Attachment predicts cortisol response and closeness in dyadic social interaction

Attachment predicts cortisol response and closeness in dyadic social interaction

Accepted Manuscript Title: Attachment predicts cortisol response and closeness in dyadic social interaction Authors: Sarah Ketay, Lindsey A. Beck PII:...

559KB Sizes 0 Downloads 80 Views

Accepted Manuscript Title: Attachment predicts cortisol response and closeness in dyadic social interaction Authors: Sarah Ketay, Lindsey A. Beck PII: DOI: Reference:

S0306-4530(17)30076-8 http://dx.doi.org/doi:10.1016/j.psyneuen.2017.03.009 PNEC 3574

To appear in: Received date: Revised date: Accepted date:

25-1-2017 4-3-2017 5-3-2017

Please cite this article as: Ketay, Sarah, Beck, Lindsey A., Attachment predicts cortisol response and closeness in dyadic social interaction.Psychoneuroendocrinology http://dx.doi.org/10.1016/j.psyneuen.2017.03.009 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Running head: ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

1

Attachment predicts cortisol response and closeness in dyadic social interaction

Sarah Ketay1 University of Hartford

Lindsey A. Beck2 Emerson College

Author Information: 1

East Hall University of Hartford 200 Bloomfield Avenue West Hartford, CT 06117 USA Tel.: 860-768-5906 Email: [email protected] 2

Walker Building Emerson College 120 Boylston Street Boston, MA 02116 USA Tel.: 617-824-3508 Email: [email protected]

Address correspondence to Sarah Ketay, East Hall, University of Hartford, 200 Bloomfield Avenue, West Hartford, CT 06117, Tel.: 860-768-5906. E-mail address: [email protected]

Highlights   

Attachment avoidance predicted cortisol response patterns. Higher cortisol response associated with a mismatch in participants’ attachment avoidance. Lower levels of attachment anxiety were associated with higher levels of closeness.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

2

Abstract The present study examined how the interplay of partners’ attachment styles influences cortisol response, actual closeness, and desired closeness during friendship initiation. Participants provided salivary cortisol samples at four timepoints throughout either a high or low closeness task that facilitated high or low levels of self-disclosure with a potential friend (i.e., another same-sex participant). Levels of actual closeness and desired closeness following the task were measured via inclusion of other in the self. Results from multi-level modeling indicated that the interaction of both participants’ attachment avoidance predicted cortisol response patterns, with participants showing the highest cortisol response when there was a mismatch between their own and their partners’ attachment avoidance. Further, the interaction between both participants’ attachment anxiety predicted actual closeness and desired closeness, with participants both feeling and wanting the most closeness with partners when both they and their partners were low in attachment anxiety.

Keywords: Cortisol, Adult Attachment; Friendship; Close Relationships; Relationship Initiation; Closeness

1. Introduction The desire for interpersonal closeness is a fundamental human drive that satisfies the need to belong (Baumeister and Leary, 1995). Closeness, in turn, has been conceptualized as inclusion of other in the self, which reflects the extent to which one overlaps with a close other in terms of self-perception (Aron et al., 1991). Despite the fundamental nature of the desire for closeness, individual differences exist, which are shaped by attachment styles. Two key dimensions of adult attachment styles are attachment avoidance and attachment anxiety (Brennan et al., 1998). Individuals high in avoidance express disinterest in closeness and prefer not to depend on others or have others depend on them. In contrast, individuals high in anxiety desire excessive closeness yet worry about others’ responsiveness to their needs. Attachment styles predict how people think, feel, and behave in close relationships, including how they

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

3

perceive partners, regulate emotions, and give and seek support (for reviews, see Mikulincer and Shaver, 2007, and Pietromonaco and Beck, 2015). Attachment styles also influence both acute cortisol responses to stress and cortisol patterns in daily life (Dewitte et al., 2010; Kidd et al., 2011; Powers et al. 2006; Quirin, et al., 2008). Furthermore, attachment styles may predict health outcomes, although the mechanisms remain largely unexplored (McWilliams and Bailey, 2010). Cortisol and Attachment Styles The hypothalamic-pituitary-adrenal (HPA) axis, a major stress response system that modulates the release of the steroid hormone cortisol, has been established as a mechanism through which stress affects health outcomes. Most research to date on the neuroendocrine correlates of attachment styles has focused on the HPA axis. For example, attachment insecurity (i.e., high attachment avoidance, high attachment anxiety, or both) has been linked to dysregulated cortisol responses before, during, and after relationship conflict (Beck et al., 2013; Brooks et al., 2011; Laurent and Powers, 2007; Powers et al., 2006). However, the majority of this research has examined cortisol responses in romantic relationships. To our knowledge, no studies have explored links between attachment styles and cortisol responses during friendship initiation. Given that the ability to connect with others significantly influences physical and emotional health (Cacioppo et al., 2006; Holt-Lunstad, et al., 2010), it is critical to examine the neuroendocrine correlates of attachment styles outside the context of romantic relationships. Such work has the potential to enhance understanding of attachment theory and of links between close relationships and physical and emotional health. Although research has yet to examine how attachment styles shape cortisol responses during friendship initiation, research has begun to examine cortisol responses during friendship initiation in general (Ketay et al., 2017; Page-Gould et al., 2008). For example, one study of early-stage friendship and hormones found that participants’ cortisol levels decreased across three laboratory-based intergroup friendship meetings (Page-Gould et al., 2008). Researchers have also found that lower basal cortisol and dynamic cortisol changes during friendship initiation are associated with higher levels of actual closeness and desired closeness (Ketay et al., 2017). However, to date no studies have considered potential

