The expression of compassionate love in everyday compassionate acts

Abstract

Compassionate love (CL) is a form of altruistic, caring love that emphasizes concern for the other’s well-being. How is CL expressed in marriage? To address this question, we adopted a behavioral acts perspective in which we examined among newlyweds the presence or absence of a series of behavioral manifestations of CL. A sample of 175 newlywed couples completed daily diaries for 2 weeks, describing their own compassionate acts and their perceptions of their partners’ compassionate acts. We found clear evidence that CL acts contributed to both spouses’ daily marital satisfaction. These effects were independent of general positivity or negativity of behavior. We also employed a quasi-signal detection analysis to determine whether acts that both partners recognize as compassionate are more influential than acts that only one of them acknowledges. This hypothesis was supported. Our findings indicate that CL-related behavior represents a beneficial form of caring interaction that merits further attention in relationship research.

Keywords

Love marriage

Scholars have long recognized that the word “love” is far too ambiguous to represent the many kinds of feelings and experiences for which people use this word. Plato, as an early example, distinguished love as sexuality, as the search for a soul mate, or as a means of contemplating the divine in another person (Aron & Davies, 2009). Over the millennia, philosophers, poets, lay writers, and scientists have proposed various typologies of love, a list that includes, in the past half century, several typologies developed by relationship researchers. In a recent review, Berscheid (2010) argued that a fourfold distinction was most theoretically coherent: companionate love, romantic love, adult attachment love, and compassionate love (CL). The first three of these have been investigated extensively (see Reis & Aron, 2008, for a review of the first two, and Mikulincer & Shaver, 2007, for a review of the third); CL, in contrast, is a relative newcomer to the relationship–science literature and has received scant attention. This article describes research intended to add to what is known about the nature and impact of CL in romantic relationships.

CL is a form of altruistic, caring love that emphasizes concern for the other’s well-being. As elaborated below, CL is freely given, involves accurate, contextually grounded understanding of the other, features validation and genuine acceptance of the other, and is expressed through openness, receptivity, and warmth (Underwood, 2009). Perhaps more than other types of love, CL emphasizes the well-being of the other, not as perceived solely from one person’s perspective but rather from the intersecting perspectives of both provider and recipient. This intrinsically relational conceptualization may be one reason why the construct appeals to relationship researchers. Although existing constructs such as empathy, altruism, caregiving, passionate love, companionate love, attachment security, and social support embrace several of these components, none of them integrates the varied properties of Underwood’s definition into a single construct.

For several reasons, we believe that newlyweds offer an ideal window for examining the interpersonal expression and effects of CL. In the contemporary Western world, marital relationships provide an organizing framework for many of life’s most important and psychologically significant activities (Coontz, 2005). Indeed, for most couples, expectations about mutual caregiving and compassion are foundational as they enter marriage and the realization of these goals is associated with marital satisfaction (Clark, Lemay, Graham, Pataki, & Finkel, 2010). Yet, the first year of marriage is often filled with disappointment, as high expectations yield to the demands of everyday routines and increasing interdependence (e.g., Bradbury, Fincham, & Beach, 2004; Kurdek, 1998). Compassionate caring may serve as a buffer against this trend, by providing opportunities for partners to demonstrate benevolence toward each other (Neff & Karney, 2009). On the other hand, the absence of compassionate caregiving is associated with decreases in satisfaction (Clark et al., 2010). Marital interaction among newlyweds may therefore be an ideal context for revealing the behavioral and emotional manifestations of CL.

The research described in this article was designed to examine CL from a behavioral acts’ perspective (Buss & Craik, 1983). By behavioral acts, we refer to the everyday behaviors through which individuals manifest CL. Because CL is primarily represented as a feeling or disposition toward others, existing research has tended to focus on its association with attitudes, beliefs, and affects toward both others in general and particular relationship partners (Fehr & Sprecher, 2009; see e.g., Fehr, Sprecher, & Underwood, 2009). Nonetheless, to have its theorized beneficial impact on others, CL must be expressed somehow in behavior; otherwise there would be no apparent vehicle for its transmission from one person to another (as would be the case for any type of love). In other words, we examined the ways in which relationship partners display CL to each other—be it in an expression of caring or concern, a willingness to do what a partner asks or wants, or a warm smile at the end of a long day.

