Volatility in daily relationship quality

Abstract

Previous research on attachment orientations has focused on how attachment is associated with levels of relationship quality; however, the nature of associations with variability over time (volatility) on relationship quality remains unclear. Couples who are higher in volatility have poorer relationship outcomes, thus it is important to understand factors that influence volatility. We used 7-day daily diaries with both members of 157 heterosexual couples to analyze associations between actor and partner reports of attachment anxiety and avoidance and gender in associations with both general levels of relationship quality and volatility in daily relationship quality. Overall, we found that regardless of gender, attachment avoidance was linked with decreased levels of relationship quality for both actors and partners. Gender differences in volatility of daily relationship quality emerged such that women’s attachment influenced both their own and their male partner’s volatility. Women’s attachment anxiety was positively associated with volatility for both their own and their partner’s relationship quality. Women’s attachment avoidance was also negatively associated with volatility in their partner’s relationship quality. We discuss how attachment avoidance is a greater predictor for average levels of daily relationship quality, whereas attachment anxiety drives volatility in daily feelings about the relationship. Further, conflict is an important factor to consider in these links between attachment anxiety and volatility in relationship quality; on days when individuals reported greater conflict than usual, they reported lower relationship quality, and this association was stronger for those whose partners were high in attachment anxiety. We explore implications for research and practice.

Keywords

APIM attachment daily diary intraindividual variability relationship quality volatility

In an effort to foster healthy relationships, much research has focused on understanding predictors of relationship quality for individuals and their partners. For example, attachment orientations are repeatedly identified as key predictors of relationship quality, such that greater anxiety and avoidance are associated with lower levels of perceived relationship quality (Mondor, McDuff, Lussier, & Wright, 2011; Shaver & Mikulincer, 2006). Although high levels of relationship quality are ideal for relationship outcomes, and low levels of perceived relationship quality are detrimental for relationship health, volatility (e.g., variability over time) in relationship perceptions may be uniquely harmful to relationship health and individual well-being (Arriaga, 2001; Arriaga, Reed, Goodfriend, & Agnew, 2006; Campbell, Simpson, Boldry, & Rubin, 2010; McDaniel, 2016). For example, daily volatility in relationship quality is associated with greater reports of distress and destructive couple interactions (Campbell et al., 2010); weekly volatility is associated with poorer relationship outcomes such as eventual breakup (Arriaga, 2001) and greater depression (Whitton & Whisman, 2010); and volatility over longer periods of time (i.e., every 4 months for 2 years) is associated with eroded commitment (Whitton, Rhoades, & Whisman, 2014).

Attachment theory is an important foundation for understanding adult romantic relationships by highlighting different patterns of behavior (Mikulincer & Shaver, 2009). Attachment behaviors are well documented for the role they play in individuals’ perceptions of general levels of relationship quality (Shaver & Mikulincer, 2006); further, the impact of attachment behaviors on the relationship varies across genders (Del Giudice, 2011). The patterns of behavior for those high in attachment anxiety and avoidance are often destructive because of overactive or deactivated responses to their partner, so they are strongly associated with perceptions of the relationship for both partners (Shaver & Mikulincer, 2006). Further, for those high in attachment anxiety, higher perceived conflict activates the attachment system, in association with lower relationship quality (Bowlby, 1980; Shaver & Mikulincer, 2006). However, examinations of how attachment are associated with daily volatility in perceptions of relationship quality are lacking.

In sum, our goal is to understand how individual characteristics, such as attachment orientations (i.e., anxiety and avoidance) and gender, predict both levels of and volatility in relationship quality. Further, given the role conflict may play in relationship quality, particularly for those high in attachment anxiety (Campbell, Simpson, Boldry, & Kashy, 2005), we examine the associations between daily conflict and relationship quality depending on attachment to help explain volatility in relationship quality. This knowledge should encourage the development of methods to address and stabilize such relationship perceptions for relational well-being. To foster this understanding, we examine associations between actor and partner reports of attachment orientations (i.e., anxiety and avoidance) and gender with both levels of and volatility in perceived daily relationship quality (satisfaction, commitment, love, closeness, and ambivalence).

Theoretical context of romantic attachment

From attachment theory and research, two insecure orientations emerge: anxiety and avoidance (Shaver & Mikulincer, 2006). High attachment anxiety is associated with fears regarding consistent care and affection, which often result in a preoccupation with whether other people are reliable sources of affection. High attachment avoidance is associated with the neglect of and distrust for relying on other people, often resulting in a learned behavior of avoiding intimate interactions (Shaver & Mikulincer, 2006). In contrast, low attachment anxiety and avoidance (i.e., attachment security) are associated with engagement with one’s partner and less extreme reactivity and sensitivity to one’s partner’s social cues (Shaver & Mikulincer, 2006). These different strategies used by individuals based on their attachment orientations have implications for both levels of and volatility in their relationship quality.

Attachment and levels of relationship quality

Individuals’ attachment orientations are predictors of long-term relationship outcomes (Simpson & Overall, 2014). Individuals who are low in either attachment anxiety or avoidance have higher reports of relationship quality, whereas those high in either attachment anxiety or avoidance report a greater number of negative feelings and fewer positive ones (Feeney, Noller, & Roberts 2000). A meta-analysis also demonstrated that high attachment anxiety or high attachment avoidance is associated with lower reports of relationship satisfaction (Li & Chan, 2012).

Further, partners are interdependent (Kelley, 1979), and one’s own attachment behavior will also impact partner responses in the relationship (Mikulincer & Shaver, 2009). By analyzing conversations between partners for 7 days, researchers found that one partner’s attachment anxiety is associated with less satisfaction, poorer relationship disclosure, and negative perceptions of interactions for both partners (Bradford, Feeney, & Campbell, 2002). In contrast, attachment avoidance by one’s partner is associated with less disclosure to that partner about one’s own feelings in the relationship, which negatively affects closeness and commitment for both partners and one’s own overall perception of the relationship (Bradford et al., 2002). Given the deleterious effects of attachment anxiety and avoidance on how partners perceive and compensate for these insecurities (Lemay & Dudley, 2011; Simpson & Overall, 2014), it is important to continue to investigate how attachment anxiety and avoidance impact both partners’ perceptions of the relationship.

Attachment and volatility in relationship quality

In much of the previous literature just described, the focus was on how attachment anxiety and avoidance were associated with levels of relationship quality, but not change in relationship quality over time. However, feelings and cognitions that prime the attachment system may fluctuate day to day. Here, we provide an important and novel extension of the attachment literature—how attachment anxiety and avoidance are also associated with volatility in daily relationship quality. Although some have found that high attachment avoidance intensifies the negative association between volatility in relationship quality and psychological distress over the course of months (Whitton et al., 2014), it is less clear how attachment avoidance can be predictive of volatility in relationship quality from day to day.

Attachment anxiety

Attachment anxiety creates a peaked state of awareness to social and emotional cues so an individual becomes more attuned to others’ responses, referred to as hypervigilance (Mikulincer & Shaver, 2009). In fact, researchers used an implicit investigation of individual reactivity to social stimuli and found that high attachment anxiety can create hypervigilance in monitoring social cues (Fraley, Niedenthal, Marks, Brumbaugh, & Vicary, 2006). Further, similar implicit methods (described as those measuring relatively unconscious and automatic evaluations) demonstrate that self-doubt, which contributes to skewed and reactive beliefs about other people’s perceptions, is readily triggered by those high in attachment anxiety (Peterson, 2014).