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

4

contributions of attachment styles and closeness outcomes and their physiological correlates during friendship initiation. Attachment Styles in Relationship Initiation Although the hormonal correlates of attachment in friendship initiation have yet to be explored, research has begun to examine the role of attachment styles in other aspects of relationship initiation. This research suggests that people high in avoidance and/or anxiety may have difficulty during relationship initiation. For example, avoidant individuals share less personal information with (Mikulincer and Nachshon, 1991) and seek less personal information about potential partners (Aspelmeier and Kerns, 2003; Beck and Clark, 2009). In contrast, although anxious individuals share personal information with potential partners, they are less responsive to potential partners’ personal sharing than are secure individuals (Mikulincer and Nachshon, 1991). Paradoxically, anxious individuals also express less interest in potential partners when they are more interested in them (Vorauer et al., 2003). Together, these findings indicate that avoidant and anxious individuals face obstacles to relationship initiation, but that the nature of these obstacles differs. (Mis)matched Attachment in Dyads As discussed above, individuals’ attachment styles may influence relationship initiation. These findings suggest that the interplay between both partners’ attachment styles also might shape this process, including partners’ feelings of closeness to one another and their associated physiological responses. Few studies have examined how the fit between both partners’ attachment styles might shape relationship initiation, and none to our knowledge have considered these factors during friendship initiation. However, a growing body of research (e.g., Beck et al., 2013; Feeney, 2003; Mikulincer et al., 2002) considers the importance of examining the interplay between both partners’ attachment styles in established romantic relationships. Some studies have found that the interplay of partners’ attachment styles fails to contribute to relationship outcomes (Creasey, 2002; Mikulincer and Florian, 1999; Paley et al., 1999). Other studies have found that the pairing of one anxious partner with one avoidant partner, or the pairing of two anxious partners, may lead to poorer relationship functioning (see Mikulincer and Shaver, 2007, for a review). The interplay of partners’ attachment styles may also contribute to

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

5

differential physiological responses. For instance, a recent study found that both spouses in newlywed couples with anxious wives and avoidant husbands showed exaggerated cortisol reactivity in anticipation of a relationship conflict (Beck et al., 2013). To date, research on the interplay of partners’ attachment styles and their associated cortisol responses has focused exclusively on romantic partners. The Present Study Although research has begun to investigate the neuroendocrine correlates of attachment styles, questions about the role of attachment styles in cortisol responses outside a romantic context remain largely unexplored. Further, there is little knowledge about how the interplay of attachment styles in a dyad might shape both partners’ feelings of closeness. Therefore, we aim to explore how the interplay between partners’ attachment styles predicts cortisol responses during friendship initiation. We also aim to explore how the interplay between partners’ attachment styles predicts actual closeness and desired closeness, as measured by inclusion of other in the self during friendship initiation. 2. Method 2.1. Participants One-hundred sixty-eight participants (84 dyads) were recruited through the Introductory Psychology subject pool at a private university in the Northeastern U.S. Participants were paired with a same-sex stranger and received course credit for participation. Four dyads were excluded from analyses because participants knew one another prior to the study; three additional dyads were excluded from analyses because at least one participant in each dyad did not complete the attachment measure. Therefore, the final sample included 154 participants (77 dyads; 80.52% female; Mage = 20.00, SDage = 2.06). Participants were relatively diverse with respect to ethnicity (53.90% Caucasian). Participants were randomly assigned to their partner in this task. However, assignment to condition was not random in that the first group of participants completed a high closeness task (45 dyads, N = 90). The second group of participants completed a low closeness task (32 dyads, N = 64), which was added as a comparison task after completion of data collection for the high closeness task. Finally, two additional dyads were

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

6

excluded from the cortisol analyses because at least one participant in each dyad was missing more than two cortisol samples. 2.2. Procedure All laboratory sessions were conducted between 12pm and 5pm to control for the diurnal rhythm of cortisol (Granger et al., 1999; Schultheiss and Stanton, 2009). Participants were asked to refrain from eating for one hour prior to the session to prevent contamination of saliva samples. Participants arrived at the laboratory expecting to have a conversation with a same-sex stranger. After giving informed consent, participants drank a few sips of water to facilitate a clean saliva sample, which would be used to determine baseline cortisol levels. Participants then privately completed questionnaires, interacted with a same-sex stranger in the context of a high or low closeness task, and provided four saliva samples to assess cortisol levels before, during, and after the task. 2.3. Measures 2.3.1. Salivary Cortisol Cortisol takes about 20 minutes to be detected in saliva after secretion from the adrenal glands. Therefore, each sample reflects participants’ cortisol levels approximately 20 minutes prior to collection (Stansbury and Gunnar, 1994). Participants provided four saliva samples to measure cortisol levels before, during, and after the high or low closeness task. Participants provided their first saliva sample 10 minutes after arriving at the laboratory. Next, participants completed some questionnaires and then began either the high or low closeness task (described below). Participants provided a second saliva sample halfway through the task (approximately 30 minutes after arrival at the laboratory), a third saliva sample at completion of the task (approximately 50 minutes after arrival at the laboratory), and a fourth saliva sample 20 minutes after completion of the task (approximately 70 minutes after arrival at the laboratory). Saliva samples were collected by placing Salimetrics Oral Swab (Salimetrics, LLC) under the tongue for 90s. Samples were sealed in a cryogenic vial and stored in a -20C freezer. Samples from the high closeness condition were shipped on dry ice and analyzed for cortisol concentrations using Salimetrics Assay kits (Granger et al., 1999, 2004) at Salimetrics Inc. (Carlsbad, CA) via enzyme immunoassay