Another contribution of this research is our focus on everyday CL. Studies of everyday social interaction—ordinary, spontaneous behaviors, both meaningful and trivial, expressed in their natural context—provide a unique window on dyadic processes that complement laboratory experiments and general surveys (Bolger, Davis, & Rafaeli, 2003; Reis, Gable, & Maniaci, in press). A daily diary study may be informative about CL for at least four reasons. First, CL expressly concerns voluntary actions (Underwood, 2009), which are best studied when the context allows them to emerge spontaneously. Second, CL seems likely to appear in the ebb and flow of ordinary experience—for example, how spouses show kindness and understanding toward each other during routine interactions or after long, stressful days at work. This perspective may be particularly revealing in the first year of marriage, inasmuch as adjustment to the fact of increasing interdependence in ordinary life activities is one of the challenges that newlyweds face (Murray & Holmes, 2011). Third, daily diaries possess the methodological advantage of minimizing retrospection bias (Bolger et al., 2003; Schwarz, 2012), an artifact that seems particularly feasible with a subjective construct like CL. The final reason is that our focus on specific behavioral acts facilitates an interpersonal perspective, as described below.

Our daily diary approach offers one additional advantage to work on this emerging topic, by contributing to a descriptive ecology of CL. Little is currently known about how, when, and to what extent CL occurs in everyday relational life. Accurate, detailed description can meaningfully influence theory development, as it did in the biological and physical sciences. Asch (1952) pointed to a similar need in social psychology, commenting that “(b)efore we inquire into origins and functional relations, it is necessary to know the thing we are trying to explain” (p. 65). Conceptual understanding of CL may be better informed if researchers have available a descriptive data base about its frequency, patterning, and consequences (Hinde, 1995).

Characterizing CL

Before we could examine behavioral expressions of CL, it was necessary to develop a precise operational definition of this construct as it would manifest in newlywed couples’ behavior. Several somewhat varying definitions have been offered (e.g., Berscheid, 2010; Fehr et al., 2009; Underwood, 2009), and we decided to rely on four qualities that are included in each of them.

Concern for the other’s well-being

This feature refers to one partner’s willingness to promote the other’s well-being, even when it is costly to the self to do so. Rusbult, Olsen, Davis, and Hannon (2001) refer to this as a pro-relationship transformation of motivation, which encompasses familiar phenomena such as accommodation, forgiveness, and willingness to sacrifice. More generally, concern for a partner’s well-being is the basis of communal relationships (Clark & Mills, 2012) and can promote partners’ goal attainment and movement toward their ideal self (Rusbult, Kumashiro, Stocker, & Wolf, 2005). Relationship partners also express concern for each other’s well-being through tenderness and affection (Floyd, 2006; Neff & Karney, 2005).

Understanding and acceptance

A central premise of CL is that it is grounded in each partner’s genuine attempts to adopt the other’s perspective; that is, to understand circumstances as the other sees them. Several relationship theories emphasize the importance of making genuine attempts to understand a partner’s perspective, because those attempts provide a foundation for feeling understood, a key determinant of responsiveness and intimacy (Finkenauer & Righetti, 2011; Murray, Holmes, Bellavia, Griffin, & Dolderman, 2002; Reis & Clark, 2013). Moreover, people are most likely to feel securely accepted by a partner when they believe that the partner “gets the facts right” about themselves (Murray, Holmes, & Collins, 2006; Reis & Clark, 2013).

Respect and admiration

Beyond feeling understood, people wish to feel valued by their partners; after all, any sense of understanding by a scornful or disapproving partner would likely undermine feelings of security and belonging. Validation serves the important purpose of communicating positive regard for the self, arguably the primary determinant of people’s sense of security in close relationships (Murray et al., 2006). Respect and admiration, which Underwood (2009) calls “valuing the other at a fundamental level,” also support the self-concept, in the sense of validating one’s self-worth and world view (Reis & Shaver, 1988).

Openness

Perceiving that a partner is receptive to one’s needs and wishes is prerequisite to the above attributes, because an unreceptive partner would be unlikely to express concern, acceptance, or respect. Openness signals a willingness to set aside one’s own needs, preferences, or opinions and voluntarily take account of the other’s needs, preferences, and opinions. Of course, openness reflects communication; if unrecognized by one’s partner, even a high level of openness would probably be ineffective (MacGeorge, Feng, & Burlseon, 2011).

