Abstract
How people interpret the meaning of minor relationship transgressions can impact broader relationship well-being. It is proposed that picturing relationship transgressions from a third-person (vs. first-person) visual perspective prompts people to think of them in the context of their chronic relationship beliefs and goals. In doing so, individuals who are relatively anxious about their relationships become more insecure, whereas less anxious individuals find reassurance. In Study 1 participants pictured a transgression they committed against their partner. Individuals high in attachment anxiety made less positive evaluations of their relationships when picturing the event using a third-person rather than first-person perspective. Similar results were found when participants recalled transgressions committed by their partners against them (Study 2). These results have implications for understanding how partners move forward in their relationships after transgressions.
Even in satisfying close relationships, it is inevitable that partners will behave badly at least on occasion. How people interpret the meaning of these transgressions, and how they subsequently behave towards their partner, can have a significant impact on the well-being of the relationship. Individuals who are high in attachment anxiety are more vigilant about possible rejection and have a reduced threshold for detecting threats to security (Fraley & Shaver, 2000). They also experience greater emotional distress when recalling partner transgressions (Collins, Ford, Guichard, & Allard, 2006; Feeney, 2004), have a harder time repressing negative affect (Mikulincer & Orbach, 1995), and endorse more relationship-threatening attributions for a partner’s negative behavior (Collins et al., 2006; Mikulincer, 1998). As such, individuals high in attachment anxiety tend to take relationship threats to heart.
The behavioral responses of those high in attachment anxiety may be particularly problematic for relationship well-being. Individuals who doubt their partners’ love and acceptance often respond to relationship threats in a self-protective manner, becoming cold and critical toward their partners (Mikulincer, 1998; Murray, Rose, Bellavia, Holmes, & Kusche, 2002). Devaluing the source of rejection helps protect these individuals from further psychological hurt, but it also alienates their partners (Campbell, Simpson, Boldry, & Kashy, 2005; Murray, Griffin, Rose, & Bellavia, 2003; Powers, Pietromonaco, Gunlicks, & Sayer, 2006, Simpson, Rholes, and Phillips, 1996). Such overreactions can decrease their partners’ satisfaction with the relationship over the long term (Murray et al., 2003; Van Orden & Joiner, 2006). More secure individuals, on the other hand, tend to compensate for relationship threats by affirming and drawing closer to their partners. This response appears to have restorative effects on the relationship (Murray, Holmes, & Collins, 2006).
A recent review showed that anxiously attached individuals do not exhibit attachment-related distress and corresponding destructive relationship behavior unvaryingly, but rather in contexts in which cues signal potential rejection or relationship loss. Attachment anxiety seems not to moderate cognitive, affective, or behavioral responses in contexts where these cues are not present or perceived (Campbell & Marshall, 2011). We agree with the authors of this review that it is important for researchers to “test the conditions where individual differences in anxious attachment are strongly, or weakly, associated with relationship-relevant thoughts, feelings, and behaviors” (Campbell & Marshall, 2011, p. 1243). The studies reported in this article answer that call, by employing a manipulation that has been previously shown to influence the impact of individual differences on people’s reactions to autobiographical events. In our studies, all participants recalled a relationship transgression. We varied the relational context activated during recall by manipulating visual perspective to see whether that would moderate the influence of attachment anxiety on subsequent relationship evaluations.
Visual perspective in mental imagery
When people recall autobiographical events, their visual imagery may assume either a first-person or a third-person perspective (Nigro & Neisser, 1983). With a first-person perspective, people see the event from the same visual perspective they had when it originally occurred. With a third-person perspective, people see the event from an observer’s visual perspective: they see themselves as well as their surroundings in the scene. People report that they experience these perspectives spontaneously during autobiographical recall (Nigro & Neisser, 1983). Most people are able to control the perspective they use to recall past events in response to experimental instructions (e.g., Frank & Gilovich, 1989; Libby, Eibach, & Gilovich, 2005).
