Abstract
Due to the personal and relationship costs associated with attachment anxiety, there has been substantial interest in identifying factors that reduce attachment anxiety over time and buffer attachment anxiety in the moment. We integrated the Attachment Security Enhancement Model (ASEM) and the dependency paradox based in attachment theory to derive novel predictions about how perceived partner support predicts attachment anxiety prospectively (one year later) and concurrently. Newlyweds (N = 326) reported their perceptions that their partners support their personal goals (i.e., perceived support for independence needs) and their perceptions that their partners provide reassurance/responsiveness (i.e., perceived support for dependence needs). We found that greater perceived support for personal goals predicted decreases in attachment anxiety over one year only among people who also perceived high partner reassurance/responsiveness, consistent with the idea that support for independence is most effective when partners also support dependence needs. We also observed evidence that this pattern was most pronounced among people with higher attachment anxiety. Regarding concurrent links, we found that perceiving greater partner reassurance/responsiveness predicted lower concurrent attachment anxiety, whereas perceived support for independent goals did not. Together, these findings extend the ASEM by highlighting that attachment anxiety declines over time when partners are perceived to support one’s needs for independence and dependence simultaneously.
Optimal personal and relationship outcomes result from an individual’s ability to a) confidently and authentically explore the environment (e.g., pursue independent goals, accept important opportunities for growth) and to b) rely on close others for support (e.g., reassurance, practical assistance) in response to genuine need (B. C. Feeney et al., 2015; B. C. Feeney & Collins, 2015). Though some people are able to achieve this independence-dependence balance, many people are either over-reliant on close others to the detriment of their personal pursuits or are under-reliant on close others and miss opportunities for rewarding connection.
These individual differences are well-captured by attachment orientation—patterns of behavior, expectations, preferences, and goals in the context of relationships (e.g., Mikulincer & Shaver, 2016; Simpson et al., 2021). Attachment orientation is modernly defined around two dimensions of attachment insecurity called attachment anxiety and attachment avoidance (Brennan et al., 1998). Attachment anxiety—the focus of the current investigation—is characterized by excessive worry about relationships, including persistent concerns about being abandoned and rejected and fears about being unworthy of love and care (e.g., Arriaga et al., 2018; Fraley & Shaver, 2000; Mikulincer & Shaver, 2016). People with high attachment anxiety use hyperactivating strategies to safeguard their relationships, achieve desired closeness, and manage their distress; these strategies include vigilance for relationship threats, exaggeration of needs for support and protection, and “clinging” to partners. People who are high in attachment anxiety can be considered over-dependent in that they have learned that prioritizing closeness to and reliance on others (even at a cost to their independent pursuits) is the best and only way to maintain supportive relationships and their own safety (B. C. Feeney et al., 2015; Mikulincer et al., 2003). Attachment avoidance is characterized by distrust in others, discomfort with intimacy, and a resulting desire to remain emotionally distant and autonomous (e.g., Arriaga et al., 2018; Fraley & Shaver, 2000; Mikulincer & Shaver, 2016). People with high attachment avoidance can therefore be considered under-dependent, as they have learned that being dependent on others is punishing and that defensive self-reliance is the best and only way to ensure that their needs are met (B. C. Feeney et al., 2015; Mikulincer et al., 2003).
People who have absent or low levels of attachment insecurity (termed secure attachment) experience considerable personal and relationship benefits relative to those who are insecurely attached (e.g., Mikulincer & Shaver, 2019; Overall et al., 2022). For example, people who are securely attached are more able to manage their personal stressors and interpersonal conflicts and tend to experience more satisfying relationships than people with insecure attachment orientations (J. A. Feeney & Karantzas, 2017; Kirkpatrick & Davis, 1994). Compared to people with higher attachment anxiety, in particular, people with lower attachment anxiety experience less conflict, less rumination and distress, and better health (Campbell et al., 2005; Stanton & Campbell, 2014).
