i.e., loyalty, neglect) are not well known. Second, researchers have examined people’s exit, voice, loyalty, and neglect responses. Whether people store close others’ reactions to their expressions of dissatisfaction in cognitive representation, and the outcomes that are likely to be associated with those reactions, remains to be tested. Finally, research on the exit-voice-loyalty-neglect model has concentrated on romantic relationships. It is unclear whether the findings obtained for romantic partners in past studies apply to other close relationships such as close friendships. In fact, the friendship literature would lead one to expect much more passive responses to dissatisfaction than has been found in romantic relationships.

Overview and hypotheses

The purpose of this research was to examine responses to dissatisfaction in friendships and romantic relationships from an interpersonal script perspective. We conducted a two-part study. In Part I, we focused on expected close other responses to self’s expression of dissatisfaction. Our central prediction was that expected close other’s reactions would be contingent on self responses (i.e., we would find evidence of interpersonal scripts) and that dissatisfaction scripts would differ for friendships and romantic relationships. More precisely, we hypothesized that friendship scripts would be characterized by greater passivity (loyalty and neglect) compared to scripts for romantic relationships and, conversely, that active responses would be more common in scripts for romantic relationships than friendships. Thus, unlike past research on the exit-voice-loyalty-neglect model, we focused on responses in the context of friendships versus romantic relationships for the active–passive dimension.

In Part II, we extended our script analysis to include anticipated outcomes. The underlying idea was that in addition to knowledge of how a close other is likely to respond to self’s expression of dissatisfaction, people have knowledge of the likely outcomes of these responses (i.e., whether the issue will be resolved). For example, based on past experiences, a person might be tempted to lash out, but know that his or her partner will lash out in return, and that the result will be an escalation of the issue. We also predicted that the outcomes that people associate with particular scripts would vary by relationship type. More specifically, we hypothesized that passive responses would be associated with less deleterious outcomes in friendships than in romantic relationships (and conversely that active responses would be associated with more negative outcomes in friendships than romantic relationships).

Method

Participants

Introductory psychology students (N = 233; 135 women, 97 men, unreported gender for 1 participant) at a large, central North American university participated in this research for course credit. The mean age of participants was 20 (M _age = 19.89, SD = 3.91), with a range of 18 to 43 years. In terms of relationship status, 37% were seriously involved, 25% were casually dating, 32% were not currently dating, 5% did not report their relationship status or indicated “other,” and 1% were married. Those who were romantically involved reported an average relationship length of 19.5 months (1–132 months).

Part I: Interpersonal scripts questionnaire

Participants were randomly assigned to complete an interpersonal script questionnaire with reference to either a romantic relationship (N = 114) or a close, same-sex friendship (N = 119). We followed the format developed by Fehr et al. (1999) to examine anger scripts in close relationships. More specifically, participants were first asked to imagine responding to a dissatisfying issue in a romantic relationship or a close same-sex friendship in various ways (i.e., exit, voice, loyalty, and neglect). Next, participants were told that “we are interested in how a partner [friend] might react to you, depending on your response to dissatisfaction.” We then described exit, voice, loyalty, and neglect responses (taken from Rusbult, Zembrodt, and Gunn, 1982; e.g., for voice: “You respond to a dissatisfying situation by discussing the problem, suggesting solutions, trying to find a compromise, trying to change himself/herself or you”). Participants then rated (on a 7-point scale) the likelihood that a partner/friend would react to each expression of dissatisfaction with exit, voice, loyalty, and neglect (see Table 1 for the wording of these items). Thus, participants imagined responding with exit, voice, loyalty, and neglect, and, for each of these responses, rated the likelihood that a partner or friend would react with exit, voice, loyalty, or neglect.

Table 1.

Mean expected close other responses to dissatisfaction.

	Overall	Friendship	Romantic relationship	p
Voice	4.74_a (.08)	4.45_a (.11)	5.03_a (.11)	.001
Discussing the problem, suggesting solutions, trying to find a compromise, and trying to change him/herself or you
Loyalty	4.24_b (.07)	4.24_a (.10)	4.24_b (.10)	.99
Waiting and hoping that things will improve, giving things some time, and continuing to have faith in the friendship/relationship and you
Neglect	3.68_c (.08)	3.89_b (.11)	3.47_c (.12)	.009
Ignoring you, spending less time with you, refusing to discuss the problem, and complaining without offering solutions to the problem
Exit
Thinking or talking about ending the friendship/relationship, threatening to end the friendship/relationship, doing things to drive you away, and end the friendship/relationship	3.17_d (.08)	2.98_c (.11)	3.35_c (.11)	.02

Note. Within columns, mean ratings with different subscripts differ at p < .05 (with exception, the difference between loyalty and neglect for friends reached marginal significance). Numbers in parentheses represent standard errors. The Relationship Type × Expected Close Other Response interaction was significant, F(3, 224) = 10.27, p < .001, η_p ² = .12. For relationship type differences, η_p ² s ranged from <.001 to .06.