However, researchers investigating implicit responses to facial expressions suggest that this high anxiety also creates perceptual errors because a highly activated emotional response system results in stronger reactions to perceived social cues; therefore, these individuals are likely to quickly jump to emotional conclusions resulting in emotional instability (Fraley et al., 2006), potentially producing greater volatility in perceptions from day to day. Further, the emotional hypervigilance of those high in attachment anxiety creates such a focus on one’s own vulnerability that it reduces the mental resources available to provide sensitivity and compassion to others (Fraley et al., 2006). The reactive nature in which these individuals respond to social cues, positive or negative, subsequently impacts both their own and their partners’ feelings and perceptions (Bradford et al., 2002). In sum, this emotional hyperactivation and reactance may result in greater volatility for those high in attachment anxiety and less volatility for those low in attachment anxiety.

Attachment anxiety and conflict

Those high in attachment anxiety perceive more conflict and are likely to escalate the severity of conflict with their partners, and their higher perceived conflict is associated with lower relationship quality (Campbell et al., 2005). Further, researchers have found most of the variance in daily conflict is due to day-to-day variation within individuals (rather than between-person differences in overall conflict; Totenhagen, Butler, Curran, & Serido, 2016) suggesting that conflict often fluctuates from day to day. Given anxious individuals are sensitive to conflict, and conflict is a construct that is likely to fluctuate from day to day, conflict likely serves as an important construct in understanding the link between attachment anxiety and daily relationship quality. Thus, in addition to examining whether attachment anxiety is associated with greater volatility in relationship quality, we also examine whether anxious individuals are more sensitive to daily relational conflict by testing whether daily conflict predicts daily relationship quality as moderated by attachment anxiety.

Attachment avoidance

Individuals high in attachment avoidance are often more emotionally deactivated, have blunted social responsiveness, and do not report as many emotional extremes as compared to individuals’ high attachment anxiety (Bradford et al., 2002; Fraley et al, 2006; Mikulincer & Shaver, 2009). Indeed, attachment deactivation not only reduces the willingness to participate in support-seeking tendencies, creating distance, or extreme self-reliance but it is also characterized by low emotional reactivity (Mikulincer & Shaver, 2009). Although attachment avoidance does not prevent individuals from perceiving emotional cues, it does blunt the response to them (see Shaver & Mikulincer, 2006). Despite low reactivity, those high in attachment avoidance still experience the negative effects of volatility in relationship perceptions on personal well-being (Whitton et al., 2014). Attachment avoidance is also associated with less disclosure about one’s self and fewer expressed feelings about the relationship (Mikulincer & Shaver, 2012). In sum, the blunted and deactivated responses of those higher in attachment avoidance should result in less volatility in perceived relationship quality from day to day than those low in attachment avoidance.

Partner effects and attachment

The attachment of the partner influences relational perceptions of both partners in the relationship (Bradford et al., 2002; Campbell et al., 2005; Mikulincer & Shaver, 2009). Whereas attachment anxiety and avoidance are associated with individuals’ general levels of relationship quality, it is less clear how the attachment orientations of one’s partner might influence daily volatility in one’s own perceptions of relationship quality. Findings about how attachment insecurity impacts the strategies of the partner, however, provide insight into partner effects on volatility. The insecurity of individuals is both perceived by their partners and also impacts their partners’ behaviors in compensating for individuals’ insecurity.

Individuals who have partners high in attachment anxiety often respond with exaggerated positive feedback (Lemay & Dudley, 2011). Also, stronger physiological reactions and stress in the face of relationship conflict emerge for partners high in attachment anxiety, at least for men who have female partners high in attachment anxiety (Powers, Pietromonaco, Gunlicks, & Sayer, 2006). Further, individuals high in attachment anxiety are often emotionally reactive in their interactions (Fraley et al., 2006) and escalate relational conflict (Campbell et al., 2005), creating greater opportunities for their partners to experience heightened reactions to this conflict. Thus, partners of those high in attachment anxiety often react in accordance with their partners’ behavior and instability (Lemay & Dudley, 2011; Simpson & Overall, 2014). Therefore, we expect that individuals with partners high in attachment anxiety will report higher daily volatility in relationship quality because of the hyperactivation of these partners.

Similarly, it seems likely that conflict should play a role in understanding relationship quality perceptions for partners. Researchers have found partner effects for attachment such that the eroding nature of hypervigilance (anxious orientation) moderated associations between relational conflict and relational quality (satisfaction and closeness; Campbell et al., 2005). In addition to actor effects, we also examine whether partners of anxious individuals are more sensitive to daily relational conflict by testing whether daily conflict predicts daily relationship quality as moderated by the partner’s attachment anxiety.

Because individuals high in attachment avoidance are less outwardly reactive to relationship experiences (Mikulincer & Shaver, 2009) and disclose less to their partners (Bradford et al., 2002), their partners consistently receive deactivated emotional responses (Shaver & Mikulincer, 2006). Whereas attachment avoidance is associated with lower relationship quality, individuals with partners higher in attachment avoidance may experience less daily change in their interactions with their partners, resulting in less volatile perceptions of relationship quality.

Gender and attachment

Whereas the impact of insecure attachment is consistently detrimental for relationship quality, impacts for both partners may differ by gender (Del Giudice, 2011; Li & Chan, 2012). Meta-analysis reveals men are higher in attachment avoidance and women are higher in attachment anxiety (Del Giudice, 2011). Also, gender can act as a moderator for relationship perceptions, with attachment anxiety and avoidance in women being associated with less positive relationship beliefs and behaviors than that of their male counterparts (Campbell et al., 2010; Hadden, Smith, & Webster, 2014). Women and men also differ in behavioral and physical stress responses when high in attachment anxiety versus attachment avoidance, such that women higher in avoidance and men higher in anxiety experience greater physiological stress responses when in conflict with their partner versus women high in anxiety and men high in avoidance (Powers et al., 2006). Others, however, have found little to no evidence for the moderation of gender between attachment and relationship quality (Campbell et al., 2005; Li & Chan, 2012). The aforementioned studies are specific to attachment and levels of relationship quality and the role of gender. How gender differences might moderate the association between attachment and volatility in relationship quality is unknown and is examined in the current study.

The current study: Hypotheses and research questions

We examine how attachment anxiety and avoidance are associated with both levels and volatility of daily relationship quality using actor–partner interdependence models (APIMs; Cook & Kenny, 2005) to test both actor and partner effects. Consistent with prior literature, we expect higher attachment anxiety and avoidance to each be associated with lower levels of individuals’ daily relationship quality (actor effects; Hypothesis 1a). Further, we expect that when their partners report higher attachment anxiety or avoidance, individuals will report lower levels of daily relationship quality (partner effects; Hypothesis 1b).

For volatility in relationship quality, we expect that when individuals are higher in attachment anxiety, both they (actor effects; Hypothesis 2a) and their partners (partner effects, Hypothesis 2b) will have greater volatility in daily relationship quality. In contrast, we expect that when individuals are higher in attachment avoidance, both they (actor effects; Hypothesis 3a) and their partners (partner effects, Hypothesis 3b) will have lower volatility in daily relationship quality.

Given the mixed literature on gender differences and attachment (Del Giudice, 2011; Li & Chan, 2012), we pose the following research questions: Does gender moderate the impact of attachment anxiety and avoidance for individuals or their partners on levels (Research Question 1a) of and volatility (Research Question 1b) in daily relationship quality?

