Attachment styles, negotiation of goal conflict, and perceived partner support during COVID-19

Abstract

Due to the pandemic, people have been stuck indoors with their partners for months. Instead of being able to rely on multiple sources of support, many couples have to rely on each other more. We investigated whether goal conflict, successful negotiation of the conflict, and individual differences in attachment styles were associated with perceived partner support to understand factors that may enable or hinder goal pursuit during the pandemic. Participants (n = 200) completed a daily diary for a week and weekly longitudinal reports for five weeks. Results showed that higher goal conflict predicted perception of less relational catalyst (RC) support and more anti-RC support from partner, whereas more successful negotiation of goal conflict predicted higher RC support and lower anti-RC support. Attachment avoidance was directly associated with less support whereas attachment anxiety moderated the relationship between goal conflict and support. Implications for partner support during the pandemic are discussed.

1. Introduction

Because of COVID-19, many countries are engaging in social distancing measures including working from home, avoiding social situations, and shutting down schools causing wide social and economic disruptions (United Nations, 2020). Social distancing has meant that couples are staying at home without much outside support for an extended period of time. The change in circumstances is likely to increase goal conflict because couples are having to negotiate how to continue work-related tasks and pursue goals while simultaneously managing other demands. Couples also struggle to be responsive during highly stressful times (Neff & Karney, 2004). Indeed, conflict in relationships has increased since the start of the pandemic (Balzarini et al., 2020). Perceiving partners as responsive can buffer against the negative impact of pandemic-related stressors (financial strain, stress, conflict, and social isolation; Balzarini et al., 2020). Partner support may be more important than ever as couples strive to cope with demands caused by the pandemic, while at the same time goal conflict may deter partners from being as supportive as they otherwise might be. In the present daily diary and longitudinal study,1 our aim was to add to the literature by examining several novel research questions: whether goal conflict and negotiation of goal conflict predicted perceived support and whether individual differences in attachment styles prevented some individuals from perceiving partner support during COVID-19.

A recent theoretical model, thriving through relationships, describes the interpersonal process of how partners can create an optimal environment for thriving by providing relational catalyst (RC) support (Feeney & Collins, 2015). RC support involves being an active catalyst throughout the process of attaining goals including helping the recipient to recognize and view opportunities positively, providing a secure base for exploration, and helping with potential setbacks (Feeney et al., 2017). If the partner is able to provide effective RC support, the recipient is likely to perceive the partner as responsive (Feeney et al., 2017). However, if the partner is not responsive and instead is seen as intrusive, the recipient is likely to perceive them as providing anti-RC support. Feeney and Collins (2015) proposed that attachment-anxious individuals may have particular difficulty seeking RC support as pursuing independent goals may trigger anxious individuals' fear of losing their partner. Furthermore, avoidant individuals may find support seeking particularly difficult during stressful times. Therefore, we would expect individual differences in attachment styles to predict perception of partner support.

1.1. Goal conflict in relationships

While close relationship partners can help each other pursue goals, each partners' goals are likely at times to conflict with the interests of the other or the relationship. Repeated exposure to goal conflict is likely to be harmful for relationships because it continuously tests partners' commitment toward each other (Kelley & Thibaut, 1978; Rusbult & Van Lange, 2003). Goal conflict is indeed negatively associated with relational and personal well-being (Gere et al., 2011; Gere & Impett, 2018; Gere & Schimmack, 2013; Righetti et al., 2016). Only one study has examined whether goal conflict is associated with less support providing and found that more (vs. less) committed individuals support their partners only when goals do not pose a threat to the relationship (Hui et al., 2014). However, none have examined whether goal conflict predicts perceiving partners as supportive. We aim to investigate this novel question in the context of a global pandemic, a highly unusual and stressful situation. We expect that higher levels of goal conflict will be negatively associated with perceiving one's partner as providing RC support and positively associated with perceiving one's partner as providing anti-RC support (H1).