Part II: Anticipated outcomes

After completion of the Part I questionnaire, participants received a second questionnaire in which the focus was on the outcomes they anticipated as a function of expected close other responses to dissatisfaction (i.e., exit, voice, loyalty, and neglect). For example, for the neglect response, the description read “Imagine that a partner [friend] responds to your expression of dissatisfaction by avoiding the problem, spending less time with you, or ignoring you.” Following each expected close other response, participants were asked to rate the probability that the dissatisfying issue would be resolved if a partner/friend responded in that way (i.e., anticipated outcomes). To assess anticipated outcomes, we adapted a measure developed by Miller (1991) in her study of roommates’ conflict scripts. She asked participants to circle one of three possible outcomes: the problem will be totally resolved; the problem will be somewhat resolved but dissatisfaction will still linger; the problem will not be resolved. We converted this measure to a continuous scale and asked, “What is the likelihood the problem would be resolved if you responded by…,” where 1= the problem will not be resolved, 4 = the problem will be somewhat resolved but dissatisfaction will still linger, and 7 = the problem will be totally resolved. Participants completed this question four times (i.e., for exit, voice, loyalty, and neglect partner/friend reactions).

Finally, participants were asked to provide demographic information. They were also asked whether they had completed the measures with respect to a current partner [friend], former partner [friend], or with respect to partners [friends] in general.² Those who had responded with a current partner [friend] in mind were asked to indicate the closeness of their relationship [friendship] using the Inclusion of Other in the Self scale (Aron, Aron, & Smollan, 1992).

Results

Before proceeding with the analyses to test our hypotheses, we conducted a two-way ANOVA with Relationship Type and Gender as the between-person variables and ratings of closeness as the dependent variable to ensure that any differences between friendships and romantic relationships were not due to differences in closeness. The Relationship Type main effect did not reach statistical significance, F(1, 160) = 1.09, p = .30 (M = 4.48, SE = .17 for friends; M = 4.76, SE = .21 for romantic partners), nor did the Gender main effect, F(1, 160) = 2.71, p = .10 (M = 4.84, SE = .16 for women; M = 4.40, SE = .22 for men), or the interaction, F < 1. Thus, it was not the case that romantic relationships were rated as significantly closer than were friendships.

Part I: Expected close other responses

Given our interpersonal script conceptualization, it was critical that we obtain a significant interaction between self responses and expected close other reactions. We also tested subsidiary hypotheses regarding relationship type differences in interpersonal scripts. In brief, we hypothesized that participants would be more likely to expect passive reactions from a friend than a romantic partner. Conversely, we hypothesized that people in romantic relationships would expect more active responses from a partner than a friend.

The next set of analyses focused on expected partner reactions to self’s expression of dissatisfaction. Given our interpersonal script conceptualization, it was crucial that a significant interaction be obtained between self responses and expected partner reactions. Such an interaction would demonstrate that self responses to dissatisfaction are, indeed, contingent on the reactions that are expected from a partner or friend. We also tested several subsidiary hypotheses regarding gender and relationship. Our predictions were tested in a 2 (Relationship Type) × 4 (Self Response) × 4 (Expected Close Other Response) three-way mixed ANOVA.³ Relationship type was a between-person variable; self response and expected close other response were within-person variables. In all of our follow-up analyses of responses, we adjusted for multiple comparisons using a Bonferonni correction. There was a marginally significant main effect for relationship type, F(1, 226) = 3.60, p = .06, η_p ² = .02. Participants in the romantic relationship condition tended to expect higher levels of response, overall, from their partner than did those in the friendship condition (M = 4.02, SE = .05 vs. 3.89, SE = .05). The main effect for self response was significant, F(3, 224) = 27.42, p < .001, η_p ² = .27. Participants expected that a response from a close other was most likely when self engaged in exit (M = 4.17, SE = .05) and neglect (M = 4.15, SE = .06), followed by loyalty (M = 3.83, SE = .04) and voice (M = 3.68, SE = .04). (Exit and neglect did not differ significantly; loyalty and voice differed significantly from each other, and from exit and neglect, ps < .005.) In other words, participants believed that destructive responses, on their part, would be most likely to evoke a reaction from their partner or friend. A significant main effect was also found for expected partner response, F(3, 224) = 61.67, p < .001, η_p ² = .45. As shown in Table 1 (see first column), participants were most likely to expect a voice response from their partner or friend, followed by loyalty, neglect, and, then, exit. Thus, people expect that close others generally will react constructively when they express dissatisfaction.

Our key prediction was that expected close other reactions would be contingent on self responses. This prediction was supported; the Self Response × Expected Close Other Response interaction was significant, F(9, 218) = 39.94, p < .001, η_p ² = .62. We decomposed this interaction by examining repeated measures ANOVA tests for the expected close other responses for each self response. Significant effects were found for all four self responses (see Table 2). Although the focus of the present investigation was on the active–passive dimension (discussed next), it can be seen in Table 2 that, in general, people expected constructive responses when they responded constructively and destructive responses when they responded destructively (this issue is taken up in the “Discussion” section).