Finally, to provide context to our expectations for attachment anxiety and volatility and to test the possibility of conflict as a potential factor contributing to this volatility, we hypothesize that attachment anxiety will moderate the link between daily conflict and relationship quality, with this link being stronger for individuals (and their partners) higher (vs. lower) in attachment anxiety (H4).

Method

Procedure and participants

Couples were recruited via Family Studies and Human Development and Communication courses at a large Southwestern University in the United States. Students could earn extra credit through participation or by giving flyers to others (e.g., friends, parents, and roommates). Thus, the sample included both students and nonstudents. Although nonstudents did not receive compensation, the student who referred them received extra credit. Qualified participants were at least 18 years of age and had been in the committed relationship (e.g., dating, cohabiting, and married) for at least 6 weeks. Both participants had to have their own unique e-mail address.

We collected data via a secure Internet-based system. Participants received instructions on how to access the study and create a unique couple ID that linked the two partners and indicated the individual’s gender. Participants were instructed to complete surveys separately from their partners. Participants completed an initial questionnaire with demographic information, attachment, and other measures not reported in the present study. Upon completion of the initial survey, we asked participants to complete the daily surveys, including measures of daily relationship quality and conflict, at approximately the same time every day for 7 consecutive days. We chose 7 days because that length is relatively common in daily diary studies (e.g., King & DeLongis, 2013; Schumann, 2012; Young, Curran, & Totenhagen, 2013), includes a sample of both weekdays and weekends, and increases the likelihood that participants would remain in the study rather than become fatigued and drop out if asked to complete a longer period (e.g., Bolger, Davis, & Rafaeli, 2003). For these daily surveys, we instructed participants to think about each item as they had experienced it within the past 24 hours.

An advantage of the online nature of the surveys was the ability to use time and date stamps of each submission to clean the data (Ogolsky, Niehuis, & Ridley, 2009). We removed any entries that were deemed invalid due to duplicate entries (e.g., submitting more than one survey in a single day). We also used time and date stamps to identify daily entries in which both partners did not match (i.e., both did not submit an entry on a particularly day) given the dyadic nature of the study and our focus on examining both actor and partner effects.

In total, 313 couples originally agreed to participate in the study. After excluding those whose data indicated they did not meet eligibility criteria (i.e., only one partner of the couple completed surveys) and those who only completed the initial baseline survey or who did not complete at least three matching days of data with their partner (given our focus on actor and partner daily processes; see Totenhagen, Serido, Curran, & Butler, 2012), as well as same-sex couples (very few participated, preventing analysis of this sample given limited power), our final sample for the present study was 157 heterosexual couples (N = 314 individuals).

Participants were primarily Caucasian (68.2%) and Hispanic (13.4%), with 4.5% of participants reporting Asian or Pacific Islander, 3.8% African American, 1.3% American Indian or Alaska Native, and 7.0% mixed race/ethnicity. Of the participants, 61.8% were college students. The majority had completed some college or an Associate’s degree (58.6%) with participant ages ranging from 18 to 66 years (M = 25.87, SD = 10.57). Approximately 26.1% of participants were married and 21.9% of the unmarried couples were cohabiting. Lengths of relationships ranged from 2 months to 44 years (M = 5.93 years, SD = 9.07 years), and 21.0% had at least one child (see Table 1 for more information).

Table 1.

Participant demographic characteristics.

	Men	Women
Mean age (SD)	26.57 (11.04)	25.16 (10.70)
Ethnicity
Caucasian or White	104 (66.2%)	110 (70.1%)
American Indian or Alaska Native	2 (1.3%)	2 (1.3%)
Hispanic or Latin American origin	23 (14.6%)	19 (12.1%)
African American or Black	7 (4.5%)	5 (3.2%)
Asian or Pacific Islander	8 (5.1%)	6 (3.8%)
Mixed race/ethnicity	11 (7.0%)	11 (7.0%)
Education level
Less than high school	1 (0.6%)	0 (0.0%)
Some high school	0 (0.0%)	2 (1.3%)
Graduated high school	26 (16.6%)	20 (12.7%)
Some college or associate’s degree	85 (54.1%)	99 (63.1%)
Bachelor’s degree	31 (19.7%)	28 (17.8%)
Some graduate school	3 (1.9%)	1 (0.6%)
Master’s degree	7 (4.5%)	4 (2.5%)
Degree such as MD or PhD	4 (2.5%)	2 (1.3%)
Marital status (% married)	41 (26.1%)	41 (26.1%)

Note. The mean and standard deviation are included for age. Otherwise, we provide the number for each category with percent of sample.

Measures

Attachment orientation

We measured attachment anxiety and avoidance using the Experiences in Close Relationships—Short Form (Wei, Russell, Mallinckrodt, & Vogel, 2007), which participants completed as part of their initial survey. The scale is a 12-item assessment of relationship security, which is measured by the presence of attachment anxiety or avoidance (e.g., “I get uncomfortable when a romantic partner wants to be very close,” “I find it difficult to allow myself to depend on romantic partners”), with responses from disagree strongly (1) and agree strongly (7). Previous use of this scale has demonstrated reliability and validity (e.g., Wei et al., 2007). Our αs were .70 for anxiety and .77 for avoidance.

Daily relationship quality

Each day, participants rated five relational qualities on a 1 (not very much or just a little) to 7 (very much or a lot) point scale. We asked participants how they felt today with respect the following: satisfaction, commitment, love, closeness, and ambivalence (e.g., “Today, how satisfied were you with your relationship with your partner?”; “Today, how ambivalent, or uncertain did you feel about the future of your relationship with your partner?”). Ambivalence was reverse coded, and these items were averaged to create the composite of relationship quality (α = .88). To reduce negative skewness in the final scale scores, we reflected the variable and performed a square root transformation. We then rereflected the transformed variable for the ease of interpretation (Tabachnick & Fidell, 2001) so that higher scores indicate greater relationship quality.

Daily conflict

Each day, participants were asked “Today, how much conflict did you have with your partner?” (1 = not very much or just a little; 7 = very much or a lot).

Control variables

We included relationship length in months (grand mean centered),¹ parent status (have children = 1), and marital status (married = 1) as controls. These constructs are common and important controls in studies of romantic relationships (e.g., Del Giudice, 2011; Hadden et al., 2014; Ruppel & Curran, 2012; Young, Curran, & Totenhagen, 2013). For example, relationship length explains interdependence between individuals in romantic relationships as well as if individuals should move forward with their relationship (Braiker & Kelley, 1979). Further, marriage often provides protective benefits compared to other relationship statuses (e.g., Stanley, Rhoades & Markman, 2006). Parent status is also important, as having children may negatively impact romantic relationships between parents through the deterioration of relationship quality (Doss, Rhoades, Stanley, & Markman, 2009) and reduced relationship satisfaction associated with child rearing (Twenge, Campbell, & Foster, 2003).

Data analysis

In Table 2, we include descriptive information and correlations for noncategorical study variables. All variables of interest were significantly correlated. Relationship length (used as a control), however, was only significantly related to attachment avoidance for the individual (i.e., not partner attachment avoidance).

Table 2.

Correlations and descriptive statistics.