The visual perspective people use to picture an event influences the way people process events and make meaning of them (Libby & Eibach, 2011a). The first-person perspective functions as a visual analogue of the experiential self or “I.” First-person imagery invokes a bottom-up processing style in which people construe events in terms of concrete experiential details and associative evaluations (e.g., Libby, Shaeffer, & Eibach, 2009; Libby, Valenti, Hines, & Eibach, 2014). The third-person perspective functions as a visual analogue of the conceptual self or “me.” Third-person imagery broadens people’s focus by invoking a top-down processing style in which they construe events abstractly, in terms of relevant personal beliefs and theories—for example, about their traits, broader goals, and life themes (e.g., Libby et al., 2014; Libby, Valenti, Pfent, & Eibach, 2011; Valenti, Libby, & Eibach, 2011). These findings suggest that people who recall a relationship transgression from a third-person perspective will reflect on the event within the broader context of more abstract, chronic relationship beliefs and goals.
For anxiously attached individuals, this broader context includes serious concerns that their partner does not truly value them (Collins & Read, 1990; Mikulincer, Shaver, Sapir-Lavid, & Avihou-Kanza, 2009). Reflecting on the broader meaning of a transgression from the third-person perspective, individuals high in attachment anxiety are expected to interpret it as one more sign of their partner’s waning affections, confirming their overall fears. These insecurities should result in self-protective distancing from their partner. In contrast, for individuals low in attachment anxiety, the broader context induced by a third-person perspective includes the confident belief that they are loved and valued (Collins & Read, 1990; Mikulincer et al., 2009). When a specific transgression is interpreted in light of such relationship-affirming sentiments, it should lose its emotional punch and have little influence on people’s attitudes toward their relationship. Consequently, we predicted that the difference between high- and low-anxiety individuals’ reactions to relationship transgressions would be more pronounced when they pictured a transgression from a third-person as opposed to a first-person visual perspective. Essentially, when individuals high in attachment anxiety adopt a third-person perspective, the abstract construal that results is hypothesized to make them “reactive” to specific negative events, tying their satisfaction with their relationship as a whole to a particular experience.
The current studies
The studies reported in this article make an important contribution to the literature in at least three ways. First, they experimentally test the theoretical assumption of both attachment theory (Mikulincer & Shaver, 2003) and risk regulation theory (Murray et al., 2006) that differences between more or less anxiously attached individuals will be magnified to the extent that a local event is considered in a wider, more abstract frame of reference. The hypothesis would receive support if a manipulation of visual perspective—which has been demonstrated to differentially promote abstract reflection—affected reactions to relationship transgressions. Such findings would begin to answer Campbell and Marshall’s (2011) query of the conditions under which individual differences in attachment orientations do or do not predict relationship outcomes.
Second, the present research seeks to enhance our understanding of the functions of imagery. Although psychologists from diverse theoretical perspectives acknowledge the importance of mental imagery to relational knowledge (e.g., Baldwin, 1992; Baldwin & Holmes, 1987; Magai, 1999; Main, 1991), they have not studied whether different types of mental imagery might vary in their power to activate relational knowledge. By examining how the type of imagery people experience when they recall relationship events influences their interpretations of those events, we can understand when people will or will not apply their broader relational knowledge to interpret specific episodes.
Third, our studies extend the work of Kross, Ayduk, and colleagues (see Kross & Ayduk, 2011, for a review) which has reliably demonstrated that a “self-distanced” perspective, which involves third-person imagery, decreases distress reactions (both self-reported emotions and physiological indicators) when recalling upsetting interpersonal events, compared to a “self-immersed” perspective, which involves first-person imagery. At first blush, our hypothesis that third-person visual imagery will increase negative outcomes for individuals high in attachment anxiety may seem inconsistent with this work. However, a closer look at this self-distanced instructions shows that manipulation of self-distanced versus self-immersed perspectives confounds visual perspective with phrasing that implies personal distance from the past event. Whereas we ask participants to simply view the transgression from a first-person or third-person visual perspective, with no further instructions, the Kross, Ayduk, and colleagues' studies ask participants to either immerse themselves in the event (“go back to the time and place of the experience and relive the situation as if it were happening to you all over again”) or distance themselves from the event (“watch the conflict unfold as if it were happening all over again to the distant you”; emphasis added), as well as think about the reasons underlying their (or their distant self’s) feelings (Ayduk & Kross, 2008; Kross, Ayduk, & Mischel, 2005).
