University of California, Davis, USA
Bar-Ilan University, Ramat Gan, Israel
Running Head: Attachment and Social Judgment Address correspondence to:
Phillip R. Shaver Mario Mikulincer
Department of Psychology Department of Psychology
University of California, Davis Bar-Ilan University
One Shields Avenue Ramat Gan, 52900
Davis, CA 95616-8686 Israel
E-mail: email@example.com E-mail: firstname.lastname@example.org Author note: Preparation of this chapter was facilitated by a grant from the Fetzer Institute.
In recent years, attachment theory (Bowlby, 1982/1969, 1973, 1980), designed originally to characterize infant-parent emotional bonding, has been applied, first, to the study of adolescent and adult romantic relationships, and then to the study of broader social phenomena. In the present chapter we review and integrate this large and still growing body of work to demonstrate the usefulness and validity of attachment theory for explaining individual variations in a wide array of social judgments, including appraisals of self and others, appraisals of person-environment transactions, and cognitive reactions to new information, out-groups, others’ needs, and transient affective states. We also provide an updated integrative model of the dynamics of the attachment system (Shaver & Mikulincer, in press), which explains the effects of two major individual-difference dimensions, attachment-related anxiety and avoidance, on social judgments and identifies the implicit and explicit mechanisms that mediate these effects.
Basic Concepts in Attachment Theory and Research
In his classic trilogy, Bowlby (1982/1969, 1973, 1980) developed a theoretical framework for explaining the nature of the affective ties we form with significant others and the relevance of these ties for socioemotional functioning. This theoretical framework can now be viewed as a part of evolutionary psychology (see Brewer and Haselton & Buss, this volume). Bowlby (1982/1969) argued that human infants are born with a repertoire of behaviors (attachment behaviors) aimed at attaining or maintaining proximity to supportive others (attachment figures) as a means of protecting themselves from physical and psychological threats. These proximity-seeking behaviors are organized around a psychoevolutionary adaptation (attachment behavioral system), which emerged over the course of evolution to increase the likelihood of survival and reproduction on the part of members of a species born with immature capacities for locomotion, feeding, and defense. Although the attachment system is most critical during the early years of life, Bowlby (1988) assumed that it is active over the entire life span and is manifested in thoughts and behaviors related to proximity seeking in times of need.
Beyond describing the universal aspects of the attachment system, Bowlby (1973) delineated possible individual differences in its functioning. In his view, these individual differences are derived from the reactions of significant others to attachment-system activation and from the internalization of these reactions in the form of attachment working models of self and others. On the one hand, interactions with significant others who are available and responsive to one’s needs facilitate the optimal functioning of the attachment system and promote the formation of a sense of attachment security. This sense consists of positive expectations about others’ availability in threatening situations, positive views of the self as competent and valued, a sense of optimism in dealing with threats, and increased confidence in support seeking as a primary distress-regulation strategy. The sense of attachment security also facilitates engagement in autonomy-promoting activities (e.g., exploration) and ability to make risky decisions while feeling confident that support is available if needed (Bowlby, 1988).
On the other hand, interactions with significant others who are unresponsive to one’s attachment needs foster insecurity regarding others’ goodwill and doubts about the effectiveness of proximity seeking. During these painful interactions, distress is not properly managed, insecure attachment working models are formed, and strategies of affect regulation other than support seeking are developed. Attachment theorists (e.g., Cassidy & Kobak, 1988; Main, 1990) have delineated two major insecure strategies: hyperactivation and deactivation of the attachment system. Hyperactivation is characterized by recurrent attempts to minimize distance from attachment figures and elicit support by clinging and controlling responses. Deactivation consists of attempts to maximize distance from attachment figures while adopting a self-reliant stance (Bowlby, 1988; Cassidy & Kobak, 1988).