	Correlations							Means (SDs)
	1	2	3	4	5	6	7	Women	Men
1. Relationship length in years	.998**	−.18*	.04	−.12	−.01	−.12	.04	5.83 (8.86)	6.02 (9.29)
2. Actor attachment avoidance	.08	.30**	.30**	.47**	.26**	.37**	−.58**	2.19 (1.10)	2.52 (1.05)
3. Partner attachment avoidance	−.11	.30**	.30**	.17*	.35**	.36**	−.41**	2.52 (1.05)	2.19 (1.10)
4. Actor attachment anxiety	.01	.35**	.26**	.18*	.18*	.34**	−.38**	3.48 (1.29)	3.42 (1.23)
5. Partner attachment anxiety	−.09	.17*	.47**	.18*	.18*	.17*	−.21*	3.42 (1.23)	3.48 (1.29)
6. Daily conflict^a	.05	.47**	.18*	.24**	.20*	.66**	−.64**	2.01 (1.03)	2.14 (1.10)
7. Relationship quality^a	−.14	−.64**	−.37**	−.25**	−.17*	−.52**	.63**	5.95 (1.01)	5.86 (0.97)

Note. Correlations for women are reported above the diagonal. Correlations for men are reported below the diagonal. Correlations between women and men are reported in italics on the diagonal. SD = standard deviation.

^aCorrelations and descriptive statistics of daily variables are provided for individuals’ mean scores of these daily variables.

*p < .05; ** p < .01.

We implemented APIMs using multilevel modeling (MLM) in SAS PROC MIXED (version 9.4; SAS Institute, 2011) to examine our hypotheses and research questions, as these models are the most appropriate for our nested and interdependent data (Singer & Willet, 2003). Multilevel models have many advantages, including allowing for longitudinal data and missingness on the outcome variable, while also taking into account that the variation that is observed in the outcome variable comes from multiple sources (e.g., male and female partners within couples and individuals across days; Kenny, Kashy, & Cook, 2006).

The data have three levels: days (Level 1) nested within participants (Level 2) who are further nested within dyads or couples (Level 3). Dyadic data is unique in that Level 2 has no variability once the role of the partner is taken into account; in other words, there are only two partners in each dyad (Bolger & Laurenceau, 2013). To alleviate this variability problem, we used dummy-coded variables for men and women (see Bolger & Laurenceau, 2013). We entered these dummy-coded variables on the random line in SAS PROC MIXED, which allows for interindividual differences in men’s and women’s intercepts on relationship quality. The model becomes a two-level model: day (Level 1) nested within men and women nested within couples (Level 2). In such models, within-person predictors are entered at Level 1, and between-person and between-dyad predictors are entered at Level 2.

Similar to Campbell, Simpson, Boldry, and Kashy (2005), we utilized a one-intercept model versus the two-intercept model in Bolger and Laurenceau (2013). This required us to enter an indicator of gender (which was effect coded, women = −1, men = 1). That is, we accounted for the lack of variability at Level 2 in the random effects by utilizing the dummy codes for men and women, but then rather than using the dummy codes in the fixed effects, we utilized the indicator of gender. The results were similar and more parsimonious this way. Instead of obtaining separate estimates for all fixed effects for men and women and having to test the difference between each of these, we now obtain a single estimate for each fixed effect pooled across gender as well as moderation by gender that inherently tests whether effects for men and women are significantly different from one another.

To test our first three hypotheses and research questions, we used methods by Hoffman (2007) to test for predictors of within-person residual variance. This model is an extension of the multilevel model for heterogeneous variances. This model allows for heterogeneity in the within-person (Level 1) residual variability, and we are then able to predict interindividual differences in this residual variability, whereas in a typical multilevel model, the variance of the Level 1 residual errors is constrained and not allowed to vary across participants. We included constructs of interest as both fixed effects (e.g., attachment as associated with levels of relationship quality) and as predictors of the within-person residual variance (e.g., attachment as associated with within-person variability from day to day in relationship quality). In doing so, one can model whether the remaining within-person residual variance in day-to-day scores of relationship quality, after accounting for that predicted by the fixed effects, can be predicted by constructs of interest (here, gender and attachment). To the extent that constructs of interest are significantly associated with this remaining within-person residual variance, these constructs are thus predicting volatility in perceived relationship quality. In SAS PROC MIXED, the repeated statement is what is used to permit the inclusion of between-person predictors of within-person residual variance (see Hoffman, 2007; Totenhagen et al., 2016).

For our fixed effects (attachment as associated with levels of relationship quality), we entered controls, attachment anxiety and attachment avoidance (both actor and partner), gender, and interactions with gender. In examining overall volatility in daily relationship quality, we also entered attachment anxiety and avoidance (both actor and partner), gender, and interactions with gender on the repeated line in PROC MIXED as recommended by Hoffman (2007). This allows these between-person variables to predict between-person differences in the heterogeneity in daily relationship quality (i.e., residual variability). Predictors of the residual variance component are exponentiated to allow for linear prediction of the variance. Hoffman (2007) explains that the “exponential function was used to normalize the variance so that a linear prediction model may be used as well as to eliminate the dependence of the variance on the mean” (p. 620). In the end, we trimmed all nonsignificant interactions with gender. We describe the general equations, before any trimming, for our MLM below:

Level 1:

Relationship {quality}_{t i} = β_{0 i} + β_{1 i} {day}_{t i} + e_{t i}

Level 2:

\begin{array}{l} β_{0 i} = γ_{00} + γ_{01} {gender}_{i} + γ_{02} {relationship length}_{i} + γ_{03} {parent status}_{i} + γ_{04} {marital status}_{i} + γ_{05} {anxiety}_{i} + γ_{06} {partner anxiety}_{i} + γ_{07} {avoidance}_{i} + γ_{08} {partner avoidance}_{i} + γ_{09} anxiety \times {gender}_{i} + γ_{010} partner anxiety \times {gender}_{i} + γ_{011} avoidance \times {gender}_{i} + γ_{012} partner avoidance \times {gender}_{i} + males \times μ_{0 i} + females \times μ_{0 i} \end{array}

β_{1 i} = γ_{10}

Heterogeneous variance in the Level 1 residuals

σ_{e i}^{2} = α_{0} exp (\begin{array}{l} α_{1} {gender}_{i} + α_{2} {anxiety}_{i} + α_{3} {partner anxiety}_{i} + α_{4} {avoidance}_{i} + α_{5} {partner avoidance}_{i} + α_{6} anxiety \times {gender}_{i} + α_{7} partner anxiety \times {gender}_{i} + α_{8} avoidance \times {gender}_{i} + α_{9} partner avoidance \times {gender}_{i} \end{array})

At Level 1, we have the equation describing the within-person relationship of daily relationship quality (relationship quality _ti ) to the predictors. The predicted value for relationship quality for each individual i on a given occasion t is a function of the individual’s average relationship quality on day 1 (intercept, β_0i), the linear slope of day (β_1i), and residual variation in relationship quality (e_ti).

Level 2 is of more interest. At Level 2, we entered the between-person predictors: attachment anxiety and avoidance, control variables, and between-person random effects. The average relationship quality score (β_0i) is a function of the overall sample average relationship quality score (γ₀₀), gender (γ₀₁), controls (γ₀₂ to γ₀₄), attachment anxiety and avoidance (γ₀₅ and γ₀₇), partner anxiety and avoidance (γ₀₆ and ), the interactions of anxiety and avoidance with gender (γ₀₉ and γ₀₁₁), the interactions of partner anxiety and avoidance with gender (γ₀₁₀ and γ₀₁₂), and random variation around the sample average for males and for females (males × μ_0i and females × μ_0i). The linear slope in relationship quality over days (β_1i) is the average sample linear slope across days (γ₁₀). We did not include a random effect at Level 2 for the linear slope (β_1i) as the model could not converge with this effect included and therefore had to be simplified.