The instructions framing the event as involving “the distant you” may encourage participants to adopt a theory of self-change or personal growth in relation to the pictured event (Libby et al., 2005), which may explain why participants are able to more constructively work through the past event in this condition. In other words, it may be necessary for third-person imagery to be paired with a constructive interpretive frame to produce the adaptive outcomes that typically emerge with this methodology.
In their 2011 review, Kross and Ayduk called for future research to examine when their methodology may be harmful or ineffective. We suggest that it may be harmful for individuals high in attachment anxiety when third-person imagery is not associated with self-distancing language, as broader concerns about the relationship now have room to fester. Thus, we expect that when such individuals are not provided with a constructive self-distancing frame by the experimenter, third-person imagery will only serve to amplify the effects of the negative interpretive frame they are predisposed to apply to distressing relational events. Previous research demonstrates an analogous effect in which third-person imagery increases shame about failure for low self-esteem individuals, whose general self-beliefs lead them to take failure to heart (Libby et al., 2011).
In the present two studies, we manipulated the visual perspective participants used to picture transgression events in their romantic relationships and then asked for participants’ current relationship perceptions to determine the extent to which participants took the transgression to heart—that is, allowed an isolated event to affect their current relationship evaluations. Our hypotheses were that: Attachment anxiety would be more strongly related to relationship evaluations in the third-person than in the first-person condition. Participants high in attachment anxiety would have more negative outcomes in the third-person compared to the first-person condition.
We expected the predicted patterns to emerge regardless of whether participants were the perpetrators (Study 1) or the victims (Study 2) of transgressions because attachment anxiety has been shown to moderate the strength of reactions to both types of events (Mikulincer, & Shaver, 2005).
Study 1
In Study 1, we manipulated the visual perspective that participants used to recall a relationship transgression they committed against their partner. We investigated the implications for participants’ evaluations of their relationships in general. Individuals high in attachment anxiety tend to react to their partner’s relationship-relevant distress with shame and despair, as they may attribute their hurtful behavior to personal deficiencies (Mikulincer, & Shaver, 2005). When recalling a time they hurt their partner, we expected that a visual perspective manipulation would influence the effects of attachment anxiety on subsequent relationship perceptions by varying the extent to which they recruit the broader context of their relationship beliefs. More specifically, the effect of attachment anxiety on relationship evaluations will be stronger for participants induced to use third-person imagery rather than first-person imagery (Hypothesis 1), and individuals high in attachment anxiety will find more cause for concern and subsequently feel worse about their relationship in the third-person condition than in the first-person condition (Hypothesis 2).
Method
Participants and procedure
Eighty-eight undergraduates (27 males and 61 females) in romantic relationships participated in a study of memory and relationships in exchange for course credit. Age ranged from 17 to 25 with a mean of 19 years. Relationship length ranged from 1.5 to 87 months with a mean of 18 months. The majority of participants (64) were in an “exclusive dating” relationship. The rest were “casually dating” (7), “living together” (8), “engaged” (8), or “married” (1).
For each experimental session, 1 to 4 participants worked independently on their questionnaire packages. First they completed measures of individual differences. Next they recalled a time they had upset their partner. They were randomly assigned to visualize the event from a first-person or third-person visual perspective. Then they answered questions concerning their feelings about the event and about their relationship in general, and completed demographic items. Finally, they described a recent positive event with their partner so that they would leave the lab feeling good about their relationship. The same general procedure was followed in Study 2.