Unsurprisingly then, researchers and theorists have prioritized identifying ways to a) buffer the negative consequences of attachment insecurity in the moment and b) promote attachment security (i.e., reduce forms of attachment insecurity) over time (Fraley et al., 2011; Gillath & Karantzas, 2019; Mikulincer & Shaver, 2020; Overall & Simpson, 2015). The attachment security enhancement model (ASEM; Arriaga et al., 2018) provides the most comprehensive explanation of strategies that, theoretically, should buffer attachment insecurity in the moment and reduce attachment insecurity over time. In brief, the ASEM proposes that short-term buffering is achieved when insecurely attached individuals receive support that is sensitive to their attachment-related concerns (Arriaga et al., 2018). For anxiously attached individuals, this means receiving support for their heightened dependence needs via “safe strategies”: behaviors that reaffirm the partner’s commitment to the relationship and provide reassurance, affection, or comfort to reduce distress. For avoidantly attached individuals, partner buffering means receiving support for their heightened independence needs via “soft strategies”: behaviors that defuse intense emotional interactions, allow self-reliance, and respect autonomy. Empirical evidence indeed shows that insecurely attached individuals prefer and benefit from these targeted partner buffering strategies during insecurity-triggering situations (Farrell et al., 2016; Fuentes et al., 2022; Kim et al., 2018; Lemay & Dudley, 2011; Overall et al., 2014; Simpson & Overall, 2014; Tran & Simpson, 2009).
Critically, the ASEM posits that although these buffering strategies offer immediate relief, they will not promote security in the long-term and may instead inadvertently reinforce insecure attachment tendencies. According to the ASEM, actual insecurity reduction is achieved when people have new, perhaps even uncomfortable, experiences that allow them to revise their existing working models: the underlying beliefs and expectations people have about themselves and others that drive their attachment behavior (e.g., Baldwin et al., 1996; Bowlby, 1973; Mikulincer & Shaver, 2016). Next, we review the internal working model that underlies attachment anxiety and the ASEM’s propositions related to revising it in order to reduce attachment anxiety. Although the ASEM proposes strategies to produce long-term reductions in attachment anxiety and attachment avoidance, we focus on attachment anxiety in this investigation because it is the propositions regarding attachment anxiety change, in particular, that we think should be reconsidered in light of another related theory—the dependency paradox based in attachment theory (B. C. Feeney, 2007)—which we review subsequently.
Reducing attachment anxiety
People with high attachment anxiety are excessively worried about their relationships: they fear that others will not be available, will not be responsive to their needs, and/or will not remain committed to the relationship (e.g., Arriaga et al., 2018; Fraley & Shaver, 2000; Mikulincer & Shaver, 2016). Accordingly, they are excessively vigilant for others’ disengagement and are preoccupied with staying proximal to and emotionally intimate with close others. Importantly, the anxious attachment pattern is thought to be rooted primarily in a negative working model of self, including low self-efficacy and low perceived self-value (Mikulincer & Shaver, 2016; Arriaga et al., 2018). In other words, people high in attachment anxiety believe that they lack the ability to navigate life’s challenges alone and believe that they are inherently unworthy of love and care. It is these foundational beliefs that motivate people with high attachment anxiety to prioritize proximity and dependence on close others and to be vigilant to rejection (Bartholomew & Horowitz, 1991; Simpson & Rholes, 2019).
The ASEM proposes that although a partner’s reassurance/responsiveness (e.g., availability, affection, comfort) may feel good in the moment, it may also inadvertently exacerbate attachment anxiety by reinforcing the idea that partner support is needed. Thus, the ASEM proposes that attachment anxiety will decrease when individuals pursue activities that enable them to view themselves as valuable and capable (i.e., develop a more secure model of self; Arriaga et al., 2018). It is these autonomous pursuits that have the potential to rectify the foundational “problem” of attachment anxiety: a negative model of self. When people have positive experiences apart from close others (in personal, independent domains), they can prove to themselves that they have value and are capable. That foundational shift in self-beliefs will theoretically enable people to be less concerned about rejection or abandonment because they will be more confident in their own lovability and in their ability to take care of themselves if they were to be rejected. Partners can therefore help to facilitate declines in attachment anxiety by encouraging independent goal pursuit and other forms of autonomous exploration that provide an opportunity to develop personal self-efficacy and enhance self-worth.
Some initial empirical research supports this long-term proposition and the differentiation between what is effective to buffer attachment anxiety concurrently and to produce change over time. Arriaga and colleagues (2021) demonstrated that greater parenting self-efficacy during the transition to parenthood predicted greater decreases in attachment anxiety over 6 months, whereas receiving reassurance from one’s partner did not predict decreases in attachment anxiety prospectively. Instead, reassurance predicted lower attachment anxiety concurrently (whereas parenting self-efficacy did not). In a related investigation, Arriaga et al., (2014) found that the more people perceived that their partners validated their goals, the greater decreases in attachment anxiety they reported over one year. Trust in one’s partner (i.e., the perception that the partner will accept one’s need for dependence) instead only predicted lower attachment anxiety concurrently. These findings support the ASEM’s tenets that a) partners’ reassurance buffers insecurity in the short-term but that b) partners’ encouragement and validation of independent goals, especially if new efforts enable individuals to develop self-efficacy, is the catalyst for decreases in attachment anxiety. In other words, these findings seem to suggest that dependence begets dependence whereas support for independence promotes attachment change.