The overall residual variability across days (Level 1 residual, within-person variability not accounted for by any of the fixed effects) for each individual “i” $(σ_{e i}^{2})$ , what we term volatility, was modeled by Level-2 variables, including gender (α₁), actor attachment anxiety and avoidance (α₂ and α₄), partner anxiety and avoidance (α₃ and α₅), and interactions with gender (α₆ to α₉).

Finally, to test our fourth hypothesis (anxiety as a moderator for daily conflict and daily relationship quality), we performed an additional model where we added the conditional main effect of daily conflict (person-centered so that interpretations are above or below a person’s own average) as well as the interactions of attachment anxiety by daily conflict and partner attachment anxiety by daily conflict to the fixed effects in the model described earlier. We also added the interactions of attachment avoidance by daily conflict and partner attachment avoidance by daily conflict to the fixed effects to explore whether it was the case that individuals higher in attachment anxiety (vs. attachment avoidance) specifically are more sensitive to states of daily conflict.

Results

We first tested whether gender significantly interacted with the actor and partner attachment variables and when appropriate, removed nonsignificant interactions. Several interactions between gender and attachment were significant in the exponentiated portion of the model (i.e., examination of volatility), although the same was not true for the fixed effects. Thus, we compared this full model against a nested model in which interactions of gender with actor and partner anxiety and avoidance were removed from the fixed effects. We used nested model comparisons to examine whether the full model provided significantly better fit than the more constrained model without the four interaction terms. We compared the difference in the −2 log likelihood (−2LL) for each model (Singer & Willet, 2003).

At four degrees of freedom, the difference was not significant (difference in −2LL = 3.9, p = .42). Thus, the full model did not provide a significantly better fit than the restricted model and so we removed gender interactions in the fixed effects. We retained the more parsimonious model that only contained interactions with gender in the exponentiated portion of the model (i.e., repeated line in PROC MIXED; see Table 3 for full model results). In the following paragraphs, we summarize results for each hypothesis.

Table 3.

Hypotheses 1 to 3 and Research Question 1 multilevel model (unstandardized results).

Fixed effects	Relationship quality
Intercept (γ00)	2.13***
Time (γ10)	0.001
Gender (γ01)	−0.01
Control variables
Relationship length (γ02)	−0.003
Parent status (γ03)	−0.03
Marital status (γ04)	0.003
Hypotheses 1a and 1b and Research Question 1: Actor and partner attachment anxiety and avoidance as associated with levels of relationship quality
Attachment anxiety (γ05)	−0.02
Partner attachment anxiety (γ06)	0.01
Attachment avoidance (γ07)	−0.16***
Partner attachment avoidance (γ08)	−0.07***
Gender × Attachment Anxiety (γ009)	–
Gender × Partner Attachment Anxiety (γ010)	–
Gender × Attachment Avoidance (γ011)	–
Gender × Partner Attachment Avoidance (γ012)	–
Hypotheses 2a, 2b, 3a, and 3b and Research Questions 1a and 1b: Actor and partner attachment anxiety and avoidance and interactions with gender as predictors of residual error variance (i.e., overall daily volatility)
Gender (α₁)	−0.06
Attachment anxiety (α₂)	0.18**
Partner attachment anxiety (α₃)	0.15***
Attachment avoidance, (α₄)	0.07
Partner attachment avoidance (α₅)	−0.09
Gender × Attachment Anxiety (α₆)	−0.14***
Gender × Partner Attachment Anxiety (α₇)	0.16***
Gender × Attachment Avoidance (α₈)	0.06
Gender × Partner Attachment Avoidance (α₉)	−0.22***

Note. Gender is coded −1 = female and 1 = male. Nonsignificant interactions that were trimmed after model comparisons are marked with a “–.” Control variables were coded as follows: Gender (1 = male, −1 = female), parent status (1 = yes, has children, 0 = no children), and married (1 = married, 0 = not married). We also tested the model without the control variables, found the same pattern of results, and chose to retain the controls. Simple slopes of decomposed interactions (i.e., effects for men and women, where they significantly differed) are reported in the text.

***p < .001; **p < .01; *p < .05.

Levels of relationship quality (Hypotheses 1a and 1b and Research Question 1a)

As expected, the actor (Hypothesis 1a) and partner (Hypothesis 1b) effects for attachment avoidance were both significant (b = −.16, p < .001 and b = −.07, p < .001, respectively). When individuals were higher in attachment avoidance, both their own (actor effect) and their partner’s (partner effect) reported lower levels of relationship quality. Contrary to expectations, attachment anxiety did not have a significant influence on levels of relationship quality for either actor or partner effects. Thus, Hypotheses 1a and 1b were partially supported with results as expected for attachment avoidance but not attachment anxiety. We found no significant gender differences (Research Question 1a) for the fixed effects of actor and partner attachment anxiety and avoidance.

Volatility in relationship quality (Hypotheses 2a, 2b, 3a, and 3b and Research Question 1b)

Both actor (Hypothesis 2a) and partner (Hypothesis 2b) attachment anxiety were associated with increased volatility in relationship quality, but both effects were qualified by significant interactions with gender (Research Question 1). Our expectation for Hypothesis 2a that individuals’ attachment anxiety would be associated with greater volatility in their own relationship quality (actor effect) was supported for women (b = .32, p < .001) but not for men (b = .04, p = .35). Further, we found in our examination of Hypothesis 2b (partner effects for volatility) that women’s attachment anxiety was associated with greater volatility in their male partners’ relationship quality reports (b = .31, p < .001), but men’s attachment anxiety did not predict volatility for their female partners (b = −.01, p = .87). In sum, the results showed that when women were higher in attachment anxiety, both they (actor effect) and their male partners (partner effect) reported higher daily volatility in relationship quality, but not vice versa.

For attachment avoidance and volatility in relationship quality, we hypothesized that individuals’ attachment avoidance would be associated with less volatility in their own relationship quality (actor effect; Hypothesis 3a), but this was not supported (b = .07, p = .17). With respect to the partner effect of attachment avoidance (Hypothesis 3b), however, we found some gender differences (Research Question 1b) indicating that the interaction of gender and partner avoidance was significant. In probing the interaction, we found that when women were more avoidant, their male partners demonstrated less volatility (b = −.31, p < .001), but not vice versa (b = .14, p = .07). Thus, Hypothesis 3b was supported for men’s volatility in perceived relationship quality, but not women’s.

Attachment anxiety as a moderator of daily conflict and relationship quality (H4)

In the model including conflict, we found that on days when individuals reported greater conflict than usual, they reported lower relationship quality (b = −0.07, p < .001). This conditional main effect was qualified by the higher level interaction of partner attachment anxiety by daily conflict (see Table 4). We decomposed this interaction at 1 standard deviation above and below the mean using methods outlined by Aiken and West (1991). The simple slopes analysis revealed that the association between daily conflict and daily relationship quality was stronger for those with a partner higher in attachment anxiety (b = −0.09, p < .001) compared to those with a partner lower in attachment anxiety (b = −0.06, p < .001). The interaction of actor attachment anxiety by daily conflict was not significant, nor were the interactions between actor or partner attachment avoidance and daily conflict. Thus, H4 was supported for partner attachment anxiety, but not actor attachment anxiety.

Table 4.

Does the influence of daily conflict on daily relationship quality depend on attachment anxiety (H4)? Multilevel model (unstandardized results).