Materials
Attachment anxiety
Participants completed an adapted version of Brennan, Clark, and Shaver’s (1998) Experience in Close Relationships scale. This scale assesses two dimensions of attachment style: anxiety (e.g., “I need a lot of reassurance that I am loved by my partner;” α = .83) and avoidance (e.g., “I prefer not to show my partner how I feel deep down;” α = .90). We focused on attachment anxiety because it is anxiety that is associated with individuals’ threshold for threat detection (Fraley & Shaver, 2000). 1
We reworded the questions to reflect participants’ current relationship, instead of all romantic relationships. We focused on the current relationship because individuals’ attachment styles change across relationship partners (La Guardia, Ryan, Couchman, & Deci, 2000; Pierce & Lydon, 2001). Also, relationship-specific outcomes are better predicted by relationship-specific than by general attachment styles (Holmes & Cameron, 2005; Klohnen, Weller, Luo, & Choe, 2005). We also shortened this 36-item scale to 22 items based on feedback from participants in other studies that the 36-item version was too long and repetitive. We were concerned that participants’ waning attention would add error to our data.
Event manipulation
All participants were then asked to “remember a time that you upset your partner. It may or may not have been intentional, but you caused your partner to feel hurt, disappointed, and/or angry. For example, you were inconsiderate of your partner’s feelings, you let your partner down, or you cheated on your partner.” They wrote down a few words to identify that event and noted how long ago it occurred.
Next, participants turned the page to read the manipulation of visual perspective (adapted from Libby et al., 2005). In the first-person condition (N = 43), individuals were asked to visualize the event “from the same visual perspective that you originally had, in other words, looking out at your surroundings through your own eyes.” They were encouraged to take a few moments to do this, and answered five yes-or-no questions designed to reinforce the image (e.g., “can you see your surroundings?” “Can you see who you are with?”). Participants in the third-person condition (N = 45) were asked to visualize the event “from an observer’s visual perspective; in other words, so that you can see yourself in the memory, as well as your surroundings.” They answered a comparable set of yes-or-no questions to reinforce the image (e.g., “can you see yourself?” “Can you see the expression on your face?”). Participants in both conditions were asked to hold their respective visual images in mind as they wrote a description of the event and while answering questions about the event on the pages following. 2
Relationship quality
In one set of ratings, participants were instructed to consider how they felt about their relationship “right now.” They rated 23 items on 7-point scales (1 = strongly disagree to 7 = strongly agree), including felt security (e.g., “my partner loves and accepts me unconditionally”), commitment (e.g., “I am very committed to my relationship”), and satisfaction (e.g., “I am extremely happy with my current romantic relationship”). These items formed an index of relationships ratings with α = .94 (adapted from Marigold, Holmes, & Ross, 2007).
In another set of ratings, participants rated their partners on 22 specific qualities (e.g., “warm,” “attractive,” “controlling,” and “complaining”) on a scale from 1 (not at all characteristic) to 9 (completely characteristic) (the Interpersonal Qualities Scale; Murray, Holmes, & Griffin, 1996). Negative items were reverse scored and then averaged with the positive items, α = .85.
Because the relationship ratings and partner ratings were highly correlated (r(88) = .60, p < .001) and showed similar patterns of results, we standardized the two measures and combined them into one overall composite of relationship quality, α = .93.