Reconsidering the implications of partner dependence
Despite this initial evidence in favor of the ASEM, there is reason to believe that perceiving partner support for independence alone could have unintended negative consequences that undermine attachment anxiety change. Specifically, a partner’s support for independent goals (e.g., “I think it would be great if you signed up for a night class”) could seem to signal the partner’s desire for distance in the relationship and could therefore be interpreted as a relationship threat. People who are high in attachment anxiety are especially likely to perceive support for independence as a relationship threat because they, by definition, are hyper-vigilant to signs of rejection and are highly motivated to maintain relationship closeness (Overall & Lemay, 2015). Accordingly, perceiving that a partner is encouraging one’s independent goals may lead people with moderate or high levels of attachment anxiety to reflexively cling to their partners instead of pursuing the independent experiences that could enable them to challenge their anxious attachment beliefs. Indeed, Jakubiak and Feeney (2016) found that partner support for goals was less strongly linked to next-day goal progress for people with greater attachment anxiety compared to those with lower attachment anxiety. Although the ASEM suggests that anxiously attached people need to be challenged to be independent in order to develop attachment security, this finding suggests that perceiving partner support for independence may be ineffective to bring about change if it is not accompanied by reassurance that the relationship is secure.
Attachment theory emphasizes the dynamic interplay between dependence and independence and specifically proposes that close others provide a foundation for successful self-development by being responsive, available, and reassuring in both contexts (Bowlby, 1988). In other words, “true independence and self-sufficiency emerge because of an individual’s ability to depend on close relationship partners in times of need” (B. C. Feeney, 2007, p. 268). This idea, termed the “dependency paradox,” has received support in several empirical investigations (B. C. Feeney, 2007; Hazen & Durrett, 1982; Jakubiak & Feeney, 2016). In a direct test of the dependency paradox, Feeney (2007) found that people were most able to be self-sufficient (e.g., engaged in the most independent exploration; were more likely to achieve their goals) when their partners were attentive and responsive to their needs. This research suggests that independence can best (and perhaps only) be achieved in contexts where people perceive acceptance of and support for their dependence needs (B. C. Feeney, 2007).
Accordingly, connecting this idea of a dependency paradox to the ASEM suggests that people may experience the greatest reductions in attachment anxiety over time when their partners both encourage independence and demonstrate their availability and willingness to be responsive, as needed. In fact, this is exactly how a secure base functions (B. C. Feeney, 2004; B. C. Feeney & Thrush, 2010). Close others provide a secure base for exploration (also referred to relational catalyst support) by encouraging independence while simultaneously remaining available and being non-interfering (B. C. Feeney et al., 2017; B. C. Feeney & Collins, 2015; B. C. Feeney & Thrush, 2010; Tomlinson et al., 2015). The availability component of a secure base is critical and includes offering reassurance related to exploration attempts.
The idea that support for independence should be accompanied by support for dependence is not in contrast to the ASEM (Arriaga et al., 2018), but is in contrast with empirical tests that pit short-term strategies against long-term strategies (Arriaga et al., 2014; 2021). The original ASEM explanation stated that “anxious individuals who feel distress may temporarily benefit from their partner’s reassurance, but reassurance alone might ‘feed’ insecurity” (emphasis added, Arriaga et al., 2018, p. 85). We argue that, although reassurance (i.e., an acceptance and support of dependence) may not be sufficient to produce reductions in attachment anxiety, it may nonetheless be necessary.