Table 2.

Mean expected close other responses as a function of self responses.

	Self responses
Expected close other responses	Voice	Loyalty	Neglect	Exit
Voice	5.54_a (.09)	4.82_a (.11)	4.23_b (.11)	4.31_a (.11)
Loyalty	4.47_b (.10)	4.88_a (.10)	4.02_b (.11)	3.57_b (.10)
Neglect	2.43_c (.10)	3.07_b (.11)	4.73_a (.13)	4.50_a (.13)
Exit	2.24_c (.09)	2.51_c (.10)	3.61_c (.12)	4.31_a (.12)

Note. Mean ratings (within each column) with different subscripts differ at p < .05 (for each self response, repeated measures ANOVA tests for the expected close other responses were significant, ps < .001). Numbers in parentheses represent standard errors. Bolded responses within each column represent the largest mean expected close other response (in cases where there is no statistically significant difference between the highest expected close other responses, multiple values are bolded).

We also hypothesized that interpersonal scripts would follow different tracks, depending on the relationship context. Specifically, we predicted that people would expect more passive responses from their friends and more active responses from their romantic partners. In support of this prediction, the Relationship Type × Expected Close Other Response interaction was significant, F(3, 224) = 10.27 p < .001, η_p ² = .12. As seen in Table 1, participants expected more neglect (a passive response) from a friend than a romantic partner in response to their expression of dissatisfaction, although not more loyalty. As well, people expected more voice and exit (active responses) from a romantic partner than a friend.⁴

Next, we explored whether relationship type differences in expected close other responses would be modified by self responses. There was a significant Relationship Type × by Self Response × Expected Close Other Response interaction, F(9, 218) = 2.31, p = .02, η_p ² = .09. As shown in Table 3, consistent with the idea of a culture of passivity in friendships, if self responded passively (i.e., with neglect and loyalty), friends were more likely to expect neglect in return than were romantic partners (the effect was only marginally significant when self responded with loyalty). In contrast, people in romantic relationships expected more active responses than friends across all types of self responses. Thus, people in romantic relationships expect more active responses from their partners no matter what response they use.⁵

Table 3.

Mean expected close other responses as a function of self responses and relationship type.

	Self responses
	Voice			Loyalty			Neglect			Exit
	Friend	Romantic	p	Friend	Romantic	p	Friend	Romantic	p	Friend	Romantic	p
Expected close other responses
Voice	5.31 (.13)	5.80 (.13)	.01	4.53 (.16)	5.13 (.16)	.007	3.94 (.15)	4.53 (.15)	.006	4.01 (.15)	4.63 (.16)	.004
Loyalty	4.54 (.14)	4.44 (.15)	.64	5.00 (.13)	4.75 (.14)	.17	3.95 (.15)	4.10 (.15)	.48	3.47 (.14)	3.68 (.14)	.30
Neglect	2.53 (.13)	2.31 (.14)	.29	3.25 (.15)	2.89 (.15)	.09	5.05 (.18)	4.40 (.18)	.01	4.71 (.18)	4.26 (.18)	.07
Exit	2.04 (.12)	2.44 (.13)	.02	2.20 (.14)	2.84 (.14)	<.001	3.29 (.17)	3.94 (.17)	.008	4.41 (.17)	4.20 (.18)	.32

Note. Numbers in parentheses represent standard errors. The reported p values are for the comparison of friendship versus romantic group means for each specific self response × expected response row (using a Bonferroni correction). All follow-up mixed ANOVA tests (i.e., Relationship Type × Expected Close Other Response) were significant for each self response, ps < .005. With exception, exit was marginally significant, F(3, 226) = 2.39, p = .07. For relationship type differences, η_p ² s ranged from .001 to .05.

Part II. Anticipated outcomes

In Part II of this study, we extended our script analysis beyond self–other contingencies and examined anticipated outcomes. The main idea was that people can not only forecast expected partner reactions but also have expectations about whether those reactions will lead to a positive or negative outcome. More specifically, we investigated whether anticipated outcomes would vary, depending on relationship type. We predicted that passive responses would be associated with less negative outcomes in friendships than in romantic relationships and, conversely, that active responses would be associated with more positive outcomes in romantic relationship than friendships.

We tested these predictions in a 2 (Relationship Type) × 4 (Expected Close Other Response) two-way mixed ANOVA (relationship type was a between-person variable and expected close other response was a within-person variable). The Relationship Type main effect was not significant, F < 1. The main effect for Expected Close Other Response was significant, F(3, 229) = 364.32, p < .001, η_p ² = .83. Participants believed that the active, constructive response of voice would be most likely to lead to a successful resolution of the issue, followed by loyalty (see Table 4, first column). A neglect response was seen as least likely to resolve the issue—more so than exit. These findings are best interpreted in light of the significant Relationship Type × Expected Close Other Response interaction, F(3, 226) = 9.11, p < .001, η_p ² = .11.⁶ Follow-up tests showed that friends expected better resolution of the issue if they responded passively, although the effect was only statistically significant for neglect (see Table 4). Conversely, romantic partners expected better resolution of the issue if they responded actively (with voice and exit) than did friends; however, this effect only reached statistical significance for exit.⁷

Table 4.