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

7

(EIA). Intra-assay and inter-assay CVs were low (Avg CVs = 4.9% and 6.9%, respectively). Samples from the low closeness condition were shipped on dry ice and analyzed for cortisol concentrations using commercially available kits from Salimetrics Inc. at the University of Trier (Trier, Germany) via enzyme immunoassay (EIA). Intra-assay and inter-assay CVs were low (Avg CVs = 9.3% and 6.0%, respectively). All samples were assayed in duplicate. 2.3.2. High and Low Closeness Conditions Participants in the high closeness condition worked in pairs to complete the Fast Friends procedure as developed by Aron et al. (1997). Participants in this condition took turns answering 36 questions that progressively increased in levels of self-disclosure. Participants in the low closeness condition worked in pairs to complete tasks adapted from Steele (1999) and used in Ketay et al. (2017). Participants in this condition took turns giving each other directions to different places on campus, thinking of geographical locations that began with the same first letter in the alphabet, and reading and responding to questions about a written passage. 2.3.3. Attachment Style Participants’ levels of attachment avoidance and attachment anxiety were assessed using the Relationships Questionnaire (RQ; Bartholomew and Horowitz, 1991), a four-item measure of adult attachment. Participants read four short paragraphs, each of which described a prototypical attachment pattern (i.e., secure, fearful, preoccupied, dismissing), and rated their agreement with each pattern on a 7point Likert scale. The two dimensions of adult attachment—attachment avoidance and attachment anxiety—were calculated from participants’ ratings, following Griffin and Bartholomew (1994) and Rasco and Warner (2016). 2.3.4. Post-Task Measures Participants were asked how close they felt to their partner and how close they desired to feel to their partner after the task. Actual closeness and desired closeness were measured via the Inclusion of Other in the Self scale (IOS; Aron, Aron, and Smollan, 1992). The IOS scale is a pictorial 7-point Likert scale that depicts two increasingly overlapping circles, with greater overlap representing higher levels of actual

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

8

closeness or desired closeness. Participants were asked to circle the picture that best represented their actual relationship with their partner in the task (i.e., their actual closeness) and to circle the picture that best represented their desired relationship with their partner in the task (i.e., their desired closeness). 2.3.5. Analytic Strategy We employed the Actor-Partner Interdependence Model (APIM; Kashy and Kenny, 2000; Kenny, Kashy, and Cook, 2006) and tested multilevel models using the linear mixed-effects models (MIXED) procedure in SPSS 22.0. This approach accounts for nonindependence of dyadic data by estimating actor effects (e.g., the effect of the individual’s attachment avoidance and attachment anxiety on his or her own cortisol levels) and partner effects (e.g., the effect of the partner’s attachment avoidance and attachment anxiety on the individual’s cortisol levels). These analyses modeled individual responses as nested within dyads. We treated dyads as indistinguishable (Kenny et al., 2006) because each dyad included individuals of the same gender. Continuous variables were grand-mean centered and categorical variables were dummy-coded (following Aiken and West, 1991) by coding closeness condition as 0 for the low closeness condition and 1 for the high closeness condition and by coding gender as 0 for men and 1 for women. We used three analogous models to predict whether individuals’ own attachment avoidance and attachment anxiety, their partners’ attachment avoidance and attachment anxiety, and the interactions between both participants’ attachment avoidance and attachment anxiety would shape their: (1) cortisol levels, (2) actual inclusion of other in the self (i.e., actual closeness), and (3) desired inclusion of other in the self (i.e., desired closeness). We tested each of these three models controlling for closeness condition, following Ketay et al. (2017).1 3. Results 3.1. Cortisol Levels 3.1.1.Preliminaryanalyses. The cortisol distribution was checked to ensure that cortisol values were within the normative range (i.e., ≥ 4 μg/dL; Aardal and Holm, 1995). Although all cortisol values were within the normative range, the

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

9

cortisol distribution was positively skewed. Therefore, a base-10 logarithmic transformation was used to normalize the data. Cortisol values three standard deviations or more from the mean-transformed scores were removed (treated as missing); five samples were removed across three individuals.2 As noted in the procedure, saliva samples were collected at four different time points to assess participants’ cortisol levels (1) before the activity, (2) at the beginning of the activity, (3) in the middle of the activity, and (4) at the end of the activity. We centered time at the third saliva sample by subtracting the mean value of time at the third sample from each participant’s time value, which gave the third sample a value of zero and set the time of the other samples relative to this zero point. We chose to center time at the third saliva sample so that the intercept of the model would reflect participants’ cortisol levels in the middle of the activity. This approach allowed us to assess participants’ cortisol levels from before the activity through the beginning, middle, and end of the activity. Three initial analyses tested whether participants’ cortisol levels could be predicted from (1) time, (2) gender, and (3) hormonal contraceptive use.

