Abstract
Negative spousal interaction (e.g., criticism, arguments) can undermine psychosocial and relational well-being. The present research examined whether sense of purpose buffers these effects using two national longitudinal datasets (MIDUS and HRS). In MIDUS, associations between negative spousal interaction and same-wave life satisfaction, depression, relationship satisfaction, and marital strain were weaker among people reporting higher (versus lower) sense of purpose. Using HRS data, we disaggregated between- and within-person effects and found that sense of purpose similarly attenuated two between-person associations (linking average negative spousal interaction with depression and anxiety) and one within-person association (linking wave-to-wave fluctuations in negative spousal interaction with depression). Unexpectedly, between-person and within-person links between negative spousal interaction and relational well-being (enjoyment and closeness) were also moderated by purpose but in the opposite direction. This pattern of results appeared to be driven by purpose enhancing relational well-being among people who report low strain and during low-strain waves. Together, these findings indicate that sense of purpose may protect individuals from corrosive effects of negative partner interaction and enhance relationships during harmonious periods.
Introduction
Despite couples’ best efforts, virtually every romantic relationship involves some negative spousal interaction, defined as “exchanges or behaviors that involve excessive demands, criticism, disappointment, or other unpleasantness” (Sneed & Cohen, 2014, p. 554). People commonly report that their romantic partners communicate negatively (e.g., speak unkindly), let them down (e.g., downplay something important), and get on their nerves (e.g., are messy or rude; Luchies et al., 2017). Negative spousal interaction is a common and impactful form of relational stress, similar to chronic daily hassles (e.g., traffic, work deadlines; Hill et al., 2018; Mann et al., 2024; Wright et al., 2020). Like other stressors, negative spousal interaction is associated with poorer psychosocial and relational outcomes (e.g., Curran et al., 2010; Walen & Lachman, 2000) and the current research assesses whether sense of purpose—a protective factor against stress—buffers these associations.
Negative spousal interaction and well-being
Negative spousal interaction, also called spousal or marital strain (Schuster et al., 1990; Walen & Lachman, 2000), is typically operationalized by self-reports of the frequency and intensity of “the negative side of social exchanges” within one’s marital relationship (Walen & Lachman, 2000, p. 6). High levels are indicated by a respondent’s report that their spouse often criticizes them, makes too many demands on them, gets on their nerves, argues with them, and lets them down (Grzywacz & Marks, 1999), as well as reports of relational tension, frequent disagreements, and refusals to compromise (Schuster et al., 1990).
In a broad body of research, greater negative spousal interaction is consistently associated with worse psychosocial well-being. For instance, greater negative spousal interaction is associated with higher depressive symptoms, loneliness, and negative mood as well as lower life satisfaction and positive mood (e.g., Ermer & Proulx, 2022; Walen & Lachman, 2000), with some evidence suggesting particularly strong associations among women in different-sex marriages (Garcia & Umberson, 2019). People who experience high levels of spousal strain (third tertile) also prospectively report lower positive affect, life satisfaction, and optimism as well as higher depressive symptoms, hopelessness, negative affect, and loneliness, relative to people who experience low levels of spousal strain (first tertile; de la Rosa et al., 2025). Moreover, negative spousal interaction is theorized to increase the likelihood of experiencing psychopathological disorders among people who are predisposed to such disorders (e.g., South, 2021).
Negative spousal interaction also likely contributes to poorer relationship well-being. Indeed, there is an established association between greater negative spousal interaction and marital risk (defined as the perception that one’s marriage is in trouble and will eventually dissolve), both in the United States and in Japan (Curran et al., 2010; Jiao & Grzywacz, 2024). There is also some evidence connecting negative spousal interaction with relationship satisfaction, one’s “global evaluative or attitudinal reservoir of positivity toward the relationship” (Niehuis et al., 2024, p. 389). For instance, Henry et al. (2007) found that perceptions of negative partner behavior (hostility) during a problem-solving task were associated with lower marital satisfaction, especially for older couples. Similarly, couples who report more frequent arguments and/or report behaving more negatively during arguments (e.g., having tense and cruel interactions) are more likely to be categorized as “distressed” (i.e., as having critically low relationship satisfaction) and have greater odds of divorcing or separating (e.g., Fincham et al., 2018; McGonagle et al., 1993). One recent study also demonstrated an association between greater negative spousal interaction and lower perceived emotional intimacy (e.g., feeling distant from and neglected by one’s partner; Carasso & Segal-Karpas, 2024).