Mean anticipated outcome ratings as a function of relationship type.

	Overall	Friendship	Romantic relationship	p
Expected close other response
Voice	5.62_a (.09)	5.49_a (.12)	5.75_a (.12)	.12
Loyalty	4.41_b (.09)	4.50_b (.12)	4.32_b (.12)	.29
Neglect	1.87_d (.07)	2.07_c (.10)	1.67_d (.10)	.006
Exit	2.37_c (.10)	2.09_c (.14)	2.64_c (.14)	.006

Note. Mean ratings (within each column) with different subscripts differ at p < .05. The reported p values are for the comparison of friendship versus romantic group means for each specific expected close other response row. Numbers in parentheses represent standard errors. For relationship type differences, η_p ² s ranged from .005 to .03.

In conclusion, there is considerable evidence, particularly in the marital literature, that destructive responses to dissatisfaction, conflict, and anger are associated with negative outcomes, whereas constructive responses lead to more positive outcomes. Interpersonal script models make the assumption that people internalize these patterns of relating, based on repeated similar experiences with a close other. Our results suggest that people’s relational knowledge extends beyond a simple dictum that good behaviors will lead to good outcomes and bad behaviors will lead to bad outcomes. We found that scripts for expected outcomes varied as a function of relationship type such that the passive response of neglect from a friend was seen as having a less negative effect than a neglect response from a romantic partner. In contrast, the active, destructive response of exit was seen as less detrimental in a romantic relationship than in a friendship.

Discussion

Relationship scientists have examined scripts for a variety of relational phenomena such as developing a relationship, relational boredom, and ending a relationship (e.g., Harasymchuk & Fehr, 2010; Holmberg & Mackenzie, 2002; see Honeycutt & Bryan, 2011, for a review). They have also conducted analyses more specifically focused on interpersonal scripts (i.e., self–other contingencies), such as anger interactions in close relationships (Fehr et al., 1999), intimacy interactions in friendships (Fehr, 2004), attachment relational schemas (Baldwin & Meunier, 1999), and interpersonal scripts for conflict (Miller, 1991). We extended this line of research to examine people’s interpersonal scripts for dissatisfaction in romantic relationships and friendships. Furthermore, we explored the anticipated outcomes that are linked to different expected close other responses. Overall, we found evidence that responses to dissatisfaction and their anticipated outcomes are stored in cognitive representation. Thus, one reason that people may choose to respond in particular ways is that they hold expectations, based on past experience, about how their partner or friend is likely to react, and also have knowledge of the consequences of those reactions.

Relationship type differences in interpersonal scripts for active–passive responses

In our view, the finding that interpersonal scripts follow different tracks, depending on relationship type, is the single most important contribution of the present research. The results revealed that interpersonal scripts for friendships are characterized by greater passivity than interpersonal scripts for romantic relationships. More specifically, people expected more passive responses (i.e., neglect) from a friend than a romantic partner (see also note 4). As well, we found that friends expect that passive responses will be reciprocated (especially neglect) unless they respond actively. Further, a passive, neglect response from a friend was seen as having a less destructive effect (i.e., less likely to undermine resolution of the issue) than a neglect response from a romantic partner (see also note 7). This finding is buttressed by other research showing that when describing an experience of neglect from a friend, people reported less high arousal, negative emotion than when they described an experience of neglect from a romantic partner (Fehr & Harasymchuk, 2005). Although we had hypothesized that the passive response of loyalty would also be more characteristic of friendship scripts than romantic relationship scripts, we generally did not obtain significant effects for loyalty. As mentioned earlier, in research on romantic relationships, it has been found that loyalty responses have less impact on relationships than voice responses (Drigotas et al., 1995) or no effect (Overall et al., 2010). It has been suggested that loyalty is less visible than the other responses and, therefore, has less of an effect. This may account for why loyalty does not play a prominent role in people’s dissatisfaction scripts for friendships as well.

Turning to romantic relationships, we found that romantic partners expected more active responses to dissatisfaction than did friends. Indeed, people in romantic relationships expected more active responses than friends across all self responses (whether active or passive). People also thought that the active destructive response of exit from a romantic partner would be more likely to resolve the dissatisfying issue than an exit response from a friend. There was also a tendency for people to expect better outcomes from a voice response in a romantic relationship than a friendship, although this finding did not reach statistical significance.