These analyses indicated that all three variables

significantly (for time and gender) or marginally (for hormonal contraceptive use) predicted participants’ cortisol levels; therefore, the following analyses controlled for effects of time, gender, and hormonal contraceptive use. 3.1.2. Does the interaction between participants’ attachment styles predict cortisol levels? We predicted that the interplay between both participants’ attachment styles would shape each individual’s cortisol levels. As expected, the interaction between participants’ attachment avoidance significantly predicted cortisol levels (see Table 1). Figure 1 shows participants’ levels of cortisol and attachment avoidance, plotted by high and low (+1 SD from the mean) values of their partners’ attachment avoidance. Simple slopes tests revealed that individuals lower in attachment avoidance had significantly lower cortisol levels when they had partners who also were low in attachment avoidance than did individuals higher in attachment avoidance, γ = .02, t(437.36) = 5.17, p < .001, 95% CI [.01, .03]. However, individuals lower in attachment avoidance had significantly higher cortisol levels when they had partners who were high in attachment avoidance than did individuals higher in attachment avoidance,

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

10

γ = -.01, t(467.82) = -2.69, p = .007, 95% CI [-.02, -.00]. This pattern suggests that participants showed the highest cortisol levels when there was mismatch between their own and their partners’ levels of attachment avoidance (i.e., when they were low in avoidance and their partners were high in avoidance or when they were high in avoidance and their partners were low in avoidance). 3.2. Inclusion of Other in the Self 3.2.1. Preliminary analyses. Two initial analyses tested whether participants’ gender would predict their (1) actual inclusion of other in the self (i.e., actual closeness) and (2) desired inclusion of other in the self (i.e., desired closeness). Participants’ gender did not significantly predict their actual closeness or their desired closeness (ps > .266). Therefore, the following analyses did not control for gender. 3.2.2. Does the interaction between participants’ attachment styles predict actual inclusion of other in the self? We predicted that the interplay between both participants’ attachment styles would shape each individual’s actual closeness with one another.

As expected, the interaction between participants’

attachment anxiety significantly predicted their actual closeness (see Table 2).

Figure 2 shows

participants’ actual closeness and levels of attachment anxiety, plotted by high and low (+1 SD from the mean) values of their partners’ attachment anxiety. Simple slopes tests revealed that individuals lower in attachment anxiety felt significantly closer to partners who also were low in attachment anxiety than did individuals higher in attachment anxiety, γ = -0.11, t(108.55) = -2.78, p = .006, 95% CI [-0.18, -0.03]. However, individuals’ feelings of closeness did not differ as a function of their own attachment anxiety when their partners were high in attachment anxiety, γ = .03, t(108.36) = .70, p = .489, 95% CI [-.05, .10]. This pattern suggests that participants felt the most closeness with their partners when both they and their partners were low in attachment anxiety.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

11

3.2.3. Does the interaction between participants’ attachment styles predict desired inclusion of other in the self? We also predicted that the interplay between both participants’ attachment styles would shape each individual’s desired closeness with one another. As expected, the interaction between participants’ attachment anxiety significantly predicted their desired closeness (see Table 3).

Figure 3 shows

participants’ desired closeness and levels of attachment anxiety, plotted by high and low (+1 SD from the mean) values of their partners’ attachment anxiety. Simple slopes tests revealed that individuals lower in attachment anxiety desired significantly more closeness with partners who also were low in attachment anxiety than did individuals higher in attachment anxiety, γ = -.12, t(123.32) = -3.07, p = .003, 95% CI [.19, -.04]. However, individuals’ desired closeness did not differ as a function of their own attachment anxiety when their partners were high in attachment anxiety, γ = .03, t(123.11) = .66, p = .512, 95% CI [.05, .10]. This pattern suggests that participants wanted the most closeness with their partners when both they and their partners were low in attachment anxiety. 4. Discussion Feeling close to a potential partner is a key factor in the initiation of both friendships and romantic relationships. Closeness and intimacy, in turn, are likely best achieved through the synchrony of selfdisclosure and displays of responsiveness seen in securely-attached individuals (Grabill and Kerns, 2000; Keelan et al., 1998; Mikulincer and Nachshon, 1991). The present study suggests that the interplay between both partners’ attachment styles shapes each individual’s cortisol responses and levels of felt closeness and desired closeness during friendship initiation. Cortisol and the Interplay of Attachment Styles The current work contributes to knowledge of attachment processes and their physiological correlates. Our results indicate that the interplay between partners’ attachment styles is linked to their hormonal responses during friendship initiation. Specifically, the interaction between partners’ levels of attachment avoidance predicted their cortisol patterns, with individuals showing the highest cortisol levels when there was a mismatch between their own and their partners’ levels of attachment avoidance (that is, when they