Current research
In the current research, we assessed change in attachment anxiety in a sample of newlywed couples who completed baseline and follow-up assessments one year apart, as part of a larger study on Newlywed relationships. At baseline, participants identified their most important independent goal, and at follow-up, they reported on the extent to which their partner encouraged and supported that goal over the year between baseline and follow-up. We expected to conceptually replicate Arriaga and colleagues’ previous findings (Arriaga et al., 2014; 2021) showing that perceived partner support for this independent goal predicts decreases in attachment anxiety over the year (H1). Additionally, integrating ideas from the ASEM and the dependency paradox, we made the novel prediction that perceived partner support for one’s independent goal will be more strongly associated with reductions in attachment anxiety over time if partners are also perceived to provide reassurance/responsiveness (H2). Because people with high attachment anxiety may be most likely to misinterpret a partner’s independent goal support as a relationship threat and may be most concerned about the risks of exploring away from their partner, we also assessed whether the hypothesized interaction was particularly evident among people with high attachment anxiety at baseline.
We conceptualized and measured perceived partner reassurance/responsiveness (PPR/R) at two time-points. First, we assessed PPR/R at baseline, prior to goal support provision occurring. This enabled us to assess the consequences associated with personal goal support in a context in which dependency needs were already being met to a greater or lesser extent. This approach is consistent with the idea that perceiving partner responsiveness is a prerequisite that enables one to accept support for independence. Second, we assessed the extent to which participants perceived reassurance/responsiveness in the time between the baseline and follow-up assessments through retrospective reports at follow-up. This enabled us to assess the consequences associated with perceiving support for one’s independent goal and co-occurring PPR/R. This approach is more consistent with the idea that a balance of support for dependence and independence needs promotes attachment security. We expected comparable results with both approaches. 1
As a secondary study aim, we tested how perceived partner support for independent goals and PPR/R predict concurrent attachment anxiety at baseline. Arriaga et al. (2021) found that perceived partner reassurance—but not support for independent goals—predicted concurrent attachment anxiety. However, because a balance between dependence and independence may prevent over-reliance on one’s partner, we also tested whether concurrent attachment anxiety is lower when people not only perceived that their partners provide responsive reassurance but also perceive support for independent goals. In the current research, we expected to conceptually replicate Arriaga et al. (2021) by showing that greater PPR/R predicts lower concurrent attachment anxiety (H3). Additionally, we made the novel prediction that perceiving both reassurance/responsiveness and support for independent goals will be associated with the lowest concurrent attachment anxiety (H4).
Although our theoretical model focused on attachment anxiety, 2 we also assessed whether perceived partner support for independent goals and/or PPR/R would predict concurrent attachment avoidance or change in attachment avoidance over time in exploratory models. Past empirical work has shown that perceived support for independent goals predicts lower attachment avoidance concurrently (Arriaga et al., 2014), but it is possible that a balance of support perceptions could prevent under-reliance on one’s partner and be associated with lower concurrent attachment avoidance. Accordingly, we explored whether perceived support for independent goals is most strongly associated with low attachment avoidance for people who also perceive that their partners are reassuring/responsive.
Regarding change over time, the ASEM posits that attachment avoidance will decrease when individuals come to view others as responsive and trustworthy (i.e., revise their negative model of others to develop a more secure model of others). In past work, attachment avoidance has been shown to decline when partners report receiving more responsive support than usual (Rholes et al., 2020) or when they have positive or intimate interpersonal interactions (Stanton et al., 2017). Based on the idea that people benefit from a balance of support for dependence and independence, one might expect that the association between PPR/R and declines in attachment avoidance would be stronger for people who also perceive support for their independent goals. However, we hesitate to make this prediction (and therefore opted to focus on attachment anxiety in this research) because it runs counter to theoretical explanations of the interworking of the attachment and exploration systems. Specifically, attachment is considered a prerequisite for prioritizing and sustaining exploration, making dependence a more foundational process than independence (Bowlby, 1988). In other words, although dependency enables independence (the dependency paradox, Feeney, 2007), we do not expect a comparable “independence paradox” whereby independence is necessary for people to prioritize dependence. We nonetheless explored whether PPR/R is associated with greater declines in attachment avoidance for people who also perceived a spouse to support their independent goal.