The picture that emerges for friendships is one of passivity, especially the passive, destructive response of neglect. This stands in sharp contrast to the “take action” approach that was found in people’s scripts for dissatisfaction in a romantic relationship. This difference in scripts is not attributable to differences in closeness—we did not find significant relationship type differences in terms of level of closeness. Rather, this passivity seems to reflect an undervaluing of friendships, relative to romantic relationships. Indeed, it has been argued that Western culture places importance on institutionalized relationships such as marital relationships, whereas the voluntary relationship of friendship is given short shrift. For example, Oliker (1989) conducted interviews with women and found that although female friendships were important and valued, these women accepted that no matter how close a friendship, their marital and family relationships should “come first” (e.g., one’s birthday should be spent with one’s husband, rather than a close friend). In a program of research on relational maintenance, Canary et al. (1993) found that more active maintenance and repair efforts were reported in romantic relationships than in friendships. They suggested that people may believe that romantic relationships warrant or “deserve” direct maintenance—more so than friendships. The lesser value that is placed on friendships is also apparent in people’s responses to the dissolution of these kinds of relationships. When romantic relationships end, the event is typically announced to social network members, who may respond by offering social support. The individuals involved may seek therapy to help them cope with the loss. In contrast, friendship dissolution is rarely announced, rarely elicits an outpouring of social support, and rarely involves professional intervention (see Fehr, 1996). Explanations for the undervaluing of friendships include evolutionary accounts and socialization accounts (see Fehr & Harasymchuk, 2005).

There are other possible reasons why friendship scripts are characterized by passivity. People have competing goals in close relationships. Self-disclosure (e.g., directly expressing dissatisfaction in the relationship) might conflict with other goals such as self-protection (e.g., Baxter & Montgomery, 1996). One way to manage competing goals is topic avoidance (e.g., Baxter & Wilmot, 1985; Guerrero & Afifi, 1995). For example, Afifi and Guerrero (1998) examined reasons for topic avoidance in same-sex and cross-sex friendships and found that self-protection (e.g., “It would leave me too vulnerable”; “I would be embarrassed to disclose”) was the primary reason why friends chose to avoid bringing up difficult topics. In other words, there might be risks associated with disclosing dissatisfaction to a friend (e.g., fear of rejection, concerns about opinions of the larger friendship network).

Finally, the differences between scripts for friendships and romantic relationships might be due to the exclusivity of romantic relationships. It is typical and even expected that people will have more than one close friend, but generally only one romantic partner. As such, people in friendships might not feel the need to respond actively when dissatisfied because they may have other friends to turn to for companionship and advice while the dissatisfying event “cools down.” Relatedly, compared to romantic relationships, friendships might be characterized by lower commitment and investments, which could, in turn, be associated with less active responses.

Relationship type differences in interpersonal scripts for constructive–destructive responses

Although not the focal point of this research, we also obtained findings relevant to the constructive–destructive dimension. In general, we found that people (friends and romantic partners) tended to expect reciprocal responses, especially for constructive responses. For instance, if self responded with voice or loyalty, people expected their close other to respond with voice or loyalty rather than with exit or neglect (see Table 2). In Rusbult and colleagues’ research, people in romantic relationships report that they are most likely to respond to dissatisfaction with the constructive response of voice, followed by loyalty. Our findings suggest that this positivity is also internalized in the form of interpersonal scripts for dissatisfaction for both romantic relationships and close friendships.

Rusbult and colleagues have also found evidence of accommodation (i.e., responding constructively even when the partner responds destructively). Is this reflected in scripts for dissatisfaction? For the destructive self response of exit, there was some evidence that people expect accommodation, in the form of voice; however, people also expected reciprocal destructive responses to a similar degree (see Table 2). We did not find evidence of accommodation for the destructive response neglect (i.e., when self responds with neglect, people did not expect constructive responses more than destructive ones). However, when taking into account the type of relationship, there was a consistent pattern such that when self responded destructively, people expected that their romantic partner would be more likely to respond with voice and loyalty than their friend (although this pattern did not reach significance for the loyalty responses; see also note 5). Put another way, our findings suggest that people internalize accommodation norms (i.e., “good behavior rules”) for their romantic relationships but are less likely to apply this norm to friendships.

In terms of anticipated outcomes, people expected that constructive responses of voice and loyalty, especially voice, from a close other would lead to resolution of the issue; the destructive responses of neglect and exit, especially neglect, from a close other were seen as impeding resolution of the issue. At a finer-grained level, we found that voice and loyalty did not lead to the same level of expected resolution; instead, voice was seen as producing a more positive outcome than loyalty (by over one mean point on a 7-point scale). These finding align with those obtained by Drigotas et al. (1995) in their daily diary study (in which people reported on responses to dissatisfying interactions and the impact of those responses on the problem at hand and the relationship; see also Overall et al., 2010). Interestingly, a neglect response was associated with a greater impediment to the resolution of a dissatisfying issue than was exit. In other words, when trouble is brewing in a close relationship, doing nothing is seen as even worse than lashing out and responding in a destructive manner.

In terms of relationship type, differences were limited to destructive responses as a function of whether it was an active (exit) or passive (neglect) response. For friendships (vs. romantic relationships), the active, destructive response (exit) was associated with a more detrimental outcome; for romantic relationships (vs. friendships), the passive destructive response (neglect) was associated with a more detrimental outcome.