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

12

were low in avoidance and their partners were high in avoidance or when they were high in avoidance and their partners were low in avoidance). These findings fit well with previous evidence linking insecure attachment to dysregulated cortisol responses before, during, and after relationship conflict discussions (Beck et al., 2013; Brooks et al., 2011; Laurent and Powers, 2007; Powers et al., 2006). These findings also complement recent research on how the interplay of romantic partners’ attachment styles contributes to their cortisol responses; this research found that newlywed couples with one anxious partner and one avoidant partner showed increased cortisol reactivity in anticipation of a relationship conflict (Beck et al., 2013). Together, both this research and the present research emphasize the importance of considering how the interplay between partners’ attachment styles might shape their cortisol patterns in dyadic interactions. Furthermore, the present research extends previous research by investigating these questions in a new and previously unexplored relationship context: that of friendship initiation. Closeness and the Interplay of Attachment Styles Our findings also indicate that the interplay of both partners’ attachment styles shaped each individual’s felt closeness and desired closeness during friendship initiation. Specifically, the interaction between both partners’ levels of attachment anxiety predicted how much closeness they felt with one another, as well as how much closeness they wanted to feel with one another (as measured by partners’ actual and desired inclusion of other in the self). Individuals who were lower in attachment anxiety both felt more closeness and wanted more closeness with partners who also were low in attachment anxiety, whereas individuals’ feelings of actual and desired closeness did not differ as a function of their own attachment anxiety when their partners were high in attachment anxiety. These results indicate that people both felt and wanted the most closeness with partners when both they and their partners were low in attachment anxiety. These results also support the attachment–security hypothesis (Chappell and Davis, 1998), which proposes that individuals are most attracted to secure potential partners rather than to partners with complementary attachment styles.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

13

Our findings also align well with previous studies showing the detrimental effects of high attachment anxiety on relationship initiation (although we note that these studies did not consider dyadic effects by examining the potential partner’s attachment style). For example, McClure and Lydon (2014) found that people high in attachment anxiety displayed social disengagement and manifest anxiety in the face of relational opportunities such as speed-dating events or conversations with attractive potential partners. Additional research on the role of attachment anxiety during relationship initiation suggests that more anxiously-attached individuals are less popular than their secure counterparts in speed-dating settings (McClure et al., 2010), more dominating during laboratory tasks with same-sex strangers (Roisman, 2006), more self-protective rather than affiliative with new peers (Feeney et al., 2008), and more likely to share negative information and have higher levels of dissatisfaction with dating partners (Bradford et al., 2002). Furthermore, in the present study, individuals both felt less closeness and wanted less closeness with potential partners when either the individuals, their partners, or both were high in attachment anxiety. Negative interpersonal outcomes early in relationship initiation, whether in romantic or friendship contexts, likely reinforce anxiously-attached individuals’ negative relational expectations. Further, these negative interpersonal outcomes likely hinder the formation of new relationships, which, in turn, may influence how individuals connect to social resources and seek social support. For instance, previous research indicates that attachment styles influence people’s ability to make use of social support, such that insecurely-attached individuals are less able to reach out for and use social support than are securelyattached individuals (Ciechanowski et al., 2002; Maunder and Hunter, 2001). Attachment styles also influence people’s ability to derive benefits from the social support they have received. For example, one study (Ditzen et al., 2008) found that males low in attachment anxiety and avoidance displayed the lowest levels of anxiety in response to a social stressor when they received social support from their romantic partner. Results from this study also suggest that social support may be less likely to buffer stress for individuals high in attachment anxiety and avoidance. The ability to seek and use social support is especially important, given that social support has been linked with lowered anxiety levels, reduced

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

14

cortisol responses to psychosocial stressors (Ditzen et al., 2008), and better physical health outcomes (e.g., Uchino, 2004). Future work examining the interplay of attachment styles between partners providing and receiving social support could yield a more nuanced understanding of both social support and relationship initiation processes. 5. Caveats and Conclusions The present findings complement previous theoretical and empirical work suggesting that the interplay of partners’ attachment styles influences their actual and desired closeness and cortisol patterns. However, we cannot assume a causal relationship due to the correlational nature of the data. Furthermore, we used a predominantly female sample of participants. Although we would not anticipate finding gender differences in our effects, exploring potential differences with a more gender-representative sample remains an important direction for future research. Further, although salivary samples were assayed for cortisol using the same enzyme immunoassay kit, the two conditions were analyzed in different labs. This introduces a potential for source of error in that the labs may employ different assay techniques. Future studies should aim to minimize this potential source of bias by employing the same assay technique across conditions. In addition, we investigated how the interplay between individuals’ attachment styles might shape their cortisol patterns, but related research has investigated how differences in individuals’ cortisol patterns (e.g., increasing, decreasing, or stable cortisol levels over time) might shape their outcomes (e.g., Thompson and Trevathan, 2008). Therefore, exploring how individual differences in cortisol patterns might shape friendship initiation processes could further contribute to the literature. Finally, we examined friendship initiation processes among young adults; testing whether our effects generalize to other age groups will clarify how these processes unfold at different points in the lifespan. Relationships are dyadic in nature, so it is important to consider how both partners’ attachment styles shape one another’s emotional and hormonal responses. Recent research indicates that both partners in a relationship may shape each other’s stress responses (e.g., Beck et al., 2013; Laurent and Powers, 2007; Powers et al., 2006). The present study shows that the interplay of partners’ attachment styles in a

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

15

dyad is linked with their actual closeness, desired closeness, and cortisol responses during friendship initiation.

Funding: This work was supported by a grant from the University of Hartford.

Conflicts of interest: None

Footnotes 1

We also tested each of these three models within each closeness condition (i.e., within the high

closeness condition and within the low closeness condition), following Ketay et al. (2017).