Results and discussion
Relationship quality was regressed on perspective condition (first person or third person), attachment anxiety, and the Perspective × Anxiety interaction. Perspective was effect-coded and anxiety was centered to make the mean equal to zero. The two main effects and the interaction term were entered simultaneously. Simple effects were calculated by recentering anxiety at 1 SD above the mean (for high anxiety) and 1 SD below the mean (for low anxiety) (as per Aiken & West, 1991). Analyses revealed no significant interactions between condition and gender on any of the dependent variables in the two studies, so gender will not be discussed further. 3
There was a main effect of attachment anxiety, β = −.60, t(84) = −6.95, p < .001, that was qualified by the predicted Perspective × Anxiety interaction, β = .18, t(84) = 2.11, p = .038 (see Figure 1). Overall, more anxiously attached participants reported lower relationship quality; however, consistent with our predictions, anxiety was a stronger predictor of relationship quality when participants pictured the transgression from the third-person, β = −.78, t(84) = −6.10, p < .001, than from the first-person perspective, β = −.42, t(84) = −3.61, p < .001. Also as predicted, participants with high attachment anxiety perceived lower relationship quality when they used third-person perspective than when they used the first-person perspective to picture the transgression, β = −.30, t(84) = −2.51, p = .014. Low anxiety participants’ ratings of relationship quality did not differ as a function of imagery perspective (β = .06, ns)
Relationship quality (standardized scale) as a function of condition and anxiety (±1 SD): Study 1.
Thus, as predicted, chronic differences between high and low anxiously attached individuals’ ratings of relationship quality were larger when adopting a third-person perspective on the memory of a transgression against a romantic partner. As well, participants high in attachment anxiety reported significantly lower relationship quality in the third-person perspective condition than in the first-person perspective condition. These results are consistent with previous work demonstrating that third-person imagery invokes top-down processing in which people process events abstractly in the context of relevant general beliefs (Libby et al., 2011, 2014). In the present work, participants who used a third-person perspective to reflect on the transgression were more inclined to draw on the context of their general relationship beliefs. When participants high in attachment anxiety pictured the event from the third-person perspective they seemed especially likely to take the transgression to heart, apparently seeing it as one more instance that confirmed their general belief about serious problems in their relationship.
The effects of perspective were not significant for participants low in attachment anxiety. Their confident and positive beliefs about their relationship left little room for further strengthening their attitudes. Also, perpetrators who were less anxiously attached may have been particularly likely to affirm their relationship and view it positively when recalling a transgression (see Murray et al., 2006), regardless of the visual perspective they adopted. Moreover, perpetrators tend to be more optimistic about their relationships than do victims, being especially motivated to assure themselves that the transgression did not cause significant harm to the relationship (Cameron, Ross, & Holmes, 2002).
Study 2
In Study 2, we examined the effects of visual imagery perspective in reaction to relationship transgressions in which the participant was the victim. Individuals high in attachment anxiety tend to react more negatively to a partner’s hurtful behavior. They experience more resentment, sadness, and despair (Mikulincer & Shaver, 2005) and question their partners’ love (Collins, 1996; Feeney, 2004). We expected the effects of attachment anxiety on subsequent relationship evaluations would be larger when using a third-person than a first-person visual perspective as in Study 1. Although there was no effect of condition at low levels of attachment anxiety when participants were the transgressors in Study 1, we expected there might be such an effect when participants were the victims. Recalling a victimization event is likely more threatening and thus being able to recruit their chronic security and positive relationship beliefs would be particularly beneficial for individuals low in attachment anxiety.
In Study 2, we made the same predictions as Study 1: (1) attachment anxiety would be more strongly related to relationship evaluations after thinking about the event using a third-person rather than a first-person visual perspective and (2) participants high in attachment anxiety will have more negative outcomes after using third-person imagery than after using first-person imagery, whereas participants low in attachment anxiety will either have more positive outcomes using third-person imagery, or show no difference between imagery conditions. Study 2 also included a control condition in which participants received no imagery instructions, providing a baseline against which to evaluate responses in the two perspective conditions.
Method
Participants and procedure
Ninety-two undergraduates (44 males, 47 females, 1 unreported) in romantic relationships participated in a study of memory and relationships in exchange for course credit. Age ranged from 18 to 29 with a mean of 20 years. Relationship length ranged from 2 to 67 months, with a mean of 19 months. The majority of participants (80) were in an exclusive dating relationship. The rest were either casually dating (6), living together (5), or married (1). The procedure was the same as Study 1 except that the remembered transgression was committed by the partner: “Remember a time when your partner said or did something that made you feel hurt and/or angry. For example, your partner cancelled plans with you and went out with his/her friends instead; or you caught your partner flirting with someone else.” Participants were randomly assigned to the first-person condition (N = 31), third-person condition (N = 30), or control condition (N = 31). For the first-person and third-person conditions, visualization instructions were the same as in Study 1.