Implications of our findings

Rusbult and colleagues’ program of research on the exit-voice-loyalty-neglect model has focused on romantic relationships. In that work, most of the “action” has centered on the constructive–destructive dimension (e.g., constructive responses are associated with relational health; destructive responses are associated with relationship distress). Indeed, Rusbult et al. (1987) commented that “the active-passive dimension has been less successfully predicted” (p. 295). There are some variables that are associated with the active–passive dimension, but these tend to be individual difference variables, rather than relationship variables. For example, Rusbult et al. (1987) found that passivity was correlated with low self-esteem. Kammrath and Dweck (2006) found that incremental theorists (people who believe that personality can change) used the active response of voice whereas entity theorists (people who believe that personality is fixed) used the passive response of loyalty. As well, Birditt and Fingerman (2005) found that older people are more likely to use a loyalty response in their social networks than younger people. However, to our knowledge, our findings are the first to show that the active–passive dimension is relevant to close relationships—if one looks beyond the context of romantic relationships.

Our findings also have implications for practitioners. When problems arise in relationships, the advice that people typically are given is to “talk about it.” Our findings suggest that this advice might be more fitting for romantic partners. Friends are more likely to believe the opposite, namely that actively bringing up issues will be worse for the relationship than ignoring them. Although our research does not speak to the issue of whether passivity actually results in positive outcomes, it seems important for therapists to be aware that friends tend not to subscribe to the “talk about it” heuristic.

Limitations and future directions

The present investigation has a number of limitations. One is the reliance on undergraduate students. It is possible that interpersonal scripts for dissatisfaction vary, depending on age and relationship length. However, the fact that our participants’ scripts generally aligned with research on actual experiences of dissatisfaction, conflict, and anger (e.g., active responses in romantic relationships; passive responses in friendships) suggests that these scripts may be generalizable. In support of this line of thinking, Fehr (2004) found that university students and a community sample of older adults generated very similar patterns of relating that exemplified intimacy in a same-sex friendship. She suggested that relational knowledge may become solidified in early adulthood.

In future research, interpersonal scripts for dissatisfaction should be examined across a broader array of relationship contexts, such as marital or cohabitation relationships (heterosexual and same-sex), familial relationships, and cross-sex friendships. This avenue of investigation seems particularly important, given our finding that dissatisfaction scripts follow different tracks in different relationships. It would also be beneficial to examine scripts for specific kinds of dissatisfaction in relationships. It is possible that scripts for highly charged kinds of dissatisfaction (e.g., anger and conflict) would differ from scripts for less intense, but still highly dissatisfying, experiences such as disengagement or relational boredom. In the latter case, scripts for romantic relationships might be more passive and, hence, begin to resemble the scripts held for friendships.

The present findings are also limited to a North American cultural context. Dissatisfaction scripts are likely to differ, depending on the cultural construction of romantic relationships and friendships. For example, Sinclair and Fehr (2005) found that an independent self-construal was associated with the active voice response to conflict issues in romantic relationships whereas an interdependent self-construal was associated with loyalty. Thus, interpersonal scripts for dissatisfaction in romantic relationships might involve greater passivity in collectivist cultures than individualist cultures (see also Yum, 2004). Friendship scripts also might vary by culture. For example, Adams and Plaut (2003) reported that in Ghana, West Africa, social norms dictate that one exercises extreme caution in selecting friends because in a highly interdependent culture, it is difficult to extricate oneself from relationships. Whether these cultural norms would translate into expectations that dissatisfaction be approached in a more active manner (in hopes of maintaining the relationship) or even greater neglect (the friendship is a “done deal,” no matter how dissatisfying) is an intriguing question for future research.

Another limitation is that we relied on self-reports. One avenue for future research is to bring romantic partners and friend pairs into the lab and assess how closely expected partner/friend reactions map on to the other person’s actual responses during an interaction. Reliance on self-reports also limits the findings to relational knowledge held at a conscious level. In doing so, we are not suggesting that people always consciously consult their scripts. Baldwin and others have argued that people can also be unconsciously influenced by scripts (e.g., Baldwin, 1992, 2005; Holmes, 2000). Thus, an important next step to examine whether interpersonal scripts for dissatisfaction are available at a more unconscious level. For instance, a reaction time study could be designed to assess whether people are quicker to respond to negative outcomes in the face of active responses to dissatisfaction in friendships than romantic relationships.

Finally, we studied people’s expectations for responses to dissatisfaction rather than focus on how they have actually responded (based on memories of a specific event) or how they will respond. Our findings suggest that friends “let things be” whereas romantic partners “take action” when responding to dissatisfaction. One pressing question is whether these scripts differentially affect relational well-being. That is, are passive responses to dissatisfaction as effective as active ones in friendships? The outcomes of exit-voice-loyalty-neglect responses in romantic relationships have been studied (Drigotas et al. 1995; Overall et al., 2010). A future direction is to extend this research to friendships.