Fixed effects	Relationship Quality
Intercept	2.14***
Time	0.00
Gender	−0.01
Control variables
Relationship length	−0.00
Parent status	0.01
Marital status	−0.02
Predictors
Daily conflict	−0.07***
Attachment anxiety	−0.02
Partner attachment anxiety	0.01
Attachment avoidance	−0.16***
Partner attachment avoidance	−0.07***
Attachment Anxiety × Daily Conflict	−0.00
Partner Attachment Anxiety × Daily Conflict	−0.01**
Attachment Avoidance × Daily Conflict	0.00
Partner Attachment Avoidance × Daily Conflict	0.01

Note. Gender is coded −1 = female and 1 = male. Nonsignificant interactions with gender in the fixed effects were trimmed. Control variables were coded as follows: Gender (1 = male, −1 = female), parent status (1 = yes, has children, 0 = no children), and married (1 = married, 0 = not married).

***p < .001; **p < .01; *p < .05.

Discussion

Our findings demonstrate that, irrespective of gender, attachment avoidance was linked with decreased levels of relationship quality for both partners. Gender differences in volatility of daily relationship quality emerged whereby women’s attachment influenced both their own and their male partner’s volatility. Specifically, women’s attachment anxiety was associated with greater volatility in both their own and their partner’s relationship quality. Women’s attachment avoidance was associated with lower volatility in their partner’s relationship quality. Additionally, partner attachment anxiety moderated the links between conflict and relationship quality, such that on days when individuals reported greater conflict than usual, they reported lower relationship quality, with this association being stronger for those whose partners were high in attachment anxiety.

Centrality of avoidance for levels of relationship quality

Our findings for individuals’ attachment in association with their own reported levels of relationship quality were as expected for avoidance such that greater avoidance was associated with lower relationship quality (Bradford et al., 2002; Campbell et al., 2005; Li & Chan, 2012), but this was not the case for attachment anxiety. Although previous researchers have found actor and partner effects for attachment anxiety and levels of relationship quality (Campbell et al., 2005; Mondor et al., 2011), our findings did not support this pattern. In an effort to replicate these previous findings, we investigated whether the association between attachment anxiety and levels of relationship quality would be significant when removing examinations of volatility, but our results did not change. Thus, future studies should continue to examine links between attachment anxiety and both levels of and volatility in relationship quality to determine whether these results remain consistent in other samples.

Still, the patterns in our findings underscore the centrality of avoidance for one’s perception of the levels of relationship quality. We also found evidence of partner effects for attachment avoidance. Specifically, when individuals were higher in avoidance, their partners reported lower levels of perceived relationship quality. Due to the disengaged behavior of someone high in attachment avoidance (Mikulincer & Shaver, 2012), partners may be more susceptible to negative feelings about the relationship. Attachment avoidance is a predictor of dissatisfaction in distressed couples (Mondor et al., 2011), and individuals high in attachment avoidance demonstrate less intimacy and closeness, with fewer support seeking or support giving efforts toward their partners (Mikulincer & Shaver, 2009). When individuals high in attachment avoidance fail to participate in the enhancement of their relationships, their partners may perceive this omission negatively. In fact, previous researchers have discussed the destructive nature of attachment avoidance due to the lack of attachment activation (Mondor et al., 2011).

In line with previous research (Bradford et al., 2002; Campbell et al., 2005), we expected to find that attachment anxiety also impacted actor and partner reports of relationship quality, but we only found significant effects for attachment avoidance. Although attachment anxiety is generally detrimental for relationship quality, some have proposed that avoidance is a stronger predictor for dissatisfaction in distressed couples than anxiety (Mondor et al., 2011). Our results suggest that attachment avoidance plays a unique role in predicting low levels of relationship quality. Of note, none of the associations between attachment and levels of relationship quality differed by gender despite previous literature documenting the existence of some gender differences (Del Giudice, 2011).

Attachment orientations and volatility in relationship quality

Whereas attachment avoidance drove the results for the levels of relationship quality, it was primarily attachment anxiety that was predictive of volatility in daily perceptions of relationship quality. Specifically, of eight predictors involving attachment and volatility, five were significant and four of these were specific to anxiety.

Attachment anxiety

In contrast to the lack of actor and partner findings for anxiety for levels of relationship quality, we found actor and partner effects for women’s attachment anxiety for volatility in relationship quality. When women were higher in attachment anxiety, both they and their partners had greater volatility in daily reports of relationship quality. The nature of attachment anxiety is that it manifests as an over activated awareness of social cues from their partners, leaving individuals high in attachment anxiety vulnerable to perceiving social responses as more extreme, thus experiencing greater intensity of feelings from their partners (Campbell et al., 2005).

It is intriguing that we found attachment anxiety to be more predictive of volatility in daily perceptions of relationship quality than avoidance. These results may be explained by our findings regarding the role of daily conflict (H4). We found that on days when individuals reported greater conflict than usual, they reported lower relationship quality, with this association stronger for those whose partners were high in attachment anxiety. Thus, individuals with a highly anxious partner may be more susceptible to declining relationship quality on days when they report high conflict. We did not find that the association was stronger for those who were themselves higher in attachment anxiety, nor were the interactions of daily conflict with actor or partner avoidance significant. It is plausible that partners of those high in anxiety have heightened reactions to conflict given that their anxious partners are more likely to be emotionally reactive (Fraley et al., 2006) and escalate conflict (Campbell et al., 2005), explaining why we see these effects for the partners but not the anxious individuals themselves. In sum, these results suggest that daily conflict may help explain why those with partners who are high in attachment anxiety, as opposed to attachment avoidance, are vulnerable to volatility in relationship quality. Researchers should continue to examine the role of attachment anxiety as a moderator between daily conflict and relationship quality.

Although attachment anxiety was associated with greater volatility in relationship perceptions for both actors and partners as expected, testing gender as a moderator demonstrated that these effects were specific to women’s attachment anxiety but not men’s attachment anxiety. When women were higher in attachment anxiety, both they and their male partners reported higher daily volatility in relationship quality. Previous findings document that women are higher in attachment anxiety than men (Del Giudice, 2011), and this attachment anxiety in women may be associated with more destructive interactions with partners than for anxiety in men (Li & Chan, 2012). Subsequently, when these women (vs. men) act out on their attachment anxiety through hypervigilance, our results suggest that it is associated with greater reported volatility in relationship perceptions for themselves and their male partners who are likely experiencing the negative relational effects of attachment anxiety.

To further explain the aforementioned effect, emotional hypervigilance and the occurrence of intense responses may serve to significantly change how positively the women feel about their relationship from day to day. However, we speculate that female partners of men high in attachment anxiety may come to expect hypervigilance and conflict related to their partner’s attachment insecurity, explaining why we do not find partner effects of men’s anxiety on women’s volatility. That is, women may be more skilled in buffering the deleterious impact of attachment anxiety on relationship quality reports by developing behaviors that compensate for the behaviors of their anxious partner (Lemay & Dudley, 2011). This pattern could be because attachment anxiety in their partners may activate women’s desires to care for their partners, especially because women may be expected to take on the role of maintaining the relationship (Impett & Gordon, 2008). Therefore, whereas the expectation and buffering of certain anxious attachment behaviors in their male partners may result in overall lower levels of relationship quality as demonstrated in other studies, it does not predict volatility in their daily levels of relationship quality because of the compensatory behaviors of women when men are high in attachment anxiety (Lemay & Dudley, 2011; Simpson & Overall, 2014).