Method
Participants and procedure
Newlywed couples (N = 229, all different-sex) were recruited from Pittsburgh, PA as part of a larger study on newlywed relationships. Sample size was determined by a power analysis for alternate research hypotheses. Participants separately and privately completed self-report questionnaires in the laboratory at baseline and one year later. A total of 163 couples completed the follow up measures and are included in the current sample (no participants were excluded from analyses). These participants ranged in age from 18 to 54 (M = 27.16, SD = 4.66) and varied with regard to educational attainment (9.9% completed high school or less; 24.9% had earned some college credits or an Associate’s degree; 33.1% had a Bachelor’s degree; 15.7% had completed some graduate school; and 16.3% had a professional degree). The majority of participants identified as White/Caucasian (81.9%). Additionally, some participants identified as Black/African American (6.9%), Asian (4.6%), Hispanic (2.9%), American Indian (1.0%), or something else/multiple race/ethnicities (2.7%). Participants had been married for an average of 4.9 months (SD = 3.44; range = 0-21), and they had been romantically involved with their spouses for an average of 45.8 months (SD = 28.0), or approximately 3 and a half years.
The participants who completed the follow-up reported greater relationship quality (b = 0.25, p = .018) and marginally lower attachment avoidance (b = −0.19, p = .070) at baseline than participants who did not complete the follow-up. However, participants who completed the follow-up did not differ from those who did not complete the follow-up with regard to age (b = −0.53, p = .397), time married (b = 0.02, p = .963), or time romantically involved (b = −0.67, p = .837). Critically, participants who completed the follow-up did not differ from those who did not complete the follow-up with regard to baseline attachment anxiety (b = −0.03, p = .781). 3
Measures
Although participants completed a comprehensive battery of questionnaires at both time-points, we report only those relevant to the current investigation.
Baseline and follow-up attachment orientation
Scale reliabilities, descriptive statistics, and zero-order correlations between variables.
Note. These descriptive statistics and correlates are based on measures with outliers removed. The items for baseline PPR/R were standardized prior to computing the composite because they were on different response scales. PPR/R = Perceived partner reassurance/responsiveness. *p < .05, **p < .01.
Perceived support at baseline
At baseline, participants identified the most important goal they wanted to pursue over the next year. They then completed six items assessing the extent to which their spouse would support this goal (e.g., “How encouraging and supportive do you think your spouse would be of this goal?“). We calculate the mean of these responses to form a
At baseline, participants also completed a scale assessing how their spouse is likely to respond during three specific stressful situations (i.e., safe haven support opportunities) including when “you stay home sick in bed,” “you are crying or obviously distressed about something that has upset you,” and “you are in a situation in which you feel frightened or afraid.” Among other filler items, participants rated the extent to which their partner is likely to be “available,” “unresponsive” (reverse-scored), “sensitive to my needs,” and “approachable” in each situation on a scale from 1(not at all) to 7(very much), M = 6.19, SD = 0.78, ⍺ = .90. Participants also completed a shortened version of the Caregiving Scale (Kunce & Shaver, 1994), and we used the 6 items that assess Proximity versus Distance (e.g., “When I want or need a hug, my spouse seems glad to provide it”) and the 6 items that assess Sensitivity versus Insensitivity (e.g., “My spouse is very attentive to my nonverbal signals for help and support). The Proximity items capture proximal, physically affectionate, caregiving during times of need, and the Sensitivity items capture the partner’s ability to identify need by decoding verbal and nonverbal support-seeking signals. Participants rated their agreement with each statement on a scale from 1(disagree strongly) to 6(agree strongly), M = 4.78, SD = 0.78, ⍺ = .87. We created a
Perceived support between baseline and follow-up
At follow-up, participants were reminded of the most important goal they selected at baseline, and they answered questions regarding their experience of their goal in the time since the baseline assessment. For this study, we identified all questions that related to the spouse’s support of the participant’s goal pursuit in the time since the baseline assessment (e.g., “Over the past year, to what extent do you think your spouse wanted you to pursue this goal?“; “Over the past year, how encouraging and supportive was your spouse of this goal?” and “Over the past year, how negative, critical, or unsupportive was your spouse with regard to your goal?” (reverse-coded). Altogether, 12 items were averaged to form a composite variable representing
Participants also completed a scale assessing other things their spouse provided during the past year. Three items, “comfort and reassurance,” “romantic physical contact (hugging, kissing, etc.),” and “help when you’re sick or ill,” most clearly assessed reassurance/responsiveness. Participants responded to each item from 1 (has not provided at all) to 7(provides a lot). From these items, we calculated a composite (mean) variable representing
Data analytic strategy
We used multilevel modeling in R to account for the nested structure of these data (i.e., the dyadic nature of the data from romantic partners violates the assumption of independence in the data). Prior to conducting analyses, we removed outliers from each analyzed variable. Additionally, we standardized all variables so that regression coefficients serve as a measure of effect size, comparable to Cohen’s d.