Furthermore, being able to successfully negotiate goal conflict is likely to influence perceived support. We are aware of no studies that have examined whether successful negotiation of goal conflict predicts support in romantic relationships. However, previous research examining how couples resolve general conflicts has found that more successful negotiation predicts greater relational well-being (Delatorre & Wagner, 2019; Kurdek, 1995). Given that successful negotiation of conflict predicts better outcomes and may enable partners to be more supportive, we expect that negotiation of goal conflict will be positively associated with perception of RC support and negatively associated with perception of anti-RC support (H2).

1.2. Attachment styles and goal conflict

Because of the uncertainty caused by the pandemic, need for safety and security is likely to be greater. At the same time, insecurities around partner availability may be heightened, leading more insecure individuals to perceive their partners as less supportive. Furthermore, attachment security has previously been associated with more adaptive response to stressful situations whereas attachment insecurity can impede ability to respond to stressors (for a review, see Mikulincer & Shaver, 2016). Insecure attachment is generally viewed along two dimensions: anxiety and avoidance (Brennan et al., 1998). High attachment anxiety is characterized by over-reliance on support and reassurance seeking from others because of feeling unlovable and unworthy of others (Mikulincer & Shaver, 2012). Attachment avoidance is characterized by a high degree of self-reliance and distrust of others' capacity to provide support in times of need (Mikulincer & Shaver, 2012). Therefore, we expect that higher (vs. lower) levels of attachment anxiety and higher (vs. lower) avoidance will be negatively associated with perceived RC support and positively associated with perceived anti-RC support, albeit for different reasons (Exploratory H12 ).

There are no published studies to date that have addressed whether individual differences in attachment styles moderate the association between goal conflict and support. However, theoretically we would expect attachment styles to influence how partners manage goal conflict in relationships. We expect that because attachment-anxious individuals are especially worried about maintaining closeness in relationships (Mikulincer & Shaver, 2012), they would experience goal conflict as threatening to the relationship. Therefore, we expect that when goal conflict is high, higher levels of attachment anxiety will negatively predict perception of RC support and positively predict perception of anti-RC support (Exploratory H2). We expect avoidant individuals to perceive their partners as less responsive overall regardless of goal conflict and therefore not significantly moderate the association between goal conflict and RC support.

2. Method

2.1. Participants and procedure

We preregistered the study on the Open Science Framework: https://osf.io/ght3x/?view_only=5b53d9e33690444e9cde8a6527775d2b. The data, code, and materials are here: https://osf.io/qr7cm/?view_only=365bf35f7ddd45548143b851e10cfcd9. The study received ethical approval from the institutional review board. Data were collected via Prolific. Participants who were aged 18 and above and currently living with their partner in a country in which social distancing measures were in place were invited to participate. We limited the number of participants for the quantitative surveys to 200 because of funding. Based on a simulated power analysis, data from 200 participants (up to 4200 observations) yield a power of 96.7% to estimate an average effect size in Psychology (r = 0.22; Richard et al., 2003) with an alpha level of p < .01. Participants were reimbursed £4.70 after completing the daily diary and a further £2.00 after completing all follow-ups.

Participants completed a baseline survey (demographic characteristics and attachment) on 31st March 2020, shortly after many countries had gone under lockdown. Participants then completed daily surveys for the next seven days in which they responded to questions about partner support, goal conflict, and negotiation of goal conflict from the past 24 h (see Fig. 1 for timeline). After the daily diary, participants completed a further three follow-ups, one week apart. This resulted in five weekly timepoints. Follow-up surveys asked similar questions to the daily diary but participants reported on the previous week rather than 24 h. Surveys were conducted via Qualtrics. The final sample was 200 with an attrition rate of 4% at the end of the diary and 8.5% at the end of the five weeks. All participants completed at least two timepoints and were included in the final analyses. Demographics can be found in Table 1 .

Fig. 1.

A graphical illustration of the study timeline.

Table 1.

Demographic variables.