In conclusion, this research shows that people possess sophisticated knowledge about self responses to dissatisfaction, expected close other reactions, and, importantly, contingencies between self and other. Moreover, they also have knowledge of how issues are likely to turn out (i.e., anticipated outcomes) based on how a close other responds to their expression of dissatisfaction. These scripts are complex and nuanced, depending on factors such as relationship type. We found that the scripts for dissatisfaction in friendships can be summed up as “do nothing and things will be OK,” whereas the scripts for romantic relationships can be summed up as “do something or things won’t be OK.” Our findings lend empirical support to Baldwin’s (2005) cogent observation that “the human mind is profoundly attuned to the contingencies of social life” (p. xi).

Footnotes

Notes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a Social Sciences and Humanities Research Council Research Grant awarded to the second author.

References

Abelson

R. P.

(1981). Psychological status of the script concept. American Psychologist, 36, 715–729.

Adams

Plaut

V. C.

(2003). The cultural grounding of personal relationship: Friendship in North American and West African worlds. Personal Relationships, 10, 333–347.

Afifi

W. A.

Guerrero

L. K.

(1998). Some things are better left unsaid II: Topic avoidance in friendships. Communication Quarterly, 46, 231–249.

Aron

E. N.

Smollan

(1992). Inclusion of other in the self scale and the structure of interpersonal closeness. Journal of Personality and Social Psychology, 63, 596–612.

Arriaga

X. B.

Rusbult

C. E.

(1998). Standing in my partner’s shoes: Partner perspective taking and reactions to accommodative dilemmas. Personality and Social Psychology Bulletin, 24, 927–948.

Baldwin

M. W.

(1992). Relational schemas and the processing of social information. Psychological Bulletin, 112, 461–484.

Baldwin

M. W.

(2005). Interpersonal cognition. New York, NY: Guilford Press.

Baldwin

M. W.

Meunier

(1999). The cued activation of attachment relational schemas. Social Cognition, 17, 209–227.

Baxter

L. A.

Montgomery

B. M.

(1996). Relating: Dialogues and dialectics. New York, NY: Guilford Press.

10.

Baxter

L. A.

Wilmot

W. W.

(1985). Taboo topics in close relationships. Journal of Social and Personal Relationships, 2, 253–269.

11.

Baxter

L. A.

Wilmot

W. W.

Simmons

C. A.

Swartz

(1993). Ways of doing conflict: A folk taxonomy of conflict events in personal relationships. In Kalbfleisch

P. J.

(Ed.). Interpersonal communication. Hillsdale, NJ: Lawrence Erlbaum Associates.

12.

Birditt

K. S.

Fingerman

K. L.

(2005). Do we get better at picking our battles? Age group differences in descriptions of behavioral reactions to interpersonal tensions. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 60, 121–128.

13.

Canary

D. J.

Lakey

(2013). Strategic conflict. New York, NY: Routledge.

14.

Canary

D. J.

Spitzberg

B. H.

(1987). Appropriateness and effectiveness perceptions of conflict strategies. Human Communication Research, 14, 93–118.

15.

Canary

D. J.

Stafford

Hause

K. S.

Wallace

L. A.

(1993). An inductive analysis of relational maintenance strategies: Comparisons among lovers, relatives, friends, and others. Communication Research Reports, 10, 5–14.

16.

Davis

K. E.

Todd

M. J.

(1985). Assessing friendship: Prototypes, paradigm cases and relationship description. In Duck

Perlman

(Eds.). Understanding personal relationships: An interdisciplinary approach. (pp. 17–38), London: SAGE.

17.

Drigotas

S. M.

Whitney

G. A.

Rusbult

C. E.

(1995). On the peculiarities of loyalty: A diary study of responses to dissatisfaction in everyday life. Personality and Social Psychology Bulletin, 21, 596–609.

18.

Fehr

(1996). Friendship processes. Thousand Oaks, CA: SAGE.

19.

Fehr

(2004). Intimacy expectations in same-sex friendships: A prototype interaction- pattern model. Journal of Personality and Social Psychology, 86, 265–284.

20.

Fehr

Baldwin

Collins

Patterson

Benditt

(1999). Anger in close relationships: An interpersonal script analysis. Personality and Social Psychology Bulletin, 25, 299–312.

21.

Fehr

Harasymchuk

(2005). The experience of emotion in close relationships: Toward an integration of the emotion-in-relationships and interpersonal scripts models. Personal Relationships, 12, 181–196.

22.

Gaines

S. O.

Jr. Work

Johnson

Youn

M. S. P.

Lai

(2000). Impact of attachment style and self-monitoring on individuals’ responses to accommodative dilemmas across relationship types. Journal of Social and Personal Relationships, 17, 767–789.

23.

Gergen

K. J.

Gergen

M. M.

(1988). Narrative and the self as relationship. Advances in Experimental Social Psychology, 21, 17–56.

24.