While testing these theoretical ideas in studies of adults, most researchers have focused on “styles” of attachment – systematic patterns of expectations, emotions, and behavior in close relationships that are viewed as the residue of particular kinds of attachment histories (Fraley & Shaver, 2000). These residues are thought to inhere in internal working models of self and others. Initially, individual-difference studies of attachment styles in adulthood were based on Ainsworth, Blehar, Waters, and Wall’s (1978) tripartite typology of attachment styles in infancy – secure, anxious, and avoidant – and on Hazan and Shaver’s (1987) conceptualization of adult parallels to these styles in the marital/romantic domain. Subsequent studies (e.g., Bartholomew & Horowitz, 1991; Brennan, Clark, & Shaver, 1998; Fraley & Waller, 1998) revealed that adult attachment styles are best conceptualized, not as distinct types, but as regions in a continuous two-dimensional space. The dimensions defining this space, attachment-related anxiety and attachment-related avoidance, can be measured with reliable and valid self-report scales (Brennan et al., 1998), which are associated with a wide variety of cognitions and behaviors in close relationships (see Feeney, 1999, for a review) and feelings during daily social interactions (e.g., Pietromonaco & Feldman Barrett, 1997; Tidwell, Reis, & Shaver, 1996)
In this two-dimensional space, what was formerly called the “secure style” is a region in which both anxiety and avoidance are low. This region is defined by a positive history of interactions with significant others, a sense of attachment security, and comfort with closeness and interdependence. What was called the “anxious style” refers to a region in which anxiety is high and avoidance is low. Persons high in attachment anxiety are characterized by insecurity concerning others’ goodwill and reliable support, a strong need for closeness, fear of being rejected, and reliance on hyperactivating affective strategies. What was called the “avoidant style” refers to a region in which avoidance is high. This region is defined by insecurity concerning others’ goodwill, compulsive self-reliance, and the adoption of deactivating affect-regulation strategies. In Ainsworth et al.’s (1978) early two-dimensional analysis, based on a discriminant analysis that included all of these authors’ continuous coding scales of infant behavior in a laboratory “strange situation,” avoidant infants occupied mainly the region where avoidance was high and anxiety was low. In adult attachment research, Bartholomew and Horowitz (1991) drew a distinction between “dismissing avoidants” (people who are high on avoidance and low on anxiety) and “fearful avoidants” (those high on both avoidance and anxiety).
Although attachment styles may initially be formed during early interactions with primary caregivers, Bowlby (1988) contended that meaningful interactions with significant others throughout life can update a person’s attachment working models (and associated behavioral orientation). Moreover, although attachment style is often conceptualized as a global orientation toward close relationships, there are theoretical and empirical reasons for believing that working models of attachment are part of a hierarchical cognitive network that includes a complex, heterogeneous array of episodic, relationship-specific, and generalized attachment representations (Collins & Read, 1994). These representations can be viewed as existing at different levels along an implicit-to-explicit continuum of information processing, as discussed by Brewer and others in this volume. People possess multiple attachment schemas, and both congruent and incongruent attachment-related cognitions may coexist in the cognitive network with a global attachment style (Baldwin et al., 1996). In fact, research has shown that (a) people can hold relationship-specific attachment orientations organized around experiences with a specific partner (e.g., LaGuardia, Ryan, Couchman, & Deci, 2000; Pierce & Lydon, 2001), and (b) actual or imagined encounters with supportive or non-supportive others can contextually activate congruent attachment orientations (e.g., Mikulincer et al., 2001; Mikulincer & Shaver, 2001), even if these orientations do not fit with the global attachment style.
An Integrative Model of the Dynamics of the Attachment System
Based on an extensive review of adult attachment studies, Shaver and Mikulincer (in press) proposed a model of the activation and dynamics of the attachment system. This model integrates recent findings with the earlier theoretical proposals of Bowlby (1982/1969, 1973), Ainsworth (1991), and Cassidy and Kobak (1988), and is a conceptual extension and refinement of previous control-system representations of the attachment system presented by Shaver, Hazan, and Bradshaw (1988) and Fraley and Shaver (2000).