	M	SD
Age	36.5	12.3
Relationship length	11.1	9.32

	N	%
Gender  Woman  Man  Other Sexual orientation  Heterosexual  Bisexual  Lesbian/gay  Other Relationship status  Married  Cohabiting Children  No  Yes Ethnicity  White  Black  Asian  Mixed Education  Graduated high school  Some college  Undergraduate  Postgraduate  Other Employment status  Employed full-time  Employed part-time  Self-employed  Student  Unemployed  Retired Employment changed  No  Yes Usually work from home  No  Yes Country  UK  USA  Other Keyworker  No  Yes Coronavirus symptomsa  No  Yes	105 93 2 182 9 7 2 102 98 95 105 184 5 6 2 28 38 74 52 8 121 23 26 4 7 9 153 47 138 62 119 17 64 166 34 179 21	52.5 46.5 1.0 91.0 4.5 3.5 1.0 51.0 49.0 47.5 52.5 92.0 2.5 3.0 1.0 14.0 19.0 37.0 26.0 4.0 60.5 11.5 13.0 2.0 3.5 4.5 76.5 23.5 69.0 31.0 59.5 8.5 32.0 83.0 17.0 89.5 10.5

^aNone had been diagnosed with coronavirus.

2.2. Measures

2.2.1. Attachment styles

Attachment was measured at baseline using the short version of the Experience in Close Relationships questionnaire (ECR-12; Lafontaine et al., 2016) which includes 12 Likert-scale items with two six-item subscales: anxiety (e.g., “I worry that my partner won't care about me as much as I care about them.”; α = 0.83) and avoidance (e.g., “I don't feel comfortable opening up to my partner.”; α = 0.85). Participants rated agreement with items on a scale from 0 (Disagree Strongly) to 10 (Agree Strongly).

2.2.2. Relational catalyst (RC) support

Partner support was measured using a shorter version of the Relational Catalyst Support Survey (Feeney & Collins, 2014). We shortened the original 32-item questionnaire to 8-items to reduce participant fatigue. Items were selected based on face validity; also, we included both positive and negative emotional and practical support items. Participants responded to items on a scale from 0 (Not at All) to 10 (Extremely). Because there are no published guidelines on the survey, we ran an exploratory factor analysis using principal axis factoring with varimax rotation to examine its factor structure. Results indicated two factors: RC support (e.g., “Has given me confidence to pursue my goals or opportunities”; α = 0.93) and anti-RC support (e.g., “Has been negative or demeaning when I am pursuing goals or opportunities”; α = 0.86).

2.2.3. Goal conflict and negotiation

At each time-point, we asked participants to list up to three goals that they had been working toward in the past 24 h (or the past week in the weekly follow-ups). The most common goals included domestic (31.4%), exercise/health (20.1%), career (16.4%), and hobbies/self-development (14.7%). Goal conflict was measured with two items: “How problematic was pursuing this goal for your partner?” and “How problematic was pursuing this goal for your relationship?” Participants were also asked “How well were you able to negotiate with your partner being able to work toward your goals?” Items were rated on a scale from 0 (Not at All) to 10 (Extremely).

2.3. Data analysis plan

We separated the within- and between-subjects' elements of the predictor variables (see Bolger & Laurenceau, 2013). The within-subjects variables show the difference in the outcome variables due to within-person elements and the between-subjects variable shows the average difference between participants' average scores in the outcome variables. Time was scaled to start at 0 and was included in both daily diary and weekly analyses as a control. Daily diary data and the weekly longitudinal data were separately analyzed using hierarchical linear modeling (Raudenbush & Bryk, 2002) with two levels (time-points nested within individuals). Attachment anxiety and avoidance were measured once at baseline and did not vary across timepoints. All models included only a random intercept as models with any random slopes failed to converge. Data were analyzed using the lme4 package in R. Several models were conducted to test the hypotheses and therefore we used an alpha level of p < .01 as a cutoff for significance. The models were run first only including the main effects and again including attachment in the model.

3. Results

We assessed two types of goal conflict: self-goals that conflict with partner's and relationship's goals.3 The results for relationship goal conflict are presented in the Supplemental material as they were similar to partner goal conflict. Descriptive statistics and correlations among all variables are presented in Table 2 .

Table 2.

Means, standard deviations, and correlations with confidence intervals for the diary.