(See

Supplementary File 1 for tables and figures for models within each closeness condition.) These models yielded results similar to the models across both closeness conditions (i.e., controlling for closeness condition), with two exceptions: (1) the interaction between participants’ attachment anxiety no longer significantly predicted participants’ actual closeness within the high closeness or low closeness condition and (2) the interaction between participants’ attachment anxiety no longer significantly predicted participants’ desired closeness within the low closeness condition. 2

The following analyses also were conducted using Winsorized cortisol values, in which values more

than three standard deviations above the mean were replaced with values three standard deviations above the mean (Reifman & Keyton, 2010); a base-10 logarithmic transformation was then used to normalize the data. Results did not differ when participants’ data were Winsorized rather than removed (treated as missing).

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

16

References Aardal E., Holm, A-C. 1995. Cortisol in saliva—Reference ranges and relation to cortisol in serum. Eur. J. Clin. Chem. Clin. Biochem. 33, 927–932. Aiken, L.S., West, S.G., 1991. Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage Publications. Aron, A., Aron, E.N., Smollan, D., 1992. Inclusion of other in the self scale and the structure of interpersonal closeness. J. Pers. Soc. Psychol. 63, 596–612. Aron, A., Aron, E., Tudor, M., Nelson, G., 1991. Close relationships as including other in the self. J. Pers. Soc. Psychol. 60, 241-253. Aron, A., Melinat, E., Aron, E. N., Vallone, R., Bator, R., 1997. The experimental generation of interpersonal closeness: A procedure and some preliminary findings. Pers. Soc. Psychol. B. 23, 363-377. Aspelmeier, J., Kerns, K., 2003. Love and School: Attachment/Exploration Dynamics in College. J. Soc. Pers. Relatsh. 20, 5-30. Bartholomew, K., Horowitz, L. M., 1991. Attachment styles among young adults: A test of a four category model. J. Pers. Soc. Psychol. 61, 226-244. Baumeister, R., Leary, M., 1995. The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychol. Bull. 117, 497-529. Beck, L., Clark, M., 2009. Choosing to Enter or Avoid Diagnostic Social Situations. Psychol. Sci. 20, 1175-1181. Beck, L., Pietromonaco, P., DeBuse, C., Powers, S., Sayer, A., 2013. Spouses’ attachment pairings predict neuroendocrine, behavioral, and psychological responses to marital conflict. J. Pers. Soc. Psychol. 105, 388-424. Bradford, S., Feeney, J., Campbell, L., 2002. Links between attachment orientations and dispositional and diary-based measures of disclosure in dating couples: A study of actor and partner effects. Pers. Relatsh. 9, 491-506.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

17

Brennan, K.A., Clark, C.L., Shaver, P.R., 1998. Self-report measurement of adult romantic attachment: An integrative overview. In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 46-76). New York: Guilford Press. Brooks, K.P., Robles, T.F., Dunkel Schetter, C., 2011. Adult attachment and cortisol responses to discussions with a romantic partner. Pers. Relatsh. 18, 302–320. Cacioppo J.T., Hughes, M.E., Waite, L.J., Hawkley, L.C., Thisted, R.A., 2006. Loneliness as a specific risk factor for depressive symptoms: Cross-sectional and longitudinal analyses. Psychol. Aging. 21, 140–151. Chappell, K. D., Davis, K. E. 1998. Attachment, partner choice, and perception of romantic partners: An experimental test of the attachment-security hypothesis. Pers. Relatsh. 5, 327-342. Ciechanowski, P.S., Walker, E.A., Katon, W.J., Russo, J.E., 2002. Attachment theory: a model for health care utilization and somatization. Psychosom. Med. 64, 660–7. Creasey, G., 2002. Associations between working models of attachment and conflict management behavior in romantic couples. J. Counsel. Psychol. 49, 365–375. Dewitte M., De Houwer, J., Goubert, L., Buysse, A., 2010. A multi modal approach to the study of attachment related distress. Biol. Psychol. 85, 149–162. Ditzen B., Schmidt S., Strauss B., Nater U.M., Ehlert U., Heinrichs M. 2008. Adult attachment and social support interact to reduce psychological but not cortisol responses to stress. J. Psychom. Res. 64, 479–486. Feeney, J.A., 2003. The systemic nature of couple relationships: An attachment perspective. In P. Erdman & T. Caffery (Eds.), Attachment and family systems: Conceptual, empirical, and therapeutic relatedness (pp. 139–163). New York, NY: Brunner-Routledge. Feeney, B.C., Cassidy, J., Ramos-Marcuse, F., 2008. The generalization of attachment representations to new social situations: Predicting behavior during initial interactions with strangers. J. Pers. Soc. Psychol. 95, 1481-1498.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