The model (see Figure 1) includes three major components. One component concerns the monitoring and appraisal of threatening events and is responsible for activation of the attachment system. The second component involves the monitoring and appraisal of the availability and responsiveness of attachment figures who might provide support and relief, satisfy attachment needs, build the individual’s own inner resources, and broaden his or her thought-action repertoire. This component is responsible for variations in the sense of attachment security; it distinguishes between securely and insecurely attached persons, whether anxious or avoidant. The third component concerns monitoring and appraisal of the viability of proximity seeking as a means of coping with attachment insecurity and distress. This component is responsible for variations in the use of hyperactivating or deactivating strategies of affect regulation and distinguishes between anxious and avoidant people. The model also includes excitatory and inhibitory neural circuits (shown as arrows on the left side of the diagram) that result from the recurrent use of hyperactivating or deactivating strategies, which in turn affect the monitoring of threatening events and of attachment figures’ availability.
Following Bowlby’s (1982/1969) reasoning, Shaver and Mikulincer (in press) assume that the monitoring of unfolding events results in activation of the attachment system when a potential or actual threat is perceived. This activation can be viewed as part of what Chartrand and Jefferis (this volume) call automatic goal pursuit, manifested in efforts to seek and/or maintain proximity to attachment figures. Although this component of the model represents the normative operation of the attachment system, which occurs regardless of individual differences in attachment history and orientation, it is still affected by excitatory circuits resulting from the hyperactivating strategies of anxious persons and inhibitory circuits related to avoidant individuals’ deactivating strategies.
Once the attachment system is activated, an affirmative answer to the question about attachment figures’ availability results in a strong sense of attachment security and in what Shaver and Mikulincer (in press), following the lead of Fredrickson (2001), call a “broaden and build” cycle of attachment security. This cycle reflects optimal functioning of the attachment system and is characterized by distress alleviation and bolstered personal adjustment as well as facilitation of other behavioral systems, such as exploration and caregiving, which broaden a person’s perspectives and capacities. Moreover, this cycle encourages a person to openly acknowledge future threats and to rely comfortably on proximity seeking as a primary coping strategy.
It is important to note that the answer to the question about attachment-figure availability depends on the subjective appraisal of this availability and can be biased by a person’s history of interactions with attachment figures and his or her attachment working models. Attachment research has consistently reported that securely attached persons are more likely to make a positive appraisal of attachment-figure availability than insecurely attached persons (Shaver & Hazan, 1993). Despite this cognitive bias, however, reality is still important in the appraisal of attachment-figure availability. In our view, the actual presence of an available attachment figure or contextual cues that activate representations of available attachment figures can lead people to give an affirmative answer to the question of attachment-figure availability. Recent findings have consistently shown that these contextual cues can activate the “broaden and build” cycle of attachment security even among chronically insecure persons (e.g., Mikulincer et al., 2001; Mikulincer & Shaver, 2001).
Perceived unavailability of an attachment figure results in attachment insecurity, which compounds the distress initiated by the appraisal of a threat. This state of insecurity forces a decision about the viability of proximity seeking as a protective strategy. When proximity seeking is appraised as a viable option – because of attachment history, temperamental factors, or contextual cues – people adopt hyperactivating strategies, which include intense approach to attachment figures and continued reliance on others as a source of comfort. Hyperactivation of the attachment system involves excitatory circuits that increase vigilance to threat-related cues and reduce the threshold for detecting cues of attachment figures’ unavailability – the two kinds of cues that activate the attachment system (Bowlby, 1973). As a result, minimal threat-related cues are easily detected, the attachment system is chronically activated, pain related to the unavailability of attachment figures is exacerbated, and doubts about one’s ability to achieve relief and attain a sense of security are heightened. These excitatory circuits account for the psychological correlates of attachment anxiety.
Appraising proximity seeking as not viable results in the adoption of deactivating strategies, manifested in distancing from cues that activate the attachment system – cues related to threats and attachment figures – and making attempts to handle distress alone. These strategies involve inhibitory circuits that lead to the dismissal of threat- and attachment-related cues and the suppression of threat- and attachment-related thoughts, memories, and emotions. These inhibitory circuits are further reinforced by the adoption of a self-reliant attitude that decreases dependence on others and acknowledgment of personal faults or weaknesses. These inhibitory circuits account for the psychological manifestations of attachment avoidance.