Variable	M	SD	1	2	3	4	5	6	7
1. RC	6.04	2.76	–	−0.20	0.33	−0.13	−0.11	−0.43	−0.08
				[−0.24, −0.17]	[0.30, 0.37]	[−0.17, −0.09]	[−0.16, −0.08]	[−0.46, −0.40]	[−0.12, −0.04]
2. Anti-RC	1.07	1.75	−0.14	–	−0.25	0.25	0.21	0.23	0.15
			[−0.17, −0.11]		[−0.29, −0.22]	[0.21, 0.19]	[0.17, 0.25]	[0.19, 0.26]	[0.12, 0.19]
3. Negotiation	7.09	2.66	0.34	−0.20	–	−0.19	−0.19	−0.35	−0.06
			[0.31, 0.37]	[−0.23, −0.17]		[−0.22, −0.15]	[−0.23, −0.16]	[−0.38, −0.32]	[−0.10, −0.02]
4. ConflictP	1.39	2.39	−0.08	0.21	−0.14	–	0.63	0.22	0.12
			[−0.11, −0.05]	[0.17, 0.24]	[−0.17, −0.10]		[0.60, 0.65]	[0.18, 0.26]	[0.09, 0.16]
5. ConflictR	1.32	2.31	−0.11	0.20	−0.16	0.66	–	0.22	0.13
			[−0.14, −0.07]	[0.17, 0.23]	[−0.19, −0.13]	[0.64, 0.68]		[0.18, 0.25]	[0.09, 0.16]
6. Avoidance	2.27	1.74	−0.39	0.22	−0.31	0.22	0.24	–	0.16
			[−0.41, −0.36]	[0.19, 0.25]	[−0.34, −0.28]	[0.19, 0.25]	[0.21, 0.27]		[0.13, 0.20]
7. Anxiety	3.67	2.17	−0.08	0.14	−0.08	0.14	0.14	0.16	–
			[−0.11, −0.05]	[0.11, 0.17]	[−0.11, −0.05]	[0.11, 0.17]	[0.11, 0.17]	[0.13, 0.19]

Note. M = mean (across participants and time-points); SD = standard deviation. Values in square brackets indicate the 95% confidence interval for each correlation. The random measures correlations for daily diary data are presented below and for weekly data above the diagonal. Correlations for avoidance and anxiety with other study variables are zero-order correlations. All significant at p < .01.

3.1. Goal conflict and support

Consistent with our hypothesis that goal conflict would be associated with less RC support and more anti-RC support (H1; see Table 3 ), on days/weeks when participants felt their goals conflicted more with their partner's goals or relationship's goals, they reported their partner as providing less RC support and more anti-RC support. However, daily variation in goal conflict did not significantly predict a decrease in RC support for relationship goal conflict but did for partner goal conflict. The results for between-participants showed a similar pattern. We also hypothesized that negotiation of goal conflict would predict perception of support (H2; see Table 4 ). We found that on days/weeks when participants perceived that they and their partner were more successful in negotiating goal conflict they also perceived their partner as providing more RC support and on days/weeks when they perceived that negotiating goal conflict was less successful they experienced their partners as providing more anti-RC support. The results for between-participants showed a similar pattern.

Table 3.

Results from the Hierarchical Linear Modeling for Goal Conflict and Attachment as Predictors of RC and Anti-RC Support.

Predictors	RC support						Anti-RC support
	Daily			Weekly			Daily			Weekly
	Estimates	CI	p	Estimates	CI	p	Estimates	CI	p	Estimates	CI	p
Intercept	6.10	5.78–6.41	<0.001	6.25	5.93–6.57	<0.001	1.17	1.04–1.30	<0.001	0.96	0.84–1.08	<0.001
ConflictPW	−0.07	−0.10–-0.04	<0.001	−0.10	−0.13–-0.07	<0.001	0.12	0.10–0.14	<0.001	0.15	0.12–0.17	<0.001
ConflictPB	−0.14	−0.35–0.06	0.172	−0.22	−0.42–-0.01	0.040	0.71	0.63–0.79	<0.001	0.73	0.65–0.80	<0.001
Avoidance	−0.52	−0.71–-0.34	<0.001	−0.56	−0.75–-0.37	<0.001	0.01	−0.06–0.08	0.778	0.00	−0.07–0.07	0.993
Anxiety	0.01	−0.13–0.15	0.896	0.02	−0.13–0.17	0.792	0.00	−0.06–0.06	0.965	0.03	−0.03–0.08	0.349
Time	0.00	−0.02–0.03	0.747	0.00	−0.00–0.01	0.585	−0.03	−0.05–-0.02	<0.001	0.01	0.01–0.01	<0.001
ConflictPW ∗ Anxiety	0.01	0.00–0.02	0.025	0.02	0.01–0.03	0.002	−0.01	−0.01–0.00	0.188	−0.01	−0.02–−0.00	0.006
ConflictPW ∗ Avoidance	0.01	-0.00–0.02	0.088	0.01	−0.01–0.02	0.467	0.01	−0.00–0.02	0.067	0.02	0.01–0.03	0.002
Random effects
σ2	1.87			1.44			1.08			0.93
τ00	4.76ID			4.99ID			0.71ID			0.59ID
ICC	0.72			0.78			0.40			0.39
N	200ID			199ID			200ID			199ID
Observations	3769			2674			3769			2674
R2	0.135			0.165			0.434			0.491