Participants in the control condition were not given instructions on how to visualize the event. They were simply asked to “take a few moments to visualize this event in your mind, then briefly describe it below.” They then responded to a forced-choice item, “From which perspective did you see the event in memory that you just described?” after a brief explanation of what constituted a first-person and third-person memory (using similar wording as the first-person and third-person manipulations). All participants then rated their current distress about the event, and finally they rated their relationship quality.
Materials
Current distress
Four items assessed participants’ feelings about the event on a scale ranging from 1 (not at all) to 9 (extremely). They rated how upset they currently feel about it and how intense their emotions are (α = .89, M = 4.13, SD = 2.28).
Relationship ratings
Similar to Study 1, participants rated 23 items tapping into their current feelings about their relationship on a 1 to 5 scale as well as rating 16 of their partners’ personal qualities. Partner ratings and relationship ratings were again highly correlated, r(90) = .52, p < .001, and showed similar patterns of results, so we standardized the two measures and combined them into one overall composite of relationship quality, α = .85.
Results and discussion
The regression analyses were identical to those conducted in Study 1, except that the perspective factor had three levels, including the control group, and was thus effect-coded with two vectors.
Control condition
Among participants in the control condition, 13 (42%) reported that they spontaneously adopted a third-person perspective when recalling the event, and 18 reported a first-person perspective. Participants adopting a third-person perspective tended to report higher anxiety than those adopting a first-person perspective, but the difference was not significant, F(28) = 2.51, p = .12.
Current distress
A main effect of attachment anxiety, β = .34, t(86) = 3.22, p = .002, was qualified by a Perspective × Anxiety interaction, β = −.29, t(86) = −2.30, p = .024, and β = .32, t(86) = 2.53, p = .013 (see Figure 2). Overall, more anxiously attached participants reported greater distress when thinking about their partner’s transgression. However, consistent with our predictions, anxiety was a considerably stronger predictor of distress in the third-person perspective condition, β = .75, t(86) = 3.46, p < .001, than in the first-person perspective (β = .01, ns) or control conditions (β = .24, ns). Also, as predicted, participants high in attachment anxiety experienced significantly more distress in the third-person than in the first-person, β = .38, t(86) = 2.02, p = .047, or control conditions, β = .41, t(86) = 2.33, p = .022. In contrast, participants low in attachment anxiety reported significantly less distress in the third-person condition than in the first-person condition, β = −.32, t(86) = −2.03, p = .046, though they did not differ significantly from control participants (β = .06, ns).
Current distress (1–9 scale) as a function of condition and anxiety (±1 SD): Study 2.
Relationship quality
A main effect of attachment anxiety, β = −.28, t(86) = −2.56, p = .012, was qualified by a Perspective × Anxiety interaction, β = −.27, t(86) = −2.07, p = .041 (see Figure 3). Overall, more anxiously attached participants reported lower relationship quality when thinking about their partner’s transgression. However, consistent with our predictions, attachment anxiety was a considerably stronger predictor of relationship quality in the third-person perspective condition, β = −.62, t(86) = −2.80, p = .006, than in the first-person perspective (β = −.06, ns) or control conditions (β = −.15, ns). In addition, participants high in attachment anxiety reported lower relationship quality in the third-person condition than in the control condition, β = .36, t(86) = −1.98, p = .051, though not significantly different from the first-person condition (β = .24, ns) . Participants low in attachment anxiety reported somewhat higher relationship quality in the third-person perspective condition than in the first-person perspective condition, β = .30, t(86) = 1.84, p = .070. Neither of the other simple effects reached significance (p > .22).
Relationship quality (standardized scale) as a function of condition and anxiety (±1 SD): study 2.