An interpersonal perspective

For this research, we developed an inventory of everyday CL acts based on the above four criteria. As the definitions illustrate, CL is inherently interpersonal: each component depends on the motives and actions of one partner and how those motives and actions are perceived by the other. In this respect, our model of CL is similar to other fundamentally interpersonal models, such as Reis and Shaver’s (1988; Reis and Patrick, 1996) intimacy model and Rusbult’s model of mutual cyclical growth (e.g., Weiselquist, Rusbult, Foster, & Agnew, 1999). Accordingly, this inventory was intended to assess behavioral acts of CL from both partners’ corresponding perspectives: whether or not the act was performed on a given day and whether or not the other partner noticed those actions (or nonactions).

A dyadic perspective differs from other studies that investigate related processes solely from one person’s vantage point—for example, by assuming that giving support or feeling supported is beneficial, regardless of whether giver and recipient see the relevant act similarly. Underwood (2009) theorized that CL is most likely to be beneficial when partners have a shared perspective about their interaction. Yet, relationship partners often disagree about their interactions, even when focusing on relatively concrete behaviors (e.g., Baxter & Bullis, 1986; Thomas & Fletcher, 1997). These differences may reflect motivated cognition—for example, people tend to interpret partners’ behavior in a manner that reflects their motives, personality traits, or situationally specific goals (Baldwin, 1992; Murray, 1999). Differing perceptions may also reflect the complexity of discerning intent against a background of environmental “noise” (i.e., the myriad stimuli that compete for attention in daily life).

We examined the impact of CL acts in our newlywed couples by jointly considering the actor and partner perspectives on behavior. We did this by including parallel items for each action, one asking participants whether or not they had enacted a given behavior on that day and the other asking whether or not their partners had enacted a given behavior on that day. We then used these accounts in two ways. First, to distinguish the effects of one’s own and a partner’s actions on each partner’s marital outcomes, we used the actor–partner interdependence model (APIM; Kenny, Kashy, & Cook, 2006). This method identifies the impact of each partner’s reports of their own behavior (or their perceptions of their partner’s behavior) on each partner’s outcomes. In this analysis, actor effects refer to the impact of one’s actions (or perceptions of a partner’s actions) on one’s own relationship outcomes, whereas partner effects refer to the impact of one’s actions (or perceptions of a partner’s actions) on the partner’s relationship outcomes. As described above, we expected that compassionate acts would predict the daily relationship satisfaction of both provider and recipient.

Second, to better integrate actors’ reports of the behaviors they enacted with their partners’ perceptions of those same behaviors, we used the quasi-signal detection method, which examines correspondence and noncorrespondence in partners’ daily reports (Gable, Reis, & Downey, 2003; Finkenauer, Wijngaards-de Meij, Reis, & Rusbult, 2010). This paradigm uses responses from both partners to classify each behavioral report as a hit (one spouse reported enacting the behavior and the other perceived it), miss (one spouse reported enacting the behavior, while the other did not perceive it), false alarm (one spouse reported not enacting the behavior but the other perceived it), or correct rejection (one spouse reported not enacting the behavior and the other did not perceive it). In prior studies (e.g., Gable et al., 2003; Reis & Carmichael, 2006), we collected daily diary reports of various interactive behaviors over several weeks in samples of married or cohabiting adults (e.g., “I was physically affectionate toward my partner” and, in parallel, “My partner was physically affectionate toward me”). Findings in those studies revealed that for positive and supportive acts, hits predicted increased relationship well-being, whereas misses were unrelated to relationship well-being (see Gleason, Iida, Bolger, & Shrout, 2003; Gleason, Iida, Shrout, & Bolger, 2008, for related results).