Despite the relative prevalence (DeLongis et al., 2004) of negative spousal interaction and its association with poorer psychosocial and relational well-being, many partnered people are still able to maintain satisfying lives and relationships (e.g., Luchies et al., 2017). In the present research, we explore a potential factor that explains how some people experience negative spousal interaction without serious disruptions to their psychosocial or relational well-being: sense of purpose.
Sense of purpose
Sense of purpose is the perception that life is meaningful, and it is characterized by a sense of directedness towards worthwhile goals (e.g., Ryff, 1989). It is related to, but distinct from, purpose in life (a person’s specific goals and aspirations) although the extant literature sometimes uses the terms interchangeably (Pfund, 2023). Sense of purpose is considered a personality trait (Pfund, 2023) although it fluctuates meaningfully over time (Pfund et al., 2024). Sense of purpose is also associated with psychosocial and relational well-being. For instance, people with a strong sense of purpose (highest quartile) prospectively report higher life satisfaction and positive affect relative to people in the lowest quartile of purpose (Kim et al., 2022). Moreover, meta-analytic results suggest higher purpose is associated with lower depression, anxiety (Boreham & Schutte, 2023), and loneliness (Sutin et al., 2022). Regarding relational well-being, first-year college students who reported a greater sense of purpose also reported more stable relationship satisfaction with parents and college friends across their first semester (Pfund et al., 2022). In romantic relationships, purposeful people report greater relationship satisfaction and commitment than less purposeful people (Pfund et al., 2020). High-purpose individuals (versus low-purpose individuals) are also less likely to dissolve their romantic relationship over time (Pfund & Hill, 2022).
Although sense of purpose and well-being are likely bidirectionally associated, work by Gudmundsdottir et al. (2024) suggests that purpose predicts within-person changes in subjective well-being more strongly than subjective well-being predicts changes in purpose. These findings suggest that sense of purpose may confer psychosocial and relational benefits, leading researchers to propose explanations for purpose’s potential benefits. Pfund et al. (2020) posit that a strong sense of purpose may enhance psychosocial and relational well-being by promoting beneficial thoughts and behaviors (“enhancement perspective”), whereas Burrow et al. (2024) postulate that sense of purpose may protect well-being by making people less reactive to momentary situations (“homeostatic perspective”). The present research is motivated by the latter possibility. We propose that people with a strong sense of purpose may be more resilient to negative spousal interaction, a variable with known negative associations with purpose (Irani et al., 2022; Weston et al., 2021).
In describing how sense of purpose may function as a protective factor, Burrow and colleagues (2024) posit that the “prospective and overarching” (p. 2) nature of purpose produces a “homeostatic effect” whereby purposeful individuals are less reactive to momentary situations. In other words, purposeful people exhibit greater emotional regulation amid life’s ups and downs because they consciously keep their focus on their overarching meaningful and worthwhile goals. Burrow et al. (2024) proposed “recentering,” or the ability to reframe a present situation with respect to broader life aims, as a potential mechanism explaining purpose’s benefits. For instance, a purposeful individual who recenters a difficult workday with respect to their broader aims (e.g., continual self-development) may reflect on how daily challenges provide an opportunity to grow. This aspirational perspective can help a purposeful person maintain equanimity and is especially important in the context of stress, where recentering facilitates a stable psychological state, limiting the impacts of stress on well-being.
Based on the homeostatic perspective, people high in sense of purpose might recenter negative spousal interaction with respect to their broader life aims, thereby reducing the impact of these relational stressors on psychosocial and relational well-being. For example, a purposeful person may recontextualize an argument by thinking, “Disagreements are a way to work through challenges and improve important relationships. Sharing our different viewpoints gives us opportunities to grow as individuals and as a couple.” Even a partner’s criticism or annoying behavior could be less harmful to a purposeful person who consciously shifts their focus from the immediate negative interaction to the broader meaningfulness of and goals for the relationship. Consequently, a purposeful person may better maintain psychological stability and relationship quality relative to a person with a lower sense of purpose.