Attachment avoidance

We found a partner effect that differed by gender: When women were higher (vs. lower) in avoidance, their male partners demonstrated less volatility in daily reported relationship quality. With the exception of this partner effect for men’s volatility, Hypothesis 3 was largely unsupported. But, it did provide insight into our research question regarding partner and gender effects of attachment avoidance on volatility of relationship quality. Highly avoidant women’s emotional blunting (Bradford et al., 2002) may create these more stable (less volatile) perceptions of perceived relationship quality in their male partners because there are fewer emotionally dramatic moments (Mikulincer & Shaver, 2009). This finding makes theoretical sense, as attachment avoidance is associated with a lack of relationship engagement (Bradford et al., 2002). Thus, individuals higher in avoidance may enact fewer engaging behaviors, which would influence a partner’s feelings from day to day. Although these partners may be more stable in their daily feelings of relationship quality, the results from our fixed effects demonstrated that they also reported lower overall levels of relationship quality. That is, when individuals (particularly women) are higher in avoidance, their partners have more stable but also lower levels of relationship quality.

We note that our finding regarding attachment avoidance and volatility in partners’ relationship quality was specific to women’s attachment avoidance. It is unclear why similar findings were not found for men’s avoidance and their female partner’s volatility in relationship quality. Further research is needed to probe these differences in the ways in which attachment is associated with levels of relationship quality versus volatility in relationship quality, and reasons as to why women’s attachment insecurity might drive volatility in their own and their partners’ reports of relationship quality.

We did not find significant actor effects for avoidance and volatility in relationship quality. Nonsignificant findings regarding assessments of those high in attachment avoidance have previously been attributed to the stunted responsiveness of avoidant-attached individuals in reporting experiences and feelings (Bradford et al., 2002; Fraley et al, 2006). Given how robust findings were regarding attachment avoidance and fixed effects, it seems that attachment avoidance drives average levels of daily relationship quality, whereas attachment anxiety may drive volatility in daily relationship quality. Further research is needed to better assess the distinct difference in how attachment avoidance impacts daily volatility in relationship feelings compared to attachment anxiety.

Limitations, strengths, and future directions

Our results should be interpreted in light of some limitations. The majority of participants were students, and others had connections to university students (e.g., friends and parents). It is possible that there are differences between young dating relationships and longer term marriage relationships. For example, our attachment measure may capture degree of overall attachment in those who were in shorter relationships (the majority of our sample; 74% of our participants were dating, with nearly 50% of all participants in relationships 2 years or less) versus those who have been with the same partner for many years and are reflecting primarily on attachment to their current long-term partner. In fact, attachment anxiety and avoidance are more negatively associated with satisfaction and commitment as relationship length increases (Hadden et al., 2014). Further, longer relationships are more bound by rules regarding behavior than shorter relationships (Bradford et al., 2002), which could potentially result in greater stability. We did control for relationship length, but given that the majority of our sample were dating couples in shorter relationships, it would be important for future researchers to examine these changes as they unfold over the course of relationships.

While it is important to recognize the limitations of generalizability to the greater public, strengths of generalizability can be found in the variety of relationships represented. Additionally, although it was not a criteria to identify as heterosexual to participate in the study, responses by individuals who identified as lesbian, gay, bisexual, transgender, or queer (LGBTQ) were infrequent and not included in the patterns reported here. Explorations of the study questions with a LGBTQ sample would be an important area for future research. Further, the use of daily diary data with both members of dyads provides a rich source of information and insight into daily relationship fluctuations. Although it is common to collect daily data over the course of 1 week to include both weekdays and a weekend (e.g., King & DeLongis, 2013; Schumann, 2012; Young, Curran, & Totenhagen, 2013), it is possible that the particular week sampled was unique in some way for some participants, characterizing different levels of volatility in relationship quality than is typical for them. Further research should be done to examine daily volatility over longer lengths of time.

Implications and conclusions

Understanding how to promote greater relationship quality is important. Yet, given the negative impact of volatility on relationships (e.g., Arriaga, 2001; McDaniel, 2016), explaining contributing factors to volatility, such as attachment orientations, adds awareness for clinicians, and couples alike. Our study moves the field forward using dyadic data to examine the dynamic ways in which partners’ attachment influences both their levels of and volatility in relationship quality reports, yielding important distinctions. We demonstrate a pattern of gender effects that adds to the theoretical and empirical understanding of how attachment anxiety impacts relationship quality perceptions. Specifically, we found women’s attachment anxiety and avoidance to be particularly salient in associations with volatility in both their own and their partners’ perceptions of the relationship. Understanding how attachment anxiety and avoidance in women can impact volatility in relationship perceptions and relationship health (Fraley et al., 2006) has implications for improving relationships. For example, it may be advantageous for clinicians working with couples to initially assess attachment anxiety and avoidance, particularly for women, in helping clients reach more stability in their relationship (Mondor et al., 2011).

Our findings highlight the importance of being aware of attachment insecurities for both partners, how they manifest, and the different ways in which they impact relationship quality. Given that both actor and partner effects were significant in the present study, couples and clinicians might be particularly attentive to individuals’ attachment avoidance when trying to promote higher levels of relationship quality. On the other hand, recognizing and attending to the signs of attachment anxiety (e.g., hypervigilance) might help to promote more stable levels of relationship quality, especially for women and their partners. Our current results also suggest that these two things done in conjunction would produce the most ideal outcomes, likely creating more stable and high functioning relationships simultaneously. Knowledge that attachment insecurity is an important individual factor in promoting high-quality and stable relationships promotes universal assessment for clinicians and mutual awareness of attachment insecurities in romantic partners.

Footnotes

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Note

References

Aiken

L. S.

West

S. G.

(1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.

Arriaga

X. B.

(2001). The ups and downs of dating: Fluctuations in satisfaction in newly formed romantic relationships. Journal of Personality and Social Psychology, 80, 754–765. doi:10.1037/0022-3514.80.5.754

Arriaga

X. B.

Reed

J. T.

Goodfriend

Agnew

C. R.

(2006). Relationship perceptions and persistence: Do fluctuations in perceived partner commitment undermine dating relationships? Journal of Personality and Social Psychology, 91, 1045. doi:10.1037/0022-3514.91.6.1045

Bolger

Davis

Rafaeli

(2003). Diary methods: Capturing life as it is lived. Annual Review of Psychology, 54, 579–616. doi:10.1146/annurev.psych.54.101601.145030

Bolger

Laurenceau

J. P.

(2013). Intensive longitudinal methods: An introduction to diary and experience sampling research. New York, NY: Guilford.

Bowlby

(1980). Attachment and loss: Vol. 3. Loss. New York, NY: Basic Books.

Bradford

S. A.

Feeney

J. A.

Campbell

(2002). Links between attachment orientations and dispositional and diary-based measures of disclosure in dating couples: A study of actor and partner effects. Personal Relationships, 9, 491–506. doi:10.1111/1475-6811.00031

Braiker

H. B.

Kelley

H. H.

(1979). Conflict in the development of close relationships. In Burgess

Huston

(Eds.), Social exchange in developing relationships (pp. 135–167). New York, NY: Academic.

Campbell

Simpson

J. A.

Boldry

Kashy

D. A.

(2005). Perceptions of conflict and support in romantic relationships: The role of attachment anxiety. Journal of Personality and Social Psychology, 88, 510–531. doi:10.1037/0022-3514.88.3.510

10.

Campbell

Simpson

J. A.

Boldry

J. G.

Rubin

(2010). Trust, variability in relationship evaluations, and relationship processes. Journal of Personality and Social Psychology, 99, 14. doi:10.1037/a0019714

11.

Cook

W. L.

Kenny

D. A.

(2005). The actor-partner interdependence model: A model of bidirectional effects in developmental studies. International Journal of Behavioral Development, 29, 101–109. doi:10.1080/01650250444000405

12.