Note. P = goal conflicts with partner's goals, W = within-participant change, B = between-participant change, ID = participant level.

Table 4.

Results from the hierarchical linear modeling for negotiation of goal conflict as a predictor of RC and anti-RC support.

Predictors	RC Support						Anti-RC Support
	Daily			Weekly			Daily			Weekly
	Estimates	CI	p	Estimates	CI	p	Estimates	CI	p	Estimates	CI	p
Intercept	6.06	5.80–6.33	<0.001	6.20	5.93–6.47	<0.001	1.19	0.98–1.39	<0.001	0.95	0.75–1.15	<0.001
NegotiateW	0.27	0.23–0.32	<0.001	0.25	0.20–0.30	<0.001	−0.12	−0.15–-0.09	<0.001	−0.15	−0.20–-0.11	<0.001
NegotiateB	0.83	0.71–0.95	<0.001	0.87	0.75–0.99	<0.001	−0.24	−0.33–-0.15	<0.001	−0.31	−0.40–-0.22	<0.001
Time	−0.00	−0.04–0.03	0.876	0.00	−0.01–0.01	0.923	−0.04	−0.07–-0.01	0.014	0.01	0.00–0.02	0.011
Random effects
σ2	1.92			1.60			1.20			1.09
τ00	2.71ID			2.90ID			1.60ID			1.41ID
ICC	0.59			0.64			0.57			0.56
N	200ID			200ID			200ID			200ID
Observations	1357			945			1357			945
R2	0.400			0.426			0.092			0.159

Note. Attachment was not a significant predictor or moderator and therefore not included in the table but can be found on the OSF page. W = within-participant change, B = between-participant change, ID = participant level.

3.2. Attachment anxiety and avoidance

Importantly, we expected that individuals higher in attachment anxiety or avoidance would perceive their partners as less supportive (Exploratory H1; Table 2). This hypothesis was partially supported. Attachment anxiety did not significantly predict perception of support. However, individuals higher in attachment avoidance were significantly less likely to perceive their partners as providing RC support both in the daily and weekly data and explained some of the between-person variability in the data; after including attachment avoidance in the model, between-person variability in goal conflict was no longer significant.

We also expected attachment anxiety to moderate the association between goal conflict and support (Exploratory H2; see Fig. 2 for the simple slopes). Attachment anxiety significantly moderated the association between goal conflict and RC and anti-RC support in the longitudinal weekly but not in the daily diary data. However, contrary to our prediction, on weeks when goal conflict was higher, participants higher in attachment anxiety (B = −0.06 (SE = 0.02), t = −3.08) perceived their partners as providing more RC support compared to participants lower in attachment anxiety (B = −0.14 (SE = 0.02), t = −6.85). Similarly, on weeks when goal conflict was higher, participants higher in attachment anxiety (B = 0.11 (SE = 0.02), t = 7.59) perceived their partners as providing less anti-RC support compared to participants lower in attachment anxiety (B = 0.18 (SE = 0.02), t = 10.44). Overall, individuals higher in attachment anxiety perceived their partners as providing more RC support and less anti-RC support when goal conflict was high compared to individuals low in attachment anxiety. An opposite pattern was observed on low conflict weeks: Participants lower in attachment anxiety perceived their partners as providing more RC support and less anti-RC support compared to participants high in attachment anxiety. We also explored whether attachment anxiety moderated the association between negotiation and support but none of the moderator effects were significant. We did not include these results here as they were not preregistered.