The best available evidence for purpose’s proposed homeostatic effect is Hill et al.’s (2018) study of daily stress. For eight days, participants reported whether they experienced a stressor (including personal stressors such as work stress and social stressors such as arguments and social network stress), and they reported their daily affect and physical symptoms. Results indicated that while high-purpose and low-purpose individuals reported a similar number of stressors, high-purpose people were less impacted by stress. Specifically, they showed a smaller increase in negative affect and physical symptoms on stressor days relative to low-purpose individuals (Hill et al., 2018). In one subsequent study, Hill and his colleagues (2022) examined early effects of COVID-19 pandemic stressors and found that negative associations between daily stressors and positive affect were actually stronger among purposeful people, though this may be due to the unique circumstances surrounding this period (Hill et al., 2022). Thus, in the present research, we would still expect to find evidence supporting the buffering effects of purpose on stress. Specifically, we examine whether sense of purpose buffers the negative associations between negative spousal interaction (i.e., relationship stress) and poorer well-being, and we consider both psychosocial well-being and relational well-being.
The current research
The current research assesses whether sense of purpose moderates associations between negative spousal interaction and various psychosocial and relational well-being outcomes in two national studies of U.S. adults. In line with past research (e.g., Ermer & Proulx, 2022; Walen & Lachman, 2000), we expect to observe associations between negative spousal interaction (i.e., partner strain) and poorer psychosocial well-being (lower life satisfaction, greater depression, and greater anxiety) in both samples. Critically, we predict that these associations will be moderated by sense of purpose such that each association will be weaker for people reporting higher (versus lower) sense of purpose. This pattern of results would be consistent with Hill et al.’s (2018) finding that associations between daily stress and daily negative affect as well as physical symptoms are weaker for more purposeful people relative to less purposeful individuals.
We also expect to observe that negative spousal interaction is associated with poorer relational well-being. Although there is limited research linking negative spousal interaction with relational well-being, past theory suggests that behaviors in relationships contribute to global evaluations of the relationship (i.e., relationship satisfaction, relationship enjoyment, and relationship closeness) and the likelihood of relationship dissolution (i.e., marital risk). Critically, we hypothesize that the associations between negative spousal interaction and relational well-being will be weaker for people reporting higher (versus lower) sense of purpose. Overall, this pattern of results would be consistent with a “buffering” process whereby purposeful individuals maintain homeostasis when confronted with relationship stressors and are therefore able to navigate periods of negative spousal interaction without serious disruptions to their psychosocial or relational well-being (Burrow et al., 2024).
Method
Our pre-registrations for the MIDUS (https://osf.io/kfpnh) and HRS (https://osf.io/frmyn) samples are available on OSF. Analysis code and supplementary materials for the MIDUS and HRS studies, as well as a description of all deviations from the pre-registered plan using Willroth and Atherton’s (2024) template are available on OSF: (https://osf.io/6vh7d).
Participants
We examined all 3 waves of data from the publicly available Midlife in the United States (MIDUS) project (Brim et al., 2020; Ryff et al., 2019, 2021). Wave 1 data were collected in 1995–1996, Wave 2 data in 2004–2005, and Wave 3 in 2013–2014. We used the MIDUS Colectica Portal to download cleaned data: (https://midus.colectica.org). We also examined 5 waves of data from the Health and Retirement Study (HRS 2006, 2010, 2014, 2018, 2022). Wave 1 data were collected in 2006, Wave 2 in 2010, Wave 3 in 2014, Wave 4 in 2018, and Wave 5 in 2022. We used the RAND HRS Longitudinal Files. 1 The HRS is sponsored by the National Institute on Aging (grant number NIA U01AG009740) and is conducted by the University of Michigan.
Sample characteristics.
Note. Values represent N (%). We report every participant’s first value for each characteristic. * = Sexual orientation was only collected in the last two waves of the HRS; therefore, we avoid reporting percentage totals as a large portion of our sample does not have sexual orientation data.
Measures
Means, standard deviations, and between-person correlations.
Note. *p < .05 **p < .01. Values are presented as MIDUS/HRS. ‘--' indicates variable not available in that study. Age M and SD reported at baseline.