According to Shaver and Mikulincer (in press), the three major components of the model have both content and process aspects – a distinction analyzed by von Hippel, Vargas, and Sekaquaptewa (this volume). All components and circuits of the model can operate either consciously or unconsciously (see Brewer, this volume). Moreover, these components and circuits can operate either in parallel or in opposite ways at conscious and unconscious levels. This explains why some avoidant individuals experience conscious deactivation of threat-related cues while also exhibiting unconscious, including physiological, signs of distress (e.g., Dozier & Kobak, 1992; Mikulincer, 1998b; Mikulincer, Florian, & Tolmacz, 1990). It also explains why avoidant individuals exhibit heightened accessibility of attachment-related worries under cognitively demanding situations that prevent more controlled inhibition of these worries (Mikulincer, Birnbaum, Woddis, & Nachmias, 2000; see Lieberman, this volume, for a neuropsychological analysis of the effects of cognitive load).
The Dynamics of the Attachment System and Social Judgments
In our view, the formation of individual differences in attachment-system dynamics can be viewed as the prototypical precursor of variations in social judgments. These differences should play a role in theoretical models designed to explain accuracies and inaccuracies in social judgments (e.g., Funder and Kruglanski, Erb, Chun, & Pierro, this volume). According to Bowlby (1973), attachment-style differences are already present in the first year of life, in infants’ interactions with their primary caregiver, and they form a foundation for the development of specific judgments about others (beliefs about others’ availability), the self (beliefs about self-worth and self-efficacy), transactions with the environment (e.g., beliefs about the positivity of interactions with others, beliefs about the reversibility of threats), and ways of dealing with these transactions (e.g., beliefs about the efficacy of support seeking) as well as regulating cognitions in non-attachment areas (e.g., exploration). These mental products can be generalized across recurrent interactions with a relationship partner. Moreover, they can be generalized across relationships via top-down schematic processing of new partners and relationships, and then become the building blocks of a person’s global social judgments.
Shaver and Mikulincer’s (in press) model provides a guide for delineating attachment-related variations in social judgments. The module that monitors attachment-figure availability and creates a sense of attachment security is related to positive or negative working models of others, which can bias judgments of other people. Moreover, the sense of attachment security can regulate cognitions related to the exploration of new information and the provision of care for others who are in need. The module that monitors the viability of proximity seeking and determines the adoption of deactivating or hyperactivating strategies activates self-reliant or other-reliant attitudes, and can therefore bias judgments about self-worth. In addition, since these strategies are defined by the regulation of threat cues (exaggeration vs. dismissal), proximity to others (maximization vs. minimization), and affective states (perpetuation of threat-related affect vs. distancing oneself from this affect), they can bias judgments of threatening events and self-other proximity as well as cognitive reactions to affective states. In the following sections, we review research findings concerning specific links between dynamics of the attachment system and social judgments.
Attachment-Figure Availability and Social Judgments
The perceived availability of attachment figures has direct implications for the appraisal of others. An affirmative answer to the question about attachment-figure availability activates positive models of others, which may spread to positive appraisals, expectations, and explanations of others’ traits and behaviors. In contrast, a negative answer to this question activates negative models of others, which in turn may negatively bias judgments of other people. This reasoning implies that insecure attachment, either anxious or avoidant, which results from the perceived unavailability of attachment figures, should be associated with more negative judgments of others.
The perceived availability of attachment figures can also regulate attitudes related to exploration and caregiving, two of the other behavioral systems discussed by Bowlby (1982/1969). He claimed that the unavailability of attachment figures inhibits the activation of other behavioral systems, because a person without the protection and support of an attachment figure is likely to be so focused on attachment needs and feelings of distress that he or she lacks the attention and resources necessary to explore the environment and attend empathically to others’ needs. This reasoning implies that attachment insecurity should reduce or prevent exploration of new information in making social judgments, favor the formation of rigid, stereotypic judgments, and inhibit the development of a prosocial orientation and a caring attitude toward needy others.