Del Giudice

(2011). Sex differences in romantic attachment: A meta-analysis. Personality and Social Psychology Bulletin, 37, 193–214. doi:10.1177/0146167210392789

13.

Doss

B. D.

Rhoades

G. K.

Stanley

S. M.

Markman

H. J.

(2009). The effect of the transition to parenthood on relationship quality: An 8-year prospective study. Journal of Personality and Social Psychology, 96, 601–619. doi:10.1037/a0013969

14.

Feeney

J. A.

Noller

Roberts

(2000). Attachment and close relationships. In Hendrick

Hendrick

(Eds.), Close relationships: A sourcebook (pp. 185–201). Thousand Oaks, CA: Sage.

15.

Fraley

Niedenthal

P. M.

Marks

Brumbaugh

Vicary

(2006). Adult attachment and the perception of emotional expressions: Probing the hyperactivating strategies underlying anxious attachment. Journal of Personality, 74(4), 1163–1190. doi: 10.1111/j.1467-6494.2006.00406.x

16.

Hadden

B. W.

Smith

C. V.

Webster

G. D.

(2014). Relationship duration moderates associations between attachment and relationship quality meta-analytic support for the temporal adult romantic attachment model. Personality and Social Psychology Review, 18(1), 42–58. doi: 10.1177/1088868313501885

17.

Hoffman

(2007). Multilevel models for examining individual differences in within-person variation and covariation over time. Multivariate Behavioral Research, 42(4), 609–629. doi: 10.1080/00273170701710072

18.

Kelley

(1979). Personal relationships: Their structure and processes. New York, NY: Wiley.

19.

Kenny

D. A.

Kashy

D. A.

Cook

W. L

. (2006). Dyadic data analysis. Guilford Press.

20.

King

D. B.

DeLongis

(2013). Dyadic coping with stepfamily conflict demand and withdraw responses between husbands and wives. Journal of Social and Personal Relationships, 30(2), 198–206. doi: 10.1177/0265407512454524

21.

Impett

E. A.

Gordon

(2008). For the good of others: Toward a positive psychology of sacrifice. In Lopez

S. J.

(Ed.), Positive psychology: Exploring the best in people (pp. 79–100). Westport, CT: Greenwood.

22.

Lemay

E. P.

Jr. Dudley

K. L.

(2011). Caution: Fragile! Regulating the interpersonal security of chronically insecure partners. Journal of Personality and Social Psychology, 100, 681. doi:10.1037/a0021655

23.

Chan

D. K. S.

(2012). How anxious and avoidant attachment affect romantic relationship quality differently: A meta-analytic review. European Journal of Social Psychology, 42, 406–419. doi:10.1002/ejsp.1842

24.

McDaniel

B. T.

(2016). Understanding stability and change in daily coparenting: Predictors and outcomes in families with young children (Doctoral dissertation), The Pennsylvania State University, State College, PA.

25.

Mikulincer

Shaver

P. R.

(2009). An attachment and behavioral systems perspective on social support. Journal of Social and Personal Relationships, 26, 7–19. doi:10.1177/0265407509105518

26.

Mikulincer

Shaver

P. R.

(2012). Attachment theory expanded: A behavioral systems approach to personality. In Kay

Deaux

Mark

Snyder

(Eds.), Oxford handbook of personality and social psychology (Vol. 1134, pp. 1119–1137). Oxford University Press.

27.

Mondor

McDuff

Lussier

Wright

(2011). Couples in therapy: Actor-partner analysis of the relationships between adult romantic attachment and marital satisfaction. The American Journal of Family Therapy, 38, 112–123. doi:10.1080/01926187.2010.530163

28.

Ogolsky

Niehuis

Ridley

(2009). Using online methods and designs to conduct research on personal relationships. Marriage & Family Review, 45, 610–628. doi:10.1080/01494920903224202

29.

Peterson

J. L.

(2014). Explicit thoughts of security activate implicit self-doubt in anxiously attached participants. Personal Relationships, 21, 206–224. doi:10.1111/pere.12027

30.

Powers

S. I.

Pietromonaco

P. R.

Gunlicks

Sayer

(2006). Dating couples’ attachment styles and patterns of cortisol reactivity and recovery in response to a relationship conflict. Journal of Personality and Social Psychology, 90, 613. doi:10.1037/0022-3514.90.4.613

31.

Ruppel

Curran

M. A.

(2012). Relational sacrifices in romantic relationships: Satisfaction and the moderating role of attachment. Journal of Social and Personal Relationships, 29, 508–529.

32.

SAS Institute. (2011). SAS/STAT user’s guide. Cary, NC: SAS Institute, Inc.

33.

Schumann

(2012). Does love mean never having to say you’re sorry? Associations between relationship satisfaction, perceived apology sincerity, and forgiveness. Journal of Social and Personal Relationships, 29, 997–1010. doi:10.1177/0265407512448277

34.

Shaver

P. R.

Mikulincer

(2006). Attachment theory, individual psychodynamics, and relationship functioning. In Vangelisti

A. L.

Perlman

(Eds.), The Cambridge handbook of personal relationships (pp. 251–271). New York, NY: Cambridge University Press.

35.

Simpson

J. A.

Overall

N. C.

(2014). Partner buffering of attachment insecurity. Current directions in psychological science, 23, 54–59. doi:10.1177/0963721413510933

36.

Singer

J. D.

Willet

J. B.

(2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York: Oxford University Press.

37.

Stanley

S. M.

Rhoades

G. K.

Markman

H. J.

(2006). Sliding versus deciding: Inertia and the premarital cohabitation effect. Family Relations, 55, 499–509. doi:10.1111/j.1741-3729.2006.00418.x

38.

Tabachnick

B. G.

Fidell

L. S.

(2001). Using multivariate statistics (4th ed.). Boston: Allyn & Bacon.

39.

Totenhagen

C. J.

Serido

Curran

M. A.

Butler

E. A.

(2012). Daily hassles and uplifts: A diary study on understanding relationship quality. Journal of Family Psychology, 26, 719–728. doi: 10.1037/a0029628

40.

Totenhagen

C. J.

Butler

E. A.

Curran

M. A.

Serido

(2016). The calm after the storm: Relationship length as associated with couples’ daily variability. Journal of Social and Personal Relationships, 33, 768–791. doi:10.1177/0265407515597562

41.

Twenge

J. M.

Campbell

W. K.

Foster

(2003). Parenthood and marital satisfaction: A meta-analytic review. Journal of Marriage and Family, 65, 574–583. doi:10.1111/j.1741-3737.2003.00574.x

42.

Wei

Russell

D. W.

Mallinckrodt

Vogel

D. L.

(2007). The experiences in close relationship scale (ECR)-short form: Reliability, validity, and factor structure. Journal of Personality Assessment, 88, 187–204. doi:10.1080/00223890701268041

43.

Whitton

S. W.

Rhoades

G. K.

Whisman

M. A.

(2014). Fluctuation in relationship quality over time and individual well-being main, mediated, and moderated effects. Personality and Social Psychology Bulletin, 40, 858–871. doi:10.1177/0146167214528988

44.

Whitton

S. W.

Whisman

M. A.

(2010). Relationship satisfaction instability and depression. Journal of Family Psychology, 24, 791. doi:10.1037/a0021734

45.

Young

Curran

Totenhagen

(2013). A daily diary study: Working to change the relationship and relational uncertainty in understanding positive relationship quality. Journal of Social and Personal Relationships, 30, 132–148. doi:10.1177/0265407512453826

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.

0.00 MB