For the lagged hypotheses (H1, H2), we tested whether a person’s attachment anxiety at follow-up (Year 2) is predicted by perceived partner support for independent goal pursuit between the baseline and follow-up assessments (reported at Year 2), PPR/R, and their interaction. In a first model, we included our measure of PPR/R at baseline, and in a second model, we included our measure of PPR/R between baseline and follow-up. In both models, we included baseline (Year 1) attachment anxiety in the model as an additional predictor so that results represent residualized change, and we included baseline attachment avoidance in the model as a control variable.
For our concurrent hypotheses (H3, H4), we tested whether a person’s attachment anxiety at baseline (Year 1) is predicted by PPR/R at baseline (Year 1), perceived partner support for independent goal pursuit at baseline (Year 1), and their interaction. We included baseline attachment avoidance in the model as a control variable.
We constructed additional models to assess our exploratory questions regarding whether perceived support predicts concurrent or prospective attachment avoidance. These models are the same as the models described above except that attachment avoidance becomes the outcome and attachment anxiety becomes the control variable. Analytic code is also provided in the OSM, and data are available by request to the first author.
Results
Predicting change in attachment anxiety over time
Lagged models: Change in attachment anxiety over time as a function of support variables.
Perceived support for independent goal pursuit and PPR/R at baseline predict residualized change in attachment anxiety. Note. Follow-up attachment anxiety is unstandardized in the figure for ease of interpretation, though it was standardized in analyses.
The observed interaction also can be decomposed with perceived support for independent goal pursuit as the moderator. Greater PPR/R at baseline predicted declines in attachment anxiety for participants who perceived high goal support (B = −0.20, p = .008), but not for participants who perceived low goal support (B = −0.00, p = .998). The results indicate that a combination of high PPR/R at baseline and high perceived support for independent goals between baseline and follow-up predicts declines in attachment anxiety.
The second lagged model—the model testing the interaction between perceived partner support for independent goal pursuit and PPR/R between baseline and follow-up—revealed a similar pattern of results. In line with H2, we observed an interaction between perceived support for independent goal pursuit and PPR/R (both between baseline and follow-up) to predict declines in attachment anxiety (B = −.09, p = .048). However, we also observed a significant three-way interaction between perceived support for independent goal pursuit, PPR/R, and baseline attachment anxiety to predict residualized change in attachment anxiety (see Table 2 and Figure 2a). In decomposing this interaction, we found that greater support for independent goal pursuit only predicted declines in attachment anxiety for participants who had high (+1 SD) attachment anxiety at baseline and perceived high (+1 SD) reassurance/responsiveness from their partners between baseline and follow-up (B = −.24, p = .023). If participants with high attachment anxiety instead perceived low (−1 SD) in reassurance/responsiveness between baseline and follow-up, goal support was unrelated to change in attachment anxiety (B = .14, p = .124), and the direction of this simple slope became positive. All other simple slopes of perceived goal support predicting declines in attachment anxiety were non-significant. Three-way interaction between perceived goal support, PPR/R between baseline and follow-up, and baseline attachment anxiety to predict residualized change in attachment anxiety. Note. These figures depict the same 3-way interaction with perceived support for independent goal pursuit and PPR/R treated as the predictor variable in Panels A and B, respectively. Follow-up attachment anxiety is unstandardized in the figure for ease of interpretation, though it was standardized in analyses.
Using this measure of PPR/R between baseline and follow-up, we also observed a main effect of PPR/R (see Supplemental Table 1). Because PPR/R may be a stronger predictor of declines in attachment anxiety than perceived support for independent goal pursuit, we also plotted and decomposed the three-way interaction treating PPR/R as the predictor variable to help us to better understand the nature of this three-way interaction (see Figure 2b). Testing the simple slopes in this way revealed that—for participants with average (M) or high (+1 SD) attachment anxiety at baseline—greater PPR/R between baseline and follow-up was associated with declines in attachment anxiety when perceived support for independent goal pursuit was also high (B = −.26, p = .001; B = −.39, p < .001, respectively) but not when perceived support for independent goal pursuit was low (B = −0.10, p = .145; B = −0.01, p = .904, respectively). For participants with low (−1 SD) attachment anxiety, PPR/R between baseline and follow-up was unrelated to residualized change in attachment anxiety at high or low levels of goal pursuit (B = −0.13, p = .282; B = −0.19, p = .078, respectively).