18

Grabill, C., Kerns, K., 2000. Attachment style and intimacy in friendship. Pers. Relatsh. 7, 363378. Granger, D.A., Schwartz, E.B., Booth, A., Arentz, M., 1999. Salivary testosterone determination in studies of child health and development. Horm. Behav. 35, 18–27. Granger, D.A., Shirtcliff, E.A., Booth, A., Kivlighan, K., Schwartz, E.B., 2004. The trouble with salivary testosterone. Psychoneuroendocrinology. 29, 1229–1240. Griffin, D., Bartholomew, K., 1994. Models of the self and other: Fundamental dimensions underlying measures of adult attachment. J. Pers. Soc. Psychol. 67, 430- 445. Holt-Lunstad, J., Smith, T., Layton, J., 2010. Social Relationships and Mortality Risk: A Metaanalytic Review. Plos Medicine. 7(7), e1000316. Kashy, D.A., Kenny, D.A., 2000. The analysis of data from dyads and groups. In H. Reis & C. M. Judd (Eds.), Handbook of research methods in social psychology (pp. 451–477). New York, NY: Cambridge University Press. Keelan, J., Dion, K., Dion, K., 1998. Attachment style and relationship satisfaction: Test of a selfdisclosure explanation. Canadian J. Behav. Sci./Revue Canadienne Des Sciences Du Comportement, 30, 24-35. Kenny, D.A., Kashy, D.A., Cook, W.L., 2006. Dyadic data analysis. New York, NY: Guilford Press. Ketay, S., Welker, K., Slatcher, R., 2017. The roles of testosterone and cortisol in friendship formation. Psychoneuroendocrinology. 76, 88-96. Kidd T., Hamer, M., Steptoe, A., 2011. Examining the association between adult attachment style and cortisol responses to acute stress. Psychoneuroendocrinology. 36, 771–779. Laurent, H., Powers, S. I., 2007. Emotion regulation in emerging adult couples: Temperament, attachment, and HPA response to conflict. Biol. Psychol. 76, 61–71.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

19

Maunder, R.G., Hunter, J.J., 2001. Attachment and psychosomatic medicine: developmental contributions to stress and disease. Psychosom Med. 63, 556–67. McClure, M., Lydon, J., 2014. Anxiety doesn’t become you: How attachment anxiety compromises relational opportunities. J. Pers. Soc. Psychol. 106, 89-111. McClure, M., Lydon, J., Baccus, J., Baldwin, M., 2010. A signal detection analysis of chronic attachment anxiety at speed dating: Being unpopular is only the first part of the problem. Pers. Soc. Psychol. B. 36, 1024-1036. McWilliams, L.A., Bailey, S.J., 2010. Association between adult attachment ratings and health conditions: Evidence from the survey replication. Health Psychol. 29, 446–453. Mikulincer, M., Florian, V., 1999. Maternal–fetal bonding, coping strategies, and mental health during pregnancy: The contribution of attachment style. J. Soc. Clin. Psychol. 18, 255– 276. Mikulincer, M., Florian, V., Cowan, P.A., Cowan, C.P., 2002. Attachment security in couple relationships: A systemic model and its implications for family dynamics. Fam. Process. 41, 405–434. Mikulincer, M., Nachshon, O., 1991. Attachment styles and patterns of self-disclosure. J. Pers. Soc. Psychol. 61, 321-331. Mikulincer, M., Shaver, P.R., 2007. Attachment in adulthood: Structure, dynamics, and change. New York, NY: Guilford Press. Page-Gould, E., Mendoza-Denton, R., Tropp, L., 2008. With a little help from my cross-group friend: Reducing anxiety in intergroup contexts through cross-group friendship. J. Pers. Soc. Psychol. 95, 1080-1094. Paley, B., Cox, M.J., Burchinal, M.R., Payne, C., 1999. Attachment and marital functioning: Comparison of spouses with continuous-secure, earned-secure, dismissing, and preoccupied attachment stances. J. Fam. Psychol. 13, 580–597.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

20

Pietromonaco, P.R., Beck, L.A., 2015. Attachment processes in adult romantic relationships. In M. Mikulincer, P. R. Shaver, J. A. Simpson, & J. F. Dovidio (Eds.), APA handbook of personality and social psychology, Volume 3: Interpersonal relations. (pp. 33–64). Washington, DC, US: American Psychological Association. Powers, S.I., Pietromonaco, P.R., Gunlicks, M., Sayer, A., 2006. Dating couples’ attachment styles and patterns of cortisol reactivity and recovery in response to relationship conflict. J. Pers. Soc. Psychol. 90, 613–628. Quirin, M., Pruessner, J.C., Kuhl, J., 2008. HPA system regulation and adult attachment anxiety: Individual differences in reactive and awakening cortisol. Psychoneuroendocrinology. 33, 581–590. Rasco, D., Warner, R., 2016. Relationship authenticity partially mediates the effects of attachment on relationship satisfaction. J. Soc. Psychol., 1-13. Reifman, A., Keyton, K., 2010. Winsorize. In N.J. Salkind (Ed.), Encyclopedia of Research Design (pp. 1636–1638). Thousand Oaks: Sage. Roisman, G.I., 2006. The role of adult attachment security in non-romantic, non-attachmentrelated first interactions between same-sex strangers. Attach. Hum. Dev. 8, 341-352. Schultheiss, O.C., Stanton, S.J., 2009. Assessment of salivary hormones. In: Harmon-Jones, E., Beer, J.S. (Eds.), Methods in Social Neuroscience. Guilford, New York, NY, pp. 17–44. Stansbury, K., Gunnar, M., 1994. Adrenocortical activity and emotion regulation. In N. Fox (Ed.), Monographs of the Society for Research in Child Development: Vol. 59: The development of emotion regulation: Biological and behavioral consequences (pp. 108– 134). Chicago: Society for Research in Child Development. Steele, J.L., 1999. Cognitive Mechanisms of Attitude Change in Close Relationships. Unpublished doctoral Dissertation, State University of New York at Stony Brook, Stony Brook, NY.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

21

Thompson, L.A., Trevathan, W.R. 2008. Cortisol reactivity, maternal sensitivity, and learning in three-month-old infants. Infant Behav Dev. 31, 92-106. Uchino, B.N. 2004. Social Support and Physical Health: Understanding the Health Consequences of Relationships. New Haven, CT: Yale University Press. Vorauer, J., Cameron, J., Holmes, J., Pearce, D., 2003. Invisible overtures: Fears of rejection and the signal amplification bias. J. Pers. Soc. Psychol. 84, 793-812.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

22

Figure 1. Actors lower in attachment avoidance had significantly lower cortisol levels when they had partners who also were low in attachment avoidance than did actors higher in attachment avoidance.