Negative spousal interaction (NSI)
NSI (i.e., partner strain/negative social support) was measured with either 6 (MIDUS) or 4 (HRS) items from Grzywacz and Marks’ (1999) spousal affectual solidarity indicators. Participants reported how frequently their partner makes too many demands on them, criticizes them, lets them down, gets on their nerves (in both MIDUS and HRS), argues with them, and makes them feel tense (in MIDUS). Participants reported on a 1 (often/a lot) to 4 (never/not at all) scale, which was reverse-coded and averaged (MIDUS α = .88 3 ; HRS α = .78).
Sense of purpose
Sense of purpose was measured with a 3-item (MIDUS Wave 1) and 7-item (MIDUS Waves 2–3; HRS) version of the Purpose in Life Subscale of the Ryff Scales of Psychological Well-Being (Ryff, 1989). The PIL includes items such as “some people wander aimlessly through life, but I am not one of them” and is scored on a 1 (strongly agree) to 7 (strongly disagree; MIDUS) or a 1 (strongly disagree) to 6 (strongly agree; HRS) scale. Participants’ responses on each question were either summed (MIDUS Wave 1 α = .36; Wave 3 α = .72) or averaged (HRS α = .76).
Life satisfaction
Life satisfaction was measured with a single item in MIDUS (“At present, how satisfied are you with your life?”) and with the Satisfaction with Life Scale in HRS (SWLS; Diener et al., 1985). The single item was scored on a 1 (very) to 4 (not at all) scale (reverse-scored) and the SWLS was mean-scored on a 1 (strongly disagree) to 7 4 (strongly agree) scale (e.g., “I am satisfied with my life”; α = .89).
Depression
Depression was measured with seven items from MIDUS’ continuous depression measure (Wang et al., 2000) and eight items from the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977) in HRS. Both measures were scored by summing the number of “yes” responses participants provided to items such as “feeling down on yourself, no good, or worthless” (MIDUS; see Blazer et al., 1994 for test-retest reliability and clinical validity) and “you felt sad” (HRS α = .81).
Anxiety
Anxiety was measured with MIDUS’ 10-item continuous anxiety measure (Wang et al., 2000) and five items from the Beck Anxiety Inventory (Beck et al., 1988) in HRS. The MIDUS measure was scored by summing the number of times that participants reported experiencing symptoms (e.g., feeling restless because of worry) “most days” (see Blazer et al., 1994 for test-retest reliability and clinical validity). The Beck Anxiety Inventory was scored by averaging items that asked about participants’ feelings over the past week (e.g., “I was nervous”) on a 1 (never) to 4 (most of the time) scale (α = .82).
Relationship satisfaction
In MIDUS, relationship satisfaction was measured with a single item, “How would you rate your marriage or close relationship these days?” scored on a 0 (Worst) to 10 (Best) scale. Unfortunately, there was no clear measure of relationship satisfaction in the HRS. The best available item (the one that is consistent with what is assessed in satisfaction measures; Funk & Rogge, 2007) was “Overall, how enjoyable is the time you spend together with your spouse/partner?”, rated from 1 (extremely enjoyable) to 4 (not too enjoyable). Another item assessed general perceptions of closeness, which although not the same as relationship satisfaction, is another broad evaluation of the relationship. That item was, “How close is your relationship with your partner or spouse?”, rated from 1 (very close) to 4 (not at all close). We opted to assess both relationship enjoyment and closeness, both reverse-scored, in separate analyses because they measure separate constructs (this differed from our pre-registered plan; see OSF).
Marital risk
Marital risk was measured in MIDUS with either 5 (Wave 1) or 2 (Waves 2 and 3) items assessing participants’ beliefs regarding the termination of their marriage. Items included in all waves were, “During the past year, how often have you thought your relationship might be in trouble?”, scored from 1 (never) to 5 (all the time) and “… realistically, what do you think the chances are that you and your partner will eventually separate?”, reverse-scored on a 1 (very likely) to 4 (not likely at all) scale. Responses were averaged in Wave 1 (α = .69) and summed in Waves 2 and 3 (α = .72). Marital risk was not assessed in HRS.
Analytic strategy
All analyses utilized multilevel modeling with random intercepts for participants to account for repeated measures nested within participants. However, we used different models to analyze the MIDUS and HRS datasets, given design differences in the studies. Our primary goal was to assess whether the associations between NSI and various psychosocial and relational well-being outcomes were weaker among people reporting higher (versus lower) sense of purpose, consistent with the homeostatic perspective. Ideal models would estimate both between- and within-person associations linking NSI to outcomes and would test whether each is moderated by time-varying sense of purpose, given evidence that purpose changes over time (Pfund et al., 2024).