Guerrero

L. K.

Afifi

W. A.

(1995). What parents don’t know: Topic avoidance in parent-child relationships. In Socha

Stamp

(Eds.). Parents, children, and communication: Frontiers of theory and research, (pp. 219–245). Mahwah, NJ: Lawrence Erlbaum Associates.

25.

Harasymchuk

Fehr

(2010). A script analysis of relational boredom: Causes, feelings, and coping strategies. Journal of Social and Clinical Psychology, 29, 988–1019.

26.

Holmberg

MacKenzie

(2002). So far, so good: Scripts for romantic relationship development as predictors of relational well-being. Journal of Social and Personal Relationships, 19, 777–796.

27.

Holmes

J. G.

(2000). Social relationships: The nature and function of relational schemas. European Journal of Social Psychology, 30, 447–495.

28.

Honeycutt

J. M.

Bryan

S. P.

(2011). Scripts and communication for relationships. New York, NY: Peter Lang.

29.

Kammrath

L. K.

Dweck

(2006). Voicing conflict: Preferred conflict strategies among incremental and entity theorists. Personality and Social Psychology Bulletin, 32, 1497–1508.

30.

Miller

J. B.

(1991). Women’s and men’s scripts for interpersonal conflict. Psychology of Women Quarterly, 15, 15–29.

31.

Oliker

S. J.

(1989). Best friends and marriage: Exchange among women . Berkley: University of California Press.

32.

Overall

N. C.

McNulty

J. K.

(2017). What type of communication during conflict is beneficial for intimate relationships? Current Opinion in Psychology, 13, 1–5.

33.

Overall

N. C.

Sibley

C. G.

(2008). When accommodation matters: Situational dependence within daily interactions with romantic partners. Journal of Experimental Social Psychology, 44, 238–249.

34.

Overall

N. C.

Sibley

C. G.

Travaglia

L. K.

(2010). Loyal but ignored: The benefits and costs of constructive communication behavior. Personal Relationships, 17, 127–148.

35.

Planalp

(1987). Interplay between relational knowledge and events. In Burnett

McGhee

Clarke

(Eds.), Accounting for relationships: Explanation, representation and knowledge (pp. 175–191). London: Methuen.

36.

Rusbult

C. E.

Bissonnette

V. L.

Arriaga

X. B.

Cox

C. L.

(1998). Accommodation processes during the early years of marriage. In Bradbury

T.N.

(Ed.), The developmental course of marital dysfunction (pp. 74–113). New York, NY: Cambridge University Press.

37.

Rusbult

C. E.

Johnson

D. J.

Morrow

D. G.

(1986a). Determinants and consequences of exit, voice, loyalty and neglect: Responses to dissatisfaction in adult romantic involvements. Human Relations, 39, 45–63.

38.

Rusbult

C. E.

Johnson

D. J.

Morrow

G. D.

(1986b). Impact of couple patterns of problem solving on distress and nondistress in dating relationships. Journal of Personality and Social Psychology, 50, 744–753.

39.

Rusbult

C. E.

Morrow

G. D.

Johnson

D. J.

(1987). Self-esteem and problem-solving behaviour in close relationships. British Journal of Social Psychology, 26, 293–303.

40.

Rusbult

C. E.

Verette

Whitney

G. A.

Slovik

L. F.

Lipkus

(1991). Accommodation processes in close relationships: Theory and preliminary empirical evidence. Journal of Personality and Social Psychology, 60, 53–78.

41.

Rusbult

C. E.

Yovetich

N. A.

Verette

(1996). An interdependence analysis of accommodation processes. In Fletcher

Fitness

(Eds.). Knowledge structures in close relationships: A social psychological approach, (pp. 63–90). New York, NY: Psychology Press.

42.

Rusbult

C. E.

Zembrodt

I. M.

(1983). Responses to dissatisfaction in romantic involvements: A multidimensional scaling analysis. Journal of Experimental Social Psychology, 19, 274–293.

43.

Rusbult

C. E.

Zembrodt

I. M.

Gunn

L. K.

(1982). Exit, voice, loyalty, and neglect: Responses to dissatisfaction in romantic involvements. Journal of Personality and Social Psychology, 43, 1230–1242.

44.

Sillars

A. L.

(1980). Attributions and communication in roommate conflicts. Communication Monographs, 47, 180–200.

45.

Spitzberg

B. H.

Cupach

W. R.

(2013). The dark side of close relationships. New York: Routledge.

46.

Sinclair

Fehr

(2005). Voice versus loyalty: Self-construals and responses to dissatisfaction in romantic relationships. Journal of Experimental Social Psychology, 41, 298–304.

47.

Wright

P. H.

(1985). The Acquaintance description form. In Duck

Perlman

(Eds.). Understanding personal relationships: An interdisciplinary approach (pp. 39–62), London: SAGE.

48.

Yum

Y. O.

(2004). Culture and self-construal as predictors of responses to accommodative dilemmas in dating relationships. Journal of Social and Personal Relationships, 21, 817–835.