Predicting concurrent attachment anxiety
Concurrent model: Attachment anxiety and attachment avoidance at baseline as a function of support variables.
Note. In the model predicting attachment anxiety, baseline attachment avoidance is included as a predictor, and vice versa.
Predicting concurrent attachment avoidance and change in attachment avoidance
In the exploratory model predicting concurrent attachment avoidance, we observed main effects of both PPR/R and perceived goal support at baseline, indicating that people who perceived greater partner reassurance/responsiveness and greater goal support reported lower concurrent attachment avoidance (see Supplemental Table 7). The interaction term did not reach significance.
In models predicting residualized change in attachment avoidance over time, neither perceived partner support for independent goal pursuit, PPR/R, nor their interaction reached significance (see Supplemental Table 10). These null results were consistent in models that included PPR/R at baseline or between baseline and follow-up.
Discussion
In the current research, we integrated two theoretical perspectives—the attachment security enhancement model (Arriaga et al., 2018) and the dependency paradox based in attachment theory (B. C. Feeney, 2007) to develop novel hypotheses about prospective and concurrent links between perceived partner support and attachment anxiety. We found support for our primary prediction that perceived partner support for independent goal pursuit predicts decreases in attachment anxiety most strongly (and, in fact, only) when partners are also perceived to provide reassurance and responsiveness.
Primary findings
The ASEM posits that independent pursuits enable people to develop a more positive model of self (i.e., self-efficacy, self-worth), which can mitigate concerns about being unworthy of love and fears about being abandoned. However, inconsistent with past work (Arriaga et al., 2014; 2021), we did not find that perceived goal support predicted declines in attachment anxiety on its own. Instead—and consistent with the dependency paradox—we found that perceived goal support predicted declines in attachment anxiety when it was perceived in a context in which dependency needs were already being met (high baseline PPR/R) or when it was balanced with co-occurring support for dependence (high PPR/R between baseline and follow-up). These findings are in line with attachment theory’s depiction of the attachment and exploration behavioral systems as interconnected (Bowlby, 1988) and are consistent with prior work showing that reassuring/responsive partners provide the catalyst for successful autonomous pursuits (B. C. Feeney, 2004; B. C. Feeney et al., 2017; B. C. Feeney & Collins, 2015; Jakubiak & Feeney, 2016).
We also found that the balance of perceived support for dependence and independence was associated with declines in attachment anxiety most strongly for people with high baseline attachment anxiety. People with high attachment anxiety certainly had the most “room” to experience declines, but they also may have benefited most from pairing independent goal support with reassurance and comfort. Without reassurance, people with high attachment anxiety may misinterpret a partner’s independent goal support as a relationship threat and avoid goal pursuit; reassurance as well as goal support is necessary for enabling people with high attachment anxiety to pursue their goals and to reap the associated personal benefits.
Secondary and exploratory findings
Regarding our secondary study aim, we did not find support for our hypothesis that high PPR/R and perceived support for independent goal pursuit interact to predict concurrent attachment anxiety. Instead, we found only a positive main effect of PPR/R. This pattern of results is consistent with past research and the ASEM’s proposition that “safe strategies” are the best approach to buffer attachment anxiety in the moment (Arriaga et al., 2014, 2018). Although we expected that a balance between dependence and independence would prevent over-reliance on one’s partner, this finding suggests that a hyperactivated strategy of over-reliance is avoided when partners are perceived to be a reliable source of comfort and support. The lack of the significant interaction should be replicated, given that we may have been underpowered to observe the predicted interaction.
Interestingly, our exploratory analyses predicting concurrent attachment avoidance revealed that both higher PPR/R and higher perceived support for independent goal pursuit independently predict lower concurrent attachment avoidance. The association between goal support (a “soft strategy” that respects desires for autonomy) and low attachment avoidance is consistent with past research and the ASEM (Arriaga et al., 2014, 2018). The novel independent association between PPR/R and lower concurrent attachment avoidance may mean that comfort and reassurance also buffer attachment avoidance in the moment. This interpretation is consistent with research demonstrating that intimate and positive relationship experiences are associated with same-day positive relationship perceptions, even among people with high attachment avoidance (Carmichael et al., 2021; Stanton et al., 2017). Our measure of PPR/R at baseline assessed specific reassuring and responsive behaviors that a partner is perceived to provide during stress (e.g., attentiveness, physical affection). When partners provide these behaviors, individuals may be more willing to turn toward others rather than remaining emotionally distant. It is also possible that people who turn toward others (those with low attachment avoidance) simply perceive others as being more attentive, responsive, and affectionate because they engage in more direct support-seeking.