A positive between-person association would indicate that people who, on average, experience greater NSI, tend to report greater depression relative to people who experience less NSI. This association could be moderated by sense of purpose to demonstrate that experiencing NSI is less costly among people who are also reporting a higher (versus lower) sense of purpose. A positive within-person association would indicate that participants report higher depression during waves when NSI exceeds their personal average. This association could be moderated by sense of purpose to demonstrate that, even during periods of high NSI, people who report greater purpose are more resilient.
We were able to estimate both between- and within-person effects in the HRS because up to five waves of data were available, providing sufficient occasions to compute meaningful person-centered NSI scores. In contrast, a maximum of three waves (of which, we often had fewer) of data were available in the MIDUS. Therefore, for the MIDUS, we analyzed all three waves of data together without disaggregating within- and between-person variance. Our five MIDUS models tested whether NSI (grand-mean centered) at a given wave was associated with well-being at that same wave, and whether that association was moderated by same-wave sense of purpose (grand-mean centered). This approach provides a robust test of our focal moderation hypothesis in a way that uses all available data to maximize statistical power to observe the hypothesized interaction effect. However, because it does not disaggregate within-person and between-person variance, observed associations reflect both stable differences between individuals and fluctuations within individuals over time.
In the HRS, we tested moderation by sense of purpose by including two interaction terms across five models. First, we examined between-person associations by testing whether individuals higher in sense of purpose (at that wave; grand-mean centered) demonstrated buffered associations between their average NSI (across all waves; grand-mean centered) and well-being outcomes. Second, we examined within-person processes by testing whether individuals higher in sense of purpose (at that wave; grand-mean centered) were protected from well-being deficits in years they reported higher NSI relative to their own average level of NSI (person-centered). The first interaction informs whether purpose generally attenuates well-being costs associated with greater NSI while the second addresses whether purpose fosters resilience during periods of relationship discord.
Though we did not pre-register it (see OSF), we also controlled for age, sex, education, and self-rated health 5 in all models, similar to previous moderation studies involving purpose, stress, and well-being (Hill et al., 2018, 2022). POMP-scored age and health were both grand-mean centered for analyses. We selected these variables given known associations between psychosocial well-being and age (Steptoe et al., 2015), sex (Wood et al., 1989), education (Nikolaev, 2018), and health (Ngamaba et al., 2017).
When we observed significant interactions, we used the Interactions package in R (Long, 2024) to plot the interactions and to decompose the interactions by estimating simple slopes. Specifically, we estimated the slope of NSI at high (+1 SD) and low (−1 SD) levels of sense of purpose, as well as the slope of sense of purpose at high (+1 SD) and low (−1 SD) levels of NSI. Decomposing interactions in this manner allows us to observe conditional slopes of our predictor variables at specified levels (i.e., high vs. low) of the moderator (Long, 2024).
Results
MIDUS sample
Moderation results for MIDUS sample, adjusted for covariates.
Note. NSI = Negative spousal interaction. Estimates calculated while controlling for covariates, but covariate estimates omitted; see OSF for full model output. All variables POMP-scored. All predictors centered after being POMP-scored.
MIDUS interaction plots
As predicted, the negative association between NSI and life satisfaction at a given wave was weaker for people higher in sense of purpose (b = −0.15, 95% CI [−0.18, −0.13], p < .001) relative to people lower in sense of purpose (b = −0.23, 95% CI [−0.25, −0.21], p < .001). Decomposing this interaction in the opposite direction, the positive relationship between sense of purpose and life satisfaction was stronger for people reporting higher NSI (b = 0.20, 95% CI [0.17, 0.23], p < .001) relative to people reporting lower NSI (b = 0.11, 95% CI [0.08, 0.14], p < .001).
The positive association between NSI and depression at a given wave was also weaker for those higher in sense of purpose (b = .04, 95% CI [.01, .07], p = .01) relative to those lower in sense of purpose (b = .11, 95% CI [.08, .14], p < .001). Decomposing this interaction in the other direction, the negative relationship between sense of purpose and depression was stronger for those with higher NSI (b = −.14, 95% CI [−.18, −.11], p < .001) compared to those with lower NSI (b = −.07, 95% CI [−.10, −.03], p < .001).