Coming back to CL, the quasi-signal detection method can answer important questions about the role of shared perspectives. Underwood (2009) proposed that CL can only be fully expressed and experienced when the donor accurately perceives both the recipient’s needs and which acts would be most helpful. Shared perspectives about a CL act should therefore contribute to relationship well-being, above and beyond the act itself, suggesting that CL would be beneficial primarily to the extent that one partner’s compassionate intent is perceived as such by the recipient (Reis, Clark, & Holmes, 2004). In other words, we hypothesized that hits should have more beneficial impact than misses or false alarms. An alternative possibility is that misses (also called invisible CL)—acts of CL enacted by one partner that the recipient does not explicitly notice—might be more effective than hits (also called visible CL)—acts of CL enacted by one partner that are explicitly recognized by the recipient. This possibility is suggested by research showing that visible social support is associated with increased affective distress, whereas invisible support is associated with decreased distress (e.g., Bolger & Amarel, 2007; Bolger, Zuckerman, & Kessler, 2000), presumably because of the self-esteem-threatening implications of receiving help (Bolger et al., 2000; Rafaeli & Gleason, 2009) and also because support recipients may feel indebted to donors (Clark & Mills, 2012). Although prior research has shown these effects only with regard to mood and not relationship outcomes, we felt it would be useful to evaluate this alternative hypothesis in the context of CL.

The present research

The present research was designed to examine everyday acts of CL and their impact on relationship well-being in a sample of newlywed heterosexual couples. Our first set of hypotheses examined how each partner’s acts of CL on one day predicted their own and their partner’s relationship outcomes on the next day. For the reasons described above, we expected that these acts would have a beneficial effect on both partners’ reports of relationship satisfaction. Because of the possibility that CL acts are simply examples of positive couple interactions, as well as of the absence of negative interactions, we used as control variables reports of positive acts (e.g., doing something to have fun together) and negative acts (e.g., being mean or critical) that in prior work have been shown to affect relationship satisfaction (e.g., Gable et al., 2003) but that are conceptually distinct from CL. Demonstrating that CL acts predict relationship well-being over and above positive and negative acts will help establish the construct validity of CL.

Our second set of analyses used the quasi-signal detection method to examine the relative influence of hits, misses, and false alarms on relationship satisfaction. Following the logic presented earlier, we reasoned that hits would be most influential. We also expected that false alarms would have some influence, consistent with evidence about the impact of perceived partner responsiveness (Reis et al., 2004).

Method

Participants

Our sample consisted of 175 heterosexual newlywed couples recruited throughout the US and Canada from various sources (63% bridal show attendees recruited via e-mail or postcard, 9% from Craigslist postings, 10% from ads posted on Facebook, and 18% from online forums). The research was advertised as a study of everyday behavior in early marriage. Couples were offered a financial incentive of up to US$100 (depending on the number of completed diaries) in addition to entry into a raffle for four additional US$100 prizes. Couples were also promised another US$100 and entry into additional raffles for completing later follow-up assessments, which are not relevant to the current article. Interested couples completed a screening survey that eliminated couples in which either or both spouses was over the age of 50 years; lived apart from each other; reported instances of domestic violence; had been hospitalized for emotional disorders; or were drug or alcohol abusers. We also spoke to both spouses on the telephone to establish their eligibility. Of 214 eligible couples who completed the screening procedure and returned consent forms, 175 participated in the 2-week daily diary study on which this article is based. Diary participants and nonparticipants did not differ significantly on marital satisfaction or CL (p > .3).

Couples had been married an average of 7.17 months (SD = 3.5 months, range = 1–16 months), at the beginning of the study. They had known each other an average of 6.2 years (SD = 3.6 years, range = 7 months–22 years). Participants were on average 28.2 years old (SD = 5.3, range = 18–50). In all, 18 couples (10.3%) had a child living with them and another 10 were pregnant or expecting a child (5.7%). In total, 75% of the participants were Caucasian (5% Hispanic), 7% were African American, 12% were Asian, and 6% were multiracial or other. In terms of education, spouses indicated their highest level of attainment: 31% of the sample had a postbaccalaureate degree, 49% were college graduates, 15% had some college, and 5% had a high school diploma. Seventy-six percent of participants worked full time, with the rest not working or employed part time. Median annual household income fell in the range between US$80,000 and US$89,999.