Appraisal of others’ traits and behaviors. Numerous studies have provided strong support for the hypothesis that attachment-related anxiety and avoidance are associated with negative appraisals of other people. Individuals who score high on the dimensions of attachment-related anxiety and/or avoidance have been found to hold a more negative view of human nature (Collins & Read, 1990), use more negative traits to describe relationship partners (e.g., Feeney & Noller, 1991; Levy, Blatt, & Shaver, 1998), perceive these partners as less supportive (e.g., Davis, Morris, & Kraus, 1998; Ognibene & Collins, 1998), be less satisfied with the support received from others (e.g., Collins & Read, 1990; Larose & Boivin, 1997), feel less trust toward partners (e.g., Collins & Read, 1990; Mikulincer, 1998a; Simpson, 1990), and believe that partners do not truly know them (Brennan & Bosson, 1998).
Both anxiety and attachment avoidance are also associated with negative expectations concerning partner behaviors (e.g., Baldwin et al., 1993; Baldwin et al., 1996; Mikulincer & Arad, 1999). For example, Baldwin et al. (1993) examined the cognitive accessibility of expectations concerning partner’s behaviors in a lexical-decision task and found that for both anxious and avoidant persons, negative partner behaviors (e.g., partner being hurtful) were more accessible than they were among secure persons. These negative expectations have also been found in studies that assessed relationship-specific attachment orientations (e.g., Baldwin et al., 1996; Mikulincer & Arad, 1999).
Similar attachment-style differences have been found when research participants are asked to explain other people’s behavior (e.g., Collins, 1996; McCarthy & Taylor, 1999; Mikulincer, 1998a, 1998b). For example, Collins (1996) asked participants to explain hypothetical negative behaviors of a romantic partner and found that more anxious and avoidant people were more likely to provide explanations that implied lack of confidence in the partner’s love, attribute partner’s negative behaviors to stable and global causes, and view these behaviors as negatively motivated.
Exploration of new information and the rigidity of social judgments. The link between attachment security and exploration has been studied by adult-attachment researchers. Hazan and Shaver (1990) proposed that work serves as one form of exploration in adulthood and found that more anxious and avoidant individuals reported more negative attitudes toward work and were less satisfied with work activities. Moreover, whereas secure people perceived work as an opportunity for learning and advancement, anxious individuals perceived it as an opportunity for social approval, and avoidant individuals as an opportunity for evading close relationships (Hazan & Shaver, 1990). Additional studies reveal that more anxious and avoidant persons are less willing to explore the environment than secure persons (Green & Campbell, 2000) and avoidant persons report less curiosity about new information (Mikulincer, 1997, Study 1). Green and Campbell (2000, Study 2) found that the contextual priming of attachment security heightened people’s willingness to explore novel stimuli.
There is also evidence that attachment insecurity, either anxious or avoidant, inhibits engagement in cognitive exploration, leads to rejection of new information, and fosters rigid, stereotypic judgments (e.g., Green-Hennessy & Reiss, 1999; Mikulincer, 1997; Mikulincer & Arad, 1999). For example, Mikulincer (1997, Study 4) focused on the primacy effect – the tendency to make judgments on the basis of early information and to ignore later data – and found that both anxious and avoidant individuals were more likely than secure individuals to rate a target person based on the first information received. In a separate study, Mikulincer (1997, Study 5) examined stereotype-based judgments, i.e., the tendency to judge a member of a group based on a generalized notion of the group rather than on exploration of new information about the member. More anxious and avoidant individuals tended to evaluate the quality of an essay based on the supposed ethnicity of the writer: The more positive the stereotype of the writer’s ethnic group, the higher the grade assigned to the essay. In contrast, more secure individuals were less affected by ethnic stereotypes.