Replicating Study 1, and consistent with the notion that third-person imagery guides people to interpret pictured events in terms of more general relationship beliefs, chronic differences between individuals relatively high or low in attachment anxiety were greater when using third-person as opposed to first-person imagery in Study 2. Further, when participants high in attachment anxiety pictured a partner’s transgression from the third-person as opposed to first-person perspective, they reported greater distress about the specific transgression and lower relationship quality in general. Participants low in attachment anxiety reported less distress about the transgression in the third-person perspective condition than in the first-person perspective condition, although this buffering effect of a broader, affirming perspective on current feelings did not significantly affect their already very positive perceptions of relationship quality.
General discussion
Theorists have previously hypothesized that differences between secure and insecure individuals will be magnified when conditions promote the application of their more abstract, chronic relationship beliefs. Recent research on visual imagery suggested that this assumption could be directly tested by experimentally manipulating a person’s visual perspective on recalled events: third-person imagery promotes the application of abstract beliefs in interpreting pictured events. Indeed, the results of two studies, one focusing on a transgression committed by the participant, the other on a transgression committed by the partner, converged to show that individual differences in self-reported attachment anxiety predict people’s interpretations of relationship transgressions more strongly when they visualize those transgressions from a third-person perspective than when they visualize them from a first-person perspective.
When participants high in attachment anxiety recalled relationship transgressions using a third-person visual perspective, regardless of whether they were the perpetrator (Study 1) or victim (Study 2), they took the relationship threat to heart, as an instance that reminded them of their general concerns. After using the third-person perspective, individuals high in attachment anxiety reported lower relationship quality than if they had used a first-person visual perspective (Study 1) or if they had not been instructed to visualize the transgression from any particular perspective (control condition, Study 2). Individuals high in attachment anxiety also experienced more distress over their partner’s transgression when using a third-person rather than first-person visual perspective (Study 2). On the other hand, individuals low in attachment anxiety experienced less distress over their partner’s transgression when using a third-person, rather than first-person visual perspective. The general beliefs that individuals low in attachment anxiety tend to hold about how much they are loved provide a buffer against isolated negative incidents (Murray et al., 2006), and third-person imagery may enhance their use of this buffer.
Previous research has shown that when people consider past, future, or hypothetical behaviors from a third-person visual perspective they are more likely to invoke the broader context of their personality, goals, and motives (Frank & Gilovich, 1989; Libby et al., 2011, 2014; Storms, 1973; Vasquez and Buehler, 2007). Thus, the third-person perspective is associated with larger effects of individual differences in self-beliefs on peoples’ interpretations of life events. The current findings suggest that in a relationship context, visual perspective functions to determine the extent to which one’s general theories about their relationship are brought to bear on interpretations of a specific relationship event. A relationship transgression reviewed from a first-person perspective seems to be more of a concrete, isolated incident, and so no elaboration is called for. On the other hand, when reviewing the conflict from a third-person visual perspective people are more likely to reflect on it in the context of their broader relationship beliefs and goals. Doing so appears to reinforce individuals’ beliefs about their relationships, further undermining the sense of security of individuals high in attachment anxiety, while protecting lower anxiety individuals.
This present work makes a novel contribution to the literature in three ways: by specifying one of the conditions under which individual differences in attachment anxiety strongly or weakly influence responses to perceived relationship threats (cf. Campbell & Marshall, 2011); by demonstrating that different types of mental imagery vary in their power to activate relational knowledge; and by parsing out the influence of visual imagery and analytical reasoning from typically used self-distancing manipulations (cf. Kross & Ayduk, 2011). As such, we gain further insight into attachment processes as well as interventions with distressed partners.