Fig. 2.

Results of the simple slope analyses depicting the association between goal conflict and RC and anti-RC support at different levels of attachment anxiety for the longitudinal weekly data.

4. Discussion

Because partner support is especially crucial during COVID-19 as social distancing measures prevent many people from seeking support from sources outside of their romantic relationship, the aim of the present study was to understand factors that may hinder effective support, including goal conflict and attachment insecurity. Both RC and anti-RC support remained stable over the study period suggesting any variation in support was due to factors other than time. Our results suggest that higher goal conflict is likely to result in suboptimal levels of support whereas successful negotiation of goal conflict is important for partners' ability to provide an optimal environment for goal pursuit. Interestingly, goal conflict predicted a much higher increase in anti-RC support than a decrease in RC support suggesting that goal conflict likely induces feelings that the partner is trying to hinder goal pursuit rather than simply withdrawing support. In contrast, successful negotiation of goal pursuit was more strongly predictive of perceiving one's partner as providing more RC support than less anti-RC support. It appears that negotiating goal conflict is especially important in increasing feelings that a partner is supportive of one's goal pursuits.4 Previous research has found that partners will provide support to the extent that they feel the goals do not take the partner away from the relationship (Feeney et al., 2017; Hui et al., 2014). While in the present study we did not have access to partner's reports of their own support providing, we showed that during higher goal conflict, the recipient perceived less support from their partner.

Furthermore, in line with previous research (Florian et al., 1995), we found that individuals higher in attachment avoidance perceived their partner as less supportive toward their goals overall. This finding is important because during social distancing measures, support from outside sources may not be accessible and a partner may be the most important source of support. Yet, avoidant individuals are viewing their partners as less supportive of goal pursuits. This may leave avoidant individuals especially vulnerable and lacking support during the pandemic.

In contrast, attachment anxiety did not predict perceptions of support which is contrary to the prediction by Feeney and Collins (2015). Attachment anxiety was, however, a moderator of goal conflict in the weekly longitudinal data. Contrary to our hypothesis, individuals high in attachment anxiety actually perceived their partners as more supportive compared to more secure individuals when goal conflict was high. These results are similar to findings by Pietromonaco and Barrett (1997) who found that anxious individuals perceive others more positively in high (vs. low) conflict situations, which anxious individuals may interpret as an intimate interaction. However, because these findings were unexpected, they should be replicated.

Strengths of the present study include the use of longitudinal data with both daily and weekly reports over the first weeks of many countries' lockdowns; furthermore, our hypotheses and analyses were preregistered. The study had some limitations. Data were collected from individual couple members, not dyads, and were based on one partner's perception.5 Future research should investigate whether partner's actual support behaviors are influenced by goal conflict and its negotiation. Researchers could also examine whether partner's attachment style moderates the association between goal conflict and providing support. For example, partners who are more anxiously attached may experience goal conflict as a threat and withdraw support. The participants reported relatively high levels of support and low levels of anti-RC support and goal conflict, therefore the study may have captured mainly participants who were coping relatively well with lockdown. We would expect, however, that the results would be even stronger in a sample with higher levels of goal conflict. Future research should target more distressed couples. Finally, the present study does not include pre-pandemic data from participants, thus it does not allow for within-participant comparisons.

In conclusion, the present study showed that goal conflict predicts a decrease in support whereas its negotiation predicts an increase in support. We show that avoidant individuals perceive their partners as less supportive, whereas anxious individuals perceive their partners as more supportive than secure individuals but only when goal conflict is high. The study also provides a unique perspective into how individuals are coping during one of the worst global health crises the world has experienced and suggests that helping individuals in relationships negotiate goal conflicts may enable them to be more available and responsive toward each other.

CRediT authorship contribution statement

Laura M. Vowels: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writing - original draft, Writing - review & editing, Visualization, Project administration, Funding acquisition. Katherine B. Carnelley: Methodology, Writing - review & editing, Supervision, Funding acquisition.

Appendix A. Supplementary data

Supplementary material