The interaction term for the anxiety model was nonsignificant (b = .00, p = .164), but there were significant main effects of purpose (b = −.02, 95% CI [−.03, −.01], p < .001) and NSI (b = .02, 95% CI [.01, .03], p < .001) on anxiety.
The inverse association between NSI and relationship satisfaction at a given wave was weaker for people with higher sense of purpose (b = −.52, 95% CI [−.54, −.50], p < .001) relative to people with lower sense of purpose (b = −.60, 95% CI [−.62, −.58], p < .001). Decomposing this interaction in the opposite direction, the positive relationship between sense of purpose and relationship satisfaction was stronger for those reporting higher NSI (b = .14, 95% CI [.12, .16], p < .001) relative to those reporting lower NSI (b = .05, 95% CI [.02, .07], p < .001).
Finally, the positive association between NSI and marital risk at a given wave was weaker for more purposeful people (b = .53, 95% CI [.51, .55], p < .001) relative to less purposeful people (b = .58, 95% CI [.56, .60], p < .001). Decomposing this interaction in the other direction, the negative relationship between sense of purpose and marital risk was stronger for people reporting higher NSI (b = −.10, 95% CI [−.12, −.08], p < .001) relative to those reporting lower NSI (b = −.04, 95% CI [-.06, −.02], p < .001).
HRS sample
Moderation results for HRS sample, adjusted for covariates.
Note. NSI = Negative spousal interaction. Avg. = Average. PC = Person-centered. GMC = Grand-mean centered. Estimates calculated while controlling for covariates, but covariate estimates omitted; see OSF for full model output. All variables POMP-scored. All predictors centered after being POMP-scored.
HRS interaction plots for psychosocial outcomes
HRS interaction plots for relational outcomes
Regarding depression, we observed that the between-person association between average NSI and depression was weaker for participants reporting higher sense of purpose (b = 0.13, 95% CI [.11, .16], p < .001) than for participants lower in purpose (b = 0.21, 95% CI [.19, .23], p < .001). Similarly, the within-person association between NSI (person-centered) and depression was weaker for those higher in purpose (b = 0.03, 95% CI [.00, .05], p = .04) than for those lower in purpose (b = 0.09, 95% CI [.07, .12], p < .001). Decomposing these interactions in the opposite direction, the negative association between sense of purpose and depression was stronger for participants higher in average NSI (b = −.22, 95% CI [−.24, −.20], p < .001) relative to those lower in NSI (b = −.14, 95% CI [−.16, −.12], p < .001). Moreover, the negative relationship between sense of purpose and depression was stronger when participants reported higher-than-average NSI (person-centered; b = −.20, 95% CI [−.22, −.18], p < .001) relative to when participants reported lower-than-average NSI (b = −.16, 95% CI [−.18, −.14], p < .001).
Regarding anxiety, the between-person association between average NSI and anxiety was weaker for participants reporting higher purpose (b = .13, 95% CI [.10, .15], p < .001) than for participants lower in purpose (b = .19, 95% CI [.17, .21], p < .001). Decomposing this interaction in the opposite direction, the negative relationship between purpose and anxiety was stronger for participants higher in average NSI (b = −.27, 95% CI [−.29, −.24], p < .001) relative to those lower in NSI (b = −.20, 95% CI [−.22, −.18], p < .001). As the interaction term for person-centered NSI and average purpose was non-significant, we avoid interpreting the result here.
Concerning relationship enjoyment, the negative between-person association between average NSI and relationship enjoyment was stronger for participants reporting higher sense of purpose (b = −.67, 95% CI [−.70, −.63], p < .001) than for participants lower in purpose (b = −.58, 95% CI [−.61, −.55], p < .001). Similarly, the within-person association between higher-than-average NSI (person-centered) and relationship enjoyment was stronger for those reporting higher purpose (b = −.36, 95% CI [−.40, −.31], p < .001) than those reporting lower purpose (b = −.30, 95% CI [−.33, −.26], p < .001). Decomposing these interactions in the opposite direction, the association between sense of purpose and relationship enjoyment was weaker for participants higher in NSI (b = .15, 95% CI [.12, .18], p < .001) relative to those lower in NSI (b = .24, 95% CI [.21, .27], p < .001). Moreover, the relationship between sense of purpose and relationship enjoyment was weaker when participants reported higher-than-average NSI (person-centered; b = .18, 95% CI [.15, .21], p < .001) relative to when participants reported lower-than-average NSI (b = .21, 95% CI [.18, .24], p < .001).