However, actors lower in attachment avoidance had

significantly higher cortisol levels when they had partners who were high in attachment avoidance than did actors higher in attachment avoidance.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

23

Figure 2. Actors lower in attachment anxiety felt significantly closer to partners who also were low in attachment anxiety than did actors higher in attachment anxiety. However, actors’ feelings of closeness did not differ as a function of their own attachment anxiety when their partners were high in attachment anxiety.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

24

Figure 3. Actors lower in attachment anxiety desired significantly more closeness with partners who also were low in attachment anxiety than did actors higher in attachment anxiety. However, actors’ desired closeness did not differ as a function of their own attachment anxiety when their partners were high in attachment anxiety.

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

25

Table 1. Final estimation of predictors of cortisol levels Predictor Intercept Actor effects Attachment avoidance Attachment anxiety Partner effects Attachment avoidance Attachment anxiety Interaction effects Actor avoidance x partner anxiety Actor anxiety x partner avoidance Actor avoidance x partner avoidance Actor anxiety x partner anxiety Condition Time (centered at third saliva sample) Gender Hormonal contraceptive use

Estimate -0.80

SE 0.02

t(df) -33.27 (271.69)

p < .001

95% CI [-0.85, -0.75]

0.01 -0.00

0.00 0.00

2.01 (528.36) -1.57 (524.75)

.04 .12

[0.00, 0.01] [-0.01, 0.00]

0.00 -0.01

0.00 0.00

0.18 (529.09) -2.20 (524.37)

.86 .03

[-0.01, 0.01] [-0.01, -0.00]

-0.00 0.00 -0.00 0.00 0.13 -0.17 -0.11 -0.04

0.00 0.00 0.00 0.00 0.02 0.03 0.03 0.03

-0.70 (536.23) 0.24 (535.43) -5.50 (266.47) 1.26 (269.80) 6.09 (270.97) -6.85 (271.59) -4.21 (285.58) -1.30 (480.38)

.49 .81 < .001 .21 < .001 < .001 < .001 .20

[-0.00, 0.00] [-0.00, 0.00] [-0.01, -0.00] [-0.00, 0.00] [0.09, 0.17] [-0.22, -0.12] [-0.16, -0.06] [-0.09, 0.02]

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

26

Table 2. Final estimation of predictors of actual inclusion of other in the self (IOS) Predictor Intercept Actor effects Attachment avoidance Attachment anxiety Partner effects Attachment avoidance Attachment anxiety Interaction effects Actor avoidance x partner anxiety Actor anxiety x partner avoidance Actor avoidance x partner avoidance Actor anxiety x partner anxiety Condition

Estimate 3.25

SE 0.20

t(df) 15.98 (70.00)

p < .001

95% CI [2.84, 3.65]

-0.06 -0.04

0.03 0.03

-1.89 (143.74) -1.51 (143.33)

.06 .13

[-0.13, 0.00] [-0.09, 0.01]

0.03 -0.07

0.03 0.03

0.87 (143.74) -2.58 (143.33)

.38 .01

[-0.04, 0.10] [-0.12, -0.02]

-0.01 -0.00 -0.01 0.02 1.67

0.01 0.01 0.01 0.01 0.27

-1.39 (142.82) -0.29 (142.82) -0.73 (70.00) 2.40 (70.00) 6.16 (70.00)

.17 .77 .47 .02 < .001

[-0.02, 0.00] [-0.02, 0.01] [-0.03, 0.01] [0.00, 0.03] [1.13, 2.21]

ATTACHMENT PREDICTS CORTISOL RESPONSE AND CLOSENESS

27

Table 3. Final estimation of predictors of desired inclusion of other in the self (IOS) Predictor Intercept Actor effects Attachment avoidance Attachment anxiety Partner effects Attachment avoidance Attachment anxiety Interaction effects Actor avoidance x partner anxiety Actor anxiety x partner avoidance Actor avoidance x partner avoidance Actor anxiety x partner anxiety Condition

Estimate 3.65

SE 0.19

t(df) 19.39 (70.00)

p < .001

95% CI [3.27, 4.02]

-0.05 -0.05

0.03 0.03

-1.38 (140.31) -1.66 (141.34)

.17 .10

[-0.12, 0.02] [-0.10, 0.01]

0.05 -0.05

0.03 0.03

1.45 (140.31) -1.89 (141.34)

.15 .06

[-0.02, 0.12] [-0.11, 0.00]

-0.01 -0.00 -0.01 0.02 1.25

0.01 0.01 0.01 0.01 0.25

-0.82 (142.11) -0.46 (142.11) -0.73 (70.00) 2.72 (70.00) 4.97 (70.00)

.41 .65 .47 .01 < .001

[-0.02, 0.01] [-0.02, 0.01] [-0.03, 0.01] [0.00, 0.03] [0.75, 1.75]