Mikulincer and Arad (1999) examined attachment-style differences in the revision of knowledge about a relationship partner following behavior on the part of the partner that seemed inconsistent with this knowledge. Compared to secure persons, both anxious and avoidant individuals showed fewer changes in their baseline perception of the partner after being exposed to expectation-incongruent information about the partner’s behavior. They were also less capable of recalling this information. Importantly, this finding was replicated when relationship-specific attachment orientations were assessed: The higher the level of attachment anxiety or avoidance toward a specific partner, the fewer the revisions people made in their perception of this partner upon receiving expectation-incongruent information (Mikulincer & Arad, 1999, Study 2). Moreover, the contextual heightening of the sense of attachment security (visualizing a supportive other) increased cognitive openness and led even chronically anxious and avoidant people to revise their conception of a partner based on new information (Mikulincer & Arad, 1999, Study 3).
Attitudes toward others’ needs. Attachment studies have supported the hypothesis that attachment insecurity inhibits a caregiving orientation toward others who are in need. For example, attachment-related anxiety and avoidance are associated with low levels of reported responsiveness to a relationship partner’s needs (e.g., Feeney, 1996; Kunce & Shaver, 1994) and fewer supportive behaviors toward a partner under threatening conditions (e.g., Fraley & Shaver, 1998; Simpson, Rholes, & Nelligan, 1992). Westmaas and Silver (2001) reported that whereas attachment avoidance was associated with a non-supportive attitude toward a confederate who purportedly had been diagnosed with cancer, attachment anxiety was associated with the expression of high levels of distress during an interaction with this confederate. In addition, Mikulincer et al. (2001) found that both attachment anxiety and avoidance were associated with low levels of altruistic empathy to the plight of others. Moreover, whereas the contextual heightening of the sense of attachment security increased altruistic empathy, the contextual activation of attachment anxiety or avoidance reduced this prosocial attitude (Mikulincer et al., 2001).
There is also evidence that attachment insecurity inhibits the development of a prosocial orientation (e.g., Mikulincer et al., in press; Van Lange, Otten, DeBruin, & Joireman, 1997). For example, Mikulincer et al. (in press) reported that attachment avoidance was associated with low endorsement of two humanitarian values, universalism (concern for the welfare of all people) and benevolence (concern for the welfare of close persons). Moreover, contextual heightening of the sense of attachment security increased the endorsement of these values.
The Adoption of Hyperactivating or Deactivating Strategies and Social Judgments
The perceived viability of proximity seeking and the consequent adoption of hyperactivating or deactivating strategies have direct implications for self-appraisals, appraisals of person-environment transactions, judgments of self-other similarity, and cognitive reactions to affective states. With regard to attachment anxiety, the excitatory pathways running from hyperactivating strategies to the monitoring of threat-related cues cause attention to be directed to self-relevant sources of distress (personal weaknesses) and to threatening aspects of person-environment transactions. This process can foster chronic negative self-appraisals and exaggerated appraisals of external threats. These excitatory circuits also perpetuate and exacerbate negative affect, therefore encouraging judgments that are congruent with this affective state and inhibiting judgments that would be congruent with positive affect. In addition, chronic activation of the attachment system leads to recurrent attempts to minimize distance from others, which can promote a false consensus bias and overestimation of self-other similarity.
With regard to attachment avoidance, the inhibitory circuits running from deactivating strategies to monitoring of threat-related cues divert attention from self-relevant and external sources of distress, and can therefore inhibit the appraisal of negative aspects of the self as well as external threats. This cognitive bias is further reinforced by the adoption of a self-reliant attitude that requires protection and enhancement of self-worth. These inhibitory circuits also promote detachment from challenging and emotionally involving interactions that potentially constitute sources of threat. In this way, deactivating strategies favor the dismissal of the personal relevance and challenging aspects of person-environment transactions and the rejection of affective states as relevant inputs for making social judgments. In addition, deactivation of the attachment system and defensive attempts to maximize distance from others can promote a false distinctiveness bias and the overestimation of self-other dissimilarity.