One intervention which has recently been shown to have powerful effects involved a 21-min writing exercise in which participants reappraised conflict in their marriage from the perspective of a neutral third party who wanted the best for all involved, including instructions to visually image the conflict through this neutral party’s eyes. This intervention eliminated the decline in marital quality typically observed in couples over a 2-year period (Finkel, Slotter, Luchies, Walton, & Gross, 2013). The authors noted that future research would need to determine exactly how the intervention accomplished its salutary effects, as there were several components to it: a self-distanced psychological perspective (Kross et al., 2005), a third-party visual perspective (Libby & Eibach, 2011a), and an “adaptive framework” (see Libby & Eibach, 2011a, p. 234) of wanting the best for all involved and focusing on the good that could come out of the conflict. The present studies confirm that a third-party visual perspective is not likely to benefit all individuals. Our results support the idea that individuals high in attachment anxiety tend to adopt a relatively non-adaptive framework when taking a third-person visual perspective, which should hinder their recovery from relationship transgressions.
However, we suspect that individuals high in attachment anxiety could be encouraged to adopt a more adaptive framework, potentially reversing the harmful effects of the third-person perspective. Several other studies have demonstrated how individuals high in attachment anxiety in particular may be encouraged to interpret isolated incidents within a more positive, secure context. One such study asked couples to prepare for a conflict discussion by bringing to mind positive aspects of their relationship and good times they have had with their partner. This manipulation eliminated the negative effects of attachment anxiety on emotions experienced during the conflict for both partners (Ben-Naim, Hirschberger, Ein-Dor, & Mikulincer, 2013). Experimentally priming thoughts of supportive attachment figures has also been shown to buffer anxiously attached individuals’ distress reactions to a partner’s hurtful behavior (Shaver, Mikulincer, Lavy, & Cassidy, 2009). It is possible that the beneficial effects of these constructive interventions could be amplified by supplementing them with instructions to visualize the negative event from a third-person perspective.
Future research
In addition to exploring ways to reverse potentially harmful effects of third-person imagery, future research could also more thoroughly investigate which factors determine the perspective one spontaneously adopts in various memories. Individual difference variables such as high public self-consciousness and social anxiety are associated with greater use of a third-person perspective in memory, at least for certain types of events (Coles, Turk, & Heimberg, 2002; Robinson & Swanson, 1993). In the current studies, we did not find that individuals high in attachment anxiety were significantly more likely to naturally adopt a third-person visual perspective, though the data suggested a trend in this direction (p = .12) and the power to detect this relation in the control group of Study 2 was low due to sample size. Other characteristics of the person and memory (e.g., How incongruent is the past self with present self-conceptions? How far in the past is the event?) can influence visual perspective (Libby & Eibach, 2002, 2011b; Nigro & Neisser, 1983; Robinson & Swanson, 1993). In any event, the current research suggests that encouraging individuals high in attachment anxiety to adopt a first-person perspective may reduce distress and enhance ratings of the partner and the relationship.
Another worthwhile direction for future research would be to investigate the potential behavioral manifestations of the emotions and evaluations that participants reported in the present research. Evidence that the visual perspective partners use to picture relationship events influences the interpersonal dynamics in their relationships would contribute to an understanding of the role that imagery plays in relationship processes (cf. Baldwin, 1992; Magai, 1999).
Finally, future research should explore whether the present findings hold up in more diverse populations than the North American students sampled in the present research. Evidence suggests that relationship processes can vary depending on the population (Karney, 2013).
Concluding comments
When people recall transgressions in their relationships, what determines whether they will see the transgression as an isolated event or as an indicator of the broader state of their relationship? Our findings suggest that people’s interpretations of recalled transgressions depend not only on the contents of their chronic relational schemas but also on how they visually reconstruct the scene of the transgression. When they recall the scene from a third-person visual perspective the contents of their relational schemas, specifically their reported attachment anxiety, tend to shape their reactions to the memory. The simple mental trick of changing visual perspective while remembering an event can influence whether individual differences are associated with robust or negligible variations in the meaning of the episode. Meaning may thus be in the visual perspective of the beholder.
Footnotes
Funding
This research was supported by Social Sciences and Humanities Research Council grants awarded to the fourth and fifth authors.