Considering relationship closeness, the negative within-person association between NSI and relationship closeness was stronger for those reporting higher purpose (b = −.33, 95% CI [-.36, −.30], p < .001) than for those lower in purpose (b = −.29, 95% CI [-.31, −.27], p < .001). Decomposing this interaction in the opposite direction, the association between sense of purpose and relationship closeness was weaker when participants reported higher-than-average NSI (person-centered; (b = .12, 95% CI [.10, .14], p < .001) relative to when participants reported lower-than-average NSI (b = .15, 95% CI [.13, .17], p < .001). The interaction term for average NSI and sense of purpose was non-significant.
Neither of the interaction terms in the life satisfaction model were significant, although we did observe a main effect of sense of purpose (b = .28, 95% CI [.26, .30], p < .001) on life satisfaction.
Discussion
Romantic partners sometimes argue, criticize one another, and let each other down. These stressful examples of negative spousal interaction are common (Luchies et al., 2017) and can be damaging for individuals (Ermer & Proulx, 2022; Walen & Lachman, 2000) and their relationships (Fincham et al., 2018; Carasso & Segal-Karpas, 2024). The present study examined whether sense of purpose might protect against these costs of negative spousal interaction. In two national samples (MIDUS and HRS), we found evidence that people reporting a stronger sense of purpose were less vulnerable to some psychosocial and relational costs of negative spousal interaction. We also observed some disparate findings that point to a second pathway through which sense of purpose may benefit individuals and their relationships, as described below.
Consistent with Burrow and colleagues’ (2024) homeostatic account of purpose, several of our findings suggest that sense of purpose may serve as a stabilizing resource that helps individuals recenter during chronic or temporary periods of relational strain. Across both datasets, sense of purpose attenuated links between negative spousal interaction and poorer psychosocial well-being measures. In the MIDUS sample, people who reported higher (versus lower) sense of purpose experienced weaker associations between their reports of negative spousal interaction and psychosocial well-being deficits (life satisfaction and depression). In the HRS sample, sense of purpose similarly buffered between-person links between chronic (higher average) negative partner interaction and both depression and anxiety, as well as within-person associations between fluctuations in negative spousal interaction and depression. These novel results extend prior research showing that sense of purpose mitigates the associations between general stressors and daily negative affect (Hill et al., 2018) by demonstrating that the homeostatic effect of purpose extends to relational stressors. People who are purposeful appear to be more equipped to absorb the impacts of everyday spousal strain, maintaining their psychosocial equilibrium in the face of criticism, conflict, and excessive demands.
Our findings also provide novel evidence that the homeostatic effect of sense of purpose extends beyond psychosocial well-being to relational well-being. In the MIDUS sample, sense of purpose moderated links between negative spousal interaction and relational well-being such that spousal strain was less strongly linked to relationship dissatisfaction and marital risk for people reporting higher (versus lower) sense of purpose. This pattern of results suggests that sense of purpose enables people to avoid interpreting negative spousal interactions as evidence that their relationship is in trouble and likely to dissolve. Instead, consistent with the homeostatic model of purpose (Burrow et al., 2024), sense of purpose may induce a “recentering” effect and enable purposeful individuals to recontextualize criticism, disappointment, and the like, with respect to their meaningful life goals (or relationship goals), thereby reducing their reactivity to these relational stressors. More concretely, sense of purpose may enable people to make more adaptive and more constrained appraisals of negative spousal interactions, seeing them as temporary and specific disruptions rather than persistent and global relationship threats.