Self-appraisals. There is extensive evidence for the hypothesis that attachment anxiety is associated with negative self-appraisals. Compared to secure persons, anxiously attached persons report lower self-esteem (e.g., Bartholomew & Horowitz, 1991, Mickelson, Kessler, & Shaver, 1997), hold more negative perceptions of self-competence and more negative expectations of self-efficacy (e.g., Brennan & Morris, 1997; Cooper, Shaver, & Collins, 1998), incidentally recall more negative traits, and exhibit greater discrepancies between actual-self and self-standards (Mikulincer, 1995).
With regard to attachment avoidance, the findings are less consistent. Although some investigators have found that avoidant persons have higher self-esteem than anxiously attached persons (e.g., Bartholomew & Horowitz, 1991; Mikulincer, 1995), others have found that attachment avoidance is associated with low self-esteem (e.g., Brennan & Morris, 1997; Feeney & Noller, 1990). In addition, there is some evidence that attachment avoidance is related to negative appraisals of the self when these appraisals concern competence in social and interpersonal settings (e.g., Brennan & Bosson, 1998; Brennan & Morris, 1997).
These inconsistencies may be due to the fact that most of the reviewed studies have examined self-appraisals under neutral, non-threatening conditions. This is a crucial feature of these studies with respect to avoidant individuals’ self-appraisals, because avoidant individuals’ deactivating strategies are hypothesized to be active mainly in threatening contexts (Shaver & Mikulincer, in press). This feature is less important with regard to anxious individuals’ self-appraisals, because anxious individuals’ hyperactivating strategies have been found to be active even in neutral, non-threatening contexts (e.g., Mikulincer, Gillath, & Shaver, in press). In fact, in examinations of self-efficacy expectations for coping with threats, avoidant people have consistently reported more positive expectations than anxious people (e.g., Mikulincer & Florian, 1995, 1998). Accordingly, Mikulincer (1998c) reported that avoidant individuals exhibited a defensive self-enhancement reaction to threats. Specifically, Mikulincer (1998c) exposed participants to experimentally induced threatening or neutral situations and found that avoidant people made more positive self-appraisals following threatening than neutral situations.
Similar cognitive reactions have been observed in studies examining causal attribution for negative events (e.g., Kennedy, 1999; Kogot, 2001; Man & Hamid, 1998). For example, Kogot (2001) asked undergraduates to explain the reason for their failure in an academic examination and found that anxious persons displayed what Abramson, Metalsky, and Alloy (1989) called a “hopelessness-depressive” pattern of attributions. These people attributed failure to more internal, stable, global, and uncontrollable causes, while attributing success to more external, unstable, specific, and uncontrollable causes. Avoidant people exhibited a defensive, self-protective pattern of causal attributions: They attributed failure to less internal causes, dismissed the diagnosticity of the failure, and blamed others for it.
Appraisals of person-environment transactions. Attachment researchers have extensively documented an association between attachment anxiety and exaggerated appraisal of the threatening aspects of person-environment transactions. Compared to secure persons, anxious persons hold more pessimistic beliefs about close relationships (e.g., Carnelley & Janoff-Bulman, 1992; Whitaker et al., 1999), use more threat/loss frames in thinking about relationships (Boon & Griffin, 1996; Pistole, Clark, & Tubbs, 1995), appraise relational outcomes as more negative (Feeney & Noller, 1992), and endorse more dysfunctional beliefs about relationships (Whisman & Allan, 1996). Anxious individuals’ exaggerated threat appraisals have also been observed in daily social interactions (e.g., Pietromonaco & Feldman Barrett, 1997), in small group interactions (Rom & Mikulincer, 2001), and in response to stressful life events (e.g., Mikulincer & Florian, 1995, 1998, 2001).
Anxious threat exaggeration was also observed in a recent study on attitudes toward out-groups (Mikulincer & Shaver, 2001). Persons scoring high on attachment anxiety were more likely to appraise out-groups as psychologically threatening and to display more negative and hostile attitudes toward them than less anxious persons. Importantly, contextual heightening of the sense of attachment security reduced anxious individuals’ negative appraisals and responses to out-groups. These findings may be relevant to researchers who study stereotypes, prejudice, and intergroup biases, such as Galinsky and Martorana, Johnston and Miles, and von Hippel et al. (this volume).
