ATTACHMENT ANXIETY AND PARTNER GUILT 239 of each day for three wecks.This method allows predicted that such guilt-induction attempts would be successfulin ss couples'lives and provides a window to how these re sult in the partner feeling guilty.We predictec daily lives (Study 1)andur ngs w n the form we tested this prediction by assessing whether.on days when individuals should experience nd ationship er.b e we pre feeling more hurt compared to individuals lower in anxiety who edicted that the partner's ous indi als of their part s of hut this and exagge ate their hurt more to their partner.We partncrs. would s and beha Our final aim was to investigate whether partner's guilt has we contrasted hurt feelings and guilt-induction the 3-week diary was associated with relationshin satisfactio d hurt nd guilt-inducti gathered 9 months following the completion of the daily re ic the diary peric with lowe dependence,is otivated by 0 f their eaHceuiednheparncrfeligguiy in the hlater satisfacti 98:Lemay 2012N about their relations y life ly associated (see L and d that e ater partner guilt would be typicallyasso ciated w Indeed. individ tense need to er relatio ship satisfaction.How ever.becaus ould fare bette 010:Murray Holmes 009).Thu ower declin or maintain levels of satisfaction cted nd anger detrimental effects on the partner and,thus,part ners o the predicted results could b when they felt is cha rized by a deep dist of other re Method n et a Participan ities to attacm tivity to pleting all phases of the study(an initial session,a 3-week daily we expected hat the liary,and a to in serious romantic relationships (1 married.3%cohabiting. This s Study sedpweieudyi d.2013. mmitted couples to nd the threatening contexts and, when hurt, purposively communicate or exaggerate their hurt to induce guilt in their partner. We also predicted that such guilt-induction attempts would be successful in producing feelings of guilt in the partner. We also tested whether guilt experienced across couple’s daily lives (Study 1) and during conflict-related discussions (Study 2) predicted longitudinal changes in relationship secu￾rity and satisfaction. To the extent that the partner’s guilt reflects the partner’s culpability for hurtful partner transgres￾sions, the more partners tend to feel high levels of guilt across couples’ interactions, the more individuals should experience declines in partner and relationship evaluations. However, be￾cause successfully inducing guilt provides desired evidence of the partner’s care and concern, we predicted that the partner’s guilt would reassure anxious individuals of their partner’s com￾mitment and help them maintain more positive relationship evaluations. Unfortunately, we did not expect that these bene￾fits would extend to anxious intimates’ partners. Instead, be￾cause their induced guilt should be disproportionate to their intentions and behaviors, we expected that when partners of anxious individuals feel greater guilt they would experience sharper declines in satisfaction. In both studies, we contrasted hurt feelings and guilt-induction strategies to two other common responses to relationship threat: anger and hostile behavior. We isolated hurt and guilt-induction as specific responses arising from the dependence and motivation to sustain relationships central to attachment anxiety. In contrast, anger is associated with lower dependence, is motivated by a desire to attain personal control, and triggers hostile behavior that generates distance and reciprocal animosity in the partner (Gott￾man, 1998; Lemay et al., 2012). Nonetheless, despite these oppos￾ing antecedents and consequences, hurt and anger are often strongly associated (see Lemay et al., 2012), and anxious intimates report greater anger and hostility during conflict (Feeney et al., 1994; Mikulincer, 1998; Overall & Sibley, 2009; Simpson et al., 1996). Indeed, anxious individuals’ intense need to obtain close￾ness and reactivity to rejection might simultaneously generate hurt-based guilt-induction attempts and anger-based hostility (Mi￾kulincer et al., 2010; Murray & Holmes, 2009). Thus, we con￾trolled for the links across hurt and anger to illustrate the speci￾ficity of the predicted effects. Finally, we also examined whether the predicted results could be due to other forms of insecurity. Another form of attachment insecu￾rity, called avoidance, is characterized by a deep distrust of others, rigid self-reliance, and withdrawal from intimacy (Mikulincer & Shaver, 2003), and it has been shown to predict greater anger and defensive behavior during threatening interactions (Overall, Simpson, & Struthers, 2013; Simpson et al., 1996). Low self-esteem also shares similarities to attachment anxiety including sensitivity to rejection and reactivity to relationship threats (e.g., Murray, Rose, Bellavia, Hol￾mes, & Kusche, 2002). Despite positive associations with avoidance and self-esteem, we expected that the specific hyperactivating strate￾gies and resulting consequences of inducing guilt we targeted would be specific to attachment anxiety. Study 1 In Study 1, we asked both members of committed couples to report their emotions and threatening relationship events at the end of each day for three weeks. This method allows an examination of anxious reactions to relationship threat as they naturally occur across couples’ lives and provides a window to how these re￾sponses typically result in the partner feeling guilty. We predicted that individuals higher in attachment anxiety would respond to threatening relationship events with heightened hurt feelings. We also hypothesized that anxious individuals’ pronounced hurt feel￾ings would trigger hyperactivating strategies in the form of exag￾gerated expressions of hurt to induce guilt in the partner. In Study 1, we tested this prediction by assessing whether, on days when people experienced and therefore might express or exaggerate hurt, individuals higher in anxiety were perceived by their partner to be feeling more hurt compared to individuals lower in anxiety who reported equivalent levels of hurt. If partners perceive more anx￾ious individuals to feel greater hurt than low anxious individuals facing comparable levels of hurt, this suggests that anxious indi￾viduals express and exaggerate their hurt more to their partner. We also expected these guilt-inducing expressions would lead to anx￾ious individuals’ partners feeling greater guilt. Our final aim was to investigate whether partner’s guilt has more positive effects for individuals high versus low in anxiety. To do this, we tested whether the partner’s guilt assessed during the 3-week diary was associated with relationship satisfaction gathered 9 months following the completion of the daily re￾ports. By averaging across the diary period, we gathered an index of the degree to which couples’ interactions across the course of their normal life resulted in the partner feeling guilty, which should be associated with later satisfaction if guilt has important implications for how people come to feel and think about their relationship. Because greater guilt by the partner across daily life reflects more frequent and severe transgres￾sions, and signals the partner’s fault for those transgressions, we expected that greater partner guilt would be typically asso￾ciated with lower relationship satisfaction. However, because successfully inducing guilt communicates commitment and care, we expected that anxious intimates would fare better (i.e., experience lower declines or maintain levels of satisfaction) when their partner reported feeling greater guilt. However, we also expected that attempts to make partners feel guilty would have detrimental effects on the partner and, thus, partners of anxious intimates would feel less satisfied when they felt greater levels of guilt across the diary period. Method Participants. Seventy-eight heterosexual couples who replied to campus-wide advertisements were offered $90NZD for com￾pleting all phases of the study (an initial session, a 3-week daily diary, and a follow-up questionnaire 9 months later).1 Participants were on average 22.44 years of age (SD
4.81) and were involved in serious romantic relationships (11% married, 33% cohabiting, 1 This sample has been used previously to investigate the links between depressive symptoms and perceptions of daily behavior (Overall & Ham￾mond, 2013), but there is no overlap in the questions, measures, or aims of this research, and the results presented are entirely unique. The longitudinal data have not been reported before. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ATTACHMENT ANXIETY AND PARTNER GUILT 239