Not all of our results were consistent with the hypothesized buffering pattern. Most notably, in the HRS data, people higher in sense of purpose showed a stronger negative association between negative spousal interaction and poorer relational well-being (relationship enjoyment and closeness) at both the between- and within-person levels. However, inspection of the plotted interactions (see Figure 3) suggests that purposeful individuals report especially high relationship enjoyment and closeness at low levels of negative spousal interaction, rather than particularly low relationship enjoyment and closeness when negative spousal interaction is high. In other words, sense of purpose did not exacerbate spousal strain, it seems to have amplified relational thriving when relationships were harmonious. These findings—as well as the few models with main effects of sense purpose rather than moderations by purpose—are consistent with Pfund et al.’s (2020) enhancement perspective. The enhancement perspective suggests that sense of purpose promotes psychosocial and relational well-being not only by buffering stress, but also by fostering positive cognitions (e.g., hope), feelings (e.g., positive affect), and behaviors (e.g., being organized), all of which likely benefit individuals and their relationships. The enhancement perspective suggests that sense of purpose may provide benefits in both adverse and harmonious relational contexts. People high in purpose may even be able to reap the greatest relational rewards (enjoyment, closeness) in contexts of low marital strain because they are able to actively enhance their relationships through pro-relational thoughts and behaviors.
Taken together, the results presented herein support a dual-function model of purpose in relationships. During times of marital strain, sense of purpose may serve a homeostatic function, helping individuals to reframe relational stressors and preserve psychosocial and relational well-being. During times of relational harmony, purpose may instead operate as a relational enhancer by promoting pro-relational cognitions, feelings, and behaviors.
Limitations
Although we speculate about the mechanistic processes underlying our findings, a primary limitation of this work is that we have no data to establish mechanisms. Future research should measure the theorized recentering process more directly by assessing individuals’ thoughts about negative spousal interactions (e.g., attributions and reframing) and their emotion regulation abilities in that context. For example, work indicates that purposeful people are better able to self-regulate in anxiety-producing situations (Pfund et al., 2023), which might help explain the homeostatic process. Future research should also assess mechanisms suggested by the enhancement perspective to better establish how these dual functions of purpose may work in tandem.
This work was also limited by the measures available in MIDUS and HRS. Despite (and perhaps because of) the impressive scope of these projects, some measures were brief (e.g., relationship quality) or used nonideal response scales (e.g., depression measures), making them less informative and reliable relative to multiple-response items (Ali et al., 2015).
Finally, future work should consider how a person’s particular purpose in life (McKnight & Kashdan, 2009) influences these processes, something we were unable to do with these datasets. It may be the case that alignment between one’s purpose in life and a specific threatening context amplifies purpose’s homeostatic benefit. For example, purpose may buffer spousal strain most strongly when participants’ purposes in life connect to their role as romantic partner. Relationship quality and spousal support are both known sources of purpose (Nakamura et al., 2022; Pfund & Hill, 2022) and so having a relational form of purpose in life may make individuals particularly motivated to accommodate negative partner interactions in service of maintaining their relationship.
Conclusion
Despite the many benefits of partnership, romantic relationships inevitably involve criticism, excessive demands, and other negative spousal interactions that can undermine both the relationship and the well-being of the people in it. The present research demonstrated that a strong sense of purpose may help people to weather these challenges and to strengthen their relationships during more harmonious periods. By extending evidence for the moderating effects of purpose to the relational domain, this work shows that purpose’s homeostatic effects generalize from personal stressors to relational ones and from psychosocial well-being to relational well-being. Our results also support the enhancement perspective, suggesting that sense of purpose not only buffers against marital strain but amplifies closeness and relational enjoyment when strain is low or absent. Together, these findings highlight sense of purpose as a potent relational resource that protects relationships from relational strain and enriches them in times of calm.
Consent to participate
Participants in the MIDUS and HRS projects provided informed, written consent.
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Open research statement
As part of IARR's encouragement of open research practices, the author(s) have provided the following information: This research was pre-registered. The aspects of the research that were pre-registered were the hypotheses and analysis plan. The registration was submitted to: (https://osf.io/kfpnh) and (https://osf.io/frmyn). The data used in the research are available. The data can be obtained at: (https://midus.colectica.org for Study 1 and https://hrsdata.isr.umich.edu for Study 2). The materials used in the research are available. The materials can be obtained at: https://osf.io/6vh7d/.
Ethical considerations
The MIDUS project was approved by the Institutional Review Board at the University of Wisconsin-Madison, IRB#: 2016–1051. The HRS was approved by the University of Michigan Institutional Review Board.
Data Availability Statement
The MIDUS (https://midus.colectica.org) and HRS (
) data are both publicly available on their respective repositories.