With regard to attachment avoidance, studies have consistently found that avoidant persons perceive stressful events in less threatening terms than anxious persons (see Mikulincer & Florian, 2001, for a review). More important, there is some evidence that avoidant people minimize personal involvement in transactions with the environment. They dismiss the importance and relevance of close relationships, social interactions, and threatening events (e.g., Kogot, 2001; Miller, 2001; Pistole et al., 1995), and see fewer benefits and challenges in social interactions than do secure individuals (e.g., Horppu & Ikohen-Varila, 2001; Pietromonaco & Feldman Barrett, 1997; Rom & Mikulincer, 2001).
Cognitive reactions to affective states. Recent studies on the link between affect and cognition (see Forgas, this volume, for a review of this research area) provide strong evidence concerning anxious people’s perpetuation of negative affect and avoidant people’s distancing from affective states. Mikulincer and Sheffi (2000) exposed participants to positive or neutral affect inductions and then assessed their breadth of mental categorization and creative problem-solving performance. The beneficial effects of positive affect induction on problem solving and category breadth (reported by Isen & Daubman, 1984, and Isen, Daubman, & Nowicki, 1987) were observed only among secure people. For avoidant persons, no significant difference was found between positive and neutral affect conditions. For anxiously attached people, a reverse effect was found which resembled the typical effects of negative affect induction: They reacted to a positive affect induction with impaired creativity and a narrowing of mental categories.
In another study, Pereg (2001) exposed participants to a negative or neutral affect induction and then assessed recall of positive and negative information. A negative mood induction led to better recall of negative information and worse recall of positive information than a neutral condition (mood-congruent recall pattern) mainly among anxiously attached people. A negative mood induction had no effect on the recall pattern of avoidant people. Pereg (2001) conceptually replicated this finding while studying causal attributions for negative relationship events. A negative mood induction led to more stable/global attributions than a neutral condition (mood-congruent attribution pattern) among anxiously attached people. A negative mood induction had no effect on avoidant individuals’ causal attributions. Importantly, persons scoring low on both attachment anxiety and avoidance (secure style) exhibited mood-incongruent patterns of recall and attribution following negative mood induction. (Such differences in patterns of affect infusion are extensively discussed by Forgas, this volume.)
In integrating Mikulincer and Sheffi’s (2000) with Pereg’s (2001) findings, one can delineate the following patterns of cognitive reactions. Anxiously attached people exhibit cognitive reactions that maintain and even exacerbate negative mood – mood-incongruent cognitions following positive affect induction and mood-congruent cognitions following negative affect induction. Avoidant individuals tend to distance themselves from both positive and negative affect and to inhibit the cognitive impact of any affective state.
Appraisals of self-other similarity. In a series of studies, Mikulincer, Orbach, and Iavnieli (1998) provided direct evidence concerning the links between attachment insecurity and false-consensus and false-distinctiveness biases in social judgment. Whereas anxious people were more likely than secure people to perceive others as similar to themselves and to show a false consensus bias in both trait and opinion descriptions, avoidant individuals were more likely to perceive other people as dissimilar to them and to exhibit a false distinctiveness bias. Mikulincer et al. (1998) also found that anxious individuals reacted to threats by generating a self-description that was more similar to a partner’s description and by recalling more partner traits that were shared by themselves and the partner. In contrast, avoidant persons reacted to the same threats by generating a self-description that was more dissimilar to a partner’s description and by forgetting more traits that were shared by themselves and the partner.
Taken as whole, there is extensive evidence that the dynamics of the attachment system can explain variations in a wide variety of cognitions and social judgments. Figure 2 provides a schematic summary of the reviewed findings. In the next section, we discuss the possible mediational processes by which the dynamics of the attachment system affect social judgments.