Finally, the exploratory analyses assessing whether PPR/R and/or perceived support for independent goal pursuit predict declines in attachment avoidance revealed consistently null results. In this sample, attachment avoidance declined less, on average, than attachment anxiety, making it harder identify factors that contribute to its decline.
Strengths and limitations
The current research adds to a broader literature on balancing support for independence and dependence. For instance, Feeney and Collins (2015) proposed that people thrive (i.e., experience happiness, purpose in life, psychological well-being, rewarding social relationships, and physical health) when they receive 1) reassurance, comfort, and assistance during adversity as well as 2) encouragement to accept opportunities and develop individual skills in the absence of adversity. The current research expands upon these ideas and the empirical research that supports them (B. C. Feeney et al., 2017; Jakubiak & Feeney, 2016; Tomlinson et al., 2015) by demonstrating that simultaneously perceiving support for independence and dependence also predicts declines in attachment anxiety over time. This work also has important practical implications for interventions aimed at reducing attachment anxiety. Specifically, our findings highlight the need for multi-component interventions that encourage anxiously attached individuals to pursue personal goals and also ensure that needs for dependence are supported through partners’ reassurance and demonstrations of continued availability.
The current work also highlights the value in assessing attachment change during the transition to marriage. Like the transition to parenthood—a context that commonly has been considered a key opportunity for working models to be updated with new information (e.g., Arriaga et al., 2021; Rholes et al., 2020; Simpson et al., 2003)—the transition to marriage may provide opportunities for partners to navigate new challenges and stressors together and view themselves and one another differently. As partners transition to marriage, they will make decisions about whether and how to merge their financial resources, how to navigate new relationships with in-law family, and how to spend time together and apart. In each case, partners are renegotiating the balance of dependence and independence in the relationship and navigating potentially stressful discussions, so there are opportunities for updating working models.
Although this research provides an important extension of past research on change in attachment anxiety, it is nevertheless limited by a longitudinal correlational design that cannot establish causality. Future researchers may consider testing whether experimental interventions that encourage partners to balance support for dependence and independence produce subsequent decreases in attachment anxiety over time. Additionally, despite the strengths of assessing attachment change during the transition to marriage, newlywed participants are unique in that they are likely highly satisfied with their relationships and are in a relatively early stage of their relationships. Future research should address whether relationship quality or relationship stage are key determinants that provide a sensitive period for attachment. People may be most able to move toward a secure attachment when they have an overall high level of relationship satisfaction and commitment (without this foundation, any support may be interpreted less favorably), and people may be most able to change their working models earlier in relationships, before relationship patterns are firmly established. Additionally, participants in this sample were in different-sex relationships and were primarily White; future research can assess the generalizability of these findings in more diverse samples. Finally, we collected no information about participants’ gender identities (only sex), sexual orientation, or disability status; future research should collect these potentially informative participant characteristics.
Conclusion
In summary, the current research extends the ASEM model by demonstrating that support for independence is only effective to predict decreases in attachment anxiety when partners simultaneously accept dependence. Concurrently, attachment anxiety is lowest when partners offer reassurance, but attachment avoidance is lowest when partners simultaneously offer reassurance and support goals. Together, these results highlight the balancing act of effective partner support and contribute to a burgeoning literature showing links between perceived partner support and attachment insecurity.
Supplemental Material
Supplemental Material - Social support is a balancing act: Mitigating attachment anxiety by supporting independence and dependence
Supplemental Material for Social support is a balancing act: Mitigating attachment anxiety by supporting independence and dependence by Brett K Jakubiak, Julian D Fuentes, Elina R Sun and Brooke C Feeney in Journal of Social and Personal Relationships
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a grant from the National Science Foundation (BCS1650477) awarded to Dr. Brooke Feeney. Portions of this work were presented at the Society for Personality and Social Psychology annual conference in 2022.
Author’s Note
Brett K. Jakubiak is at Assistant Professor at Syracuse University; Julian D. Fuentes and Elina R. Sun are graduate students in the Social Psychology PhD program at Syracuse University; and Brooke C. Feeney is a Professor of Psychology at Carnegie Mellon University.
Supplemental Material
Supplemental material for this article is available online.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