Procedure

All data were collected via the Internet. After completing an initial questionnaire that included demographic information (as well as other measures that are not relevant to this article), we scheduled couples to begin the 2-week daily diary component of the study. At exactly 7:00 p.m. each night of the daily diary period, all participants received an e-mail with a link to that day’s diary. The survey remained open until 9:00 a.m. the next morning, at which time the survey became inaccessible. Participants who missed more than one survey received reminder e-mails and phone calls. To encourage compliance, couples’ financial incentive depended on the number of surveys completed, with each couple receiving US$100 if both partners completed at least 12 of the 14 daily diaries, US$90 for completing 9–11 surveys each, US$70 for completing 6–8 surveys, and US$50 for completing fewer than 6 surveys. Participants also received a raffle ticket for each completed diary survey, which was entered into a drawing for one of two cash prizes of US$100. Participants completed an average of 13.2 out of the 14 daily diaries.

Development of the CL acts diary

We began by writing individual items corresponding to the definitions of CL offered by Fehr et al. (2009) and Underwood (2009), as described above. We then solicited feedback and suggestions from various colleagues, including other researchers funded by this Fetzer Institute initiative. Items were written to be clear and nonredundant, to refer to only a single action, to be accessible to nonpsychologically minded individuals and persons with only a high school education, and to describe behaviors likely to show variance across days and couples (i.e., we avoided writing items likely to show ceiling or floor effects). We then conducted two focus groups with a sample of nine young couples (interviewing husbands and wives separately). In these groups, we first asked participants to discuss how they saw the early years of marriage, emphasizing factors that they believed made their relationships work better and those that may have hindered their relationships. We then asked them to discuss the concepts of compassion and CL, emphasizing how it was expressed (or not expressed) in their marriage. In the final part of these focus groups, we showed participants the items proposed for the daily diary and asked them (a) what behaviors the item suggested, (b) to describe specific, concrete examples in their lives that would lead to a “did occur” or “did not occur” response, and (c) to give specific, concrete examples of behaviors enacted by themselves or their spouses that exemplified each item. These lively sessions were recorded and transcribed and used to clarify and revise our items. The final set of diary items is listed in Table 1.

Table 1.

Proportions of days compassionate behavior was reported as enacted or perceived.

	Wives		Husbands		Enacted/perceived main effect	Gender main effect	Interaction
	Enacted	Perceived	Enacted	Perceived	Enacted/perceived main effect	Gender main effect	Interaction
Compassionate acts composite	.59_a	.60_a	.65_b	.59_a	19.89***	1.33	27.71***
I willingly put my partner's goals or wishes ahead of my own	.53_a	.56_a	.65_b	.53_a	14.43***	4.06*	33.85***
I said or did something to show that I value my partner	.78_a	.76_a	.80_b	.75_b	14.64***	0.002	4.12*
I really tried to understand my partner’s thoughts and feelings	.67_a	.60_a	.71_b	.58_b	97.65***	0.49	8.84**
I expressed a lot of tenderness and caring for my partner	.71_a	.73_a	.69_a	.68_a	0.46	3.99*	1.60
I went out of my way to “be there” for my partner	.38_a	.45_a	.49_b	.43_b	0.01	3.67^†	33.60***
I really tried to be accepting rather than judging of something about my partner	.58_a	.57_a	.64_b	.59_a	11.66***	1.83	5.20*
I willingly modified my plans or activities for my partner’s sake	.46_a	.50_a	.58_b	.46_a	12.45***	3.33^†	41.84***
I voluntarily did something special for my partner	.50_a	.55_a	.56_a	.55_b	3.93*	2.52	5.97*
I did something to show my partner that I respect and admire him/her as a person	.60_a	.60_a	.62_a	.63_a	0.001	0.83	0.33
I was genuinely open and receptive to things my partner said or asked of me	.73_a	.70_b	.75_b	.69_b	31.85***	0.16	4.27*

Note. Numbers represent the proportion of days during which the behavior was endorsed. Listed items refer to each individual’s report of behavior he or she “enacted” and began with the word “Today.” The “perceived” version of each item refers to perceptions of the partner’s behavior (e.g., “My partner willingly put my goals or wishes ahead of his/her own”). All F tests have 1, 174 degrees of freedom. Subscripts show the results of simple effects analyses comparing each participant’s reports of enacted behavior to the other partner’s perception of that behavior (different subscripts indicate a significant difference).

^† p < .10; *p < .05; **p < .01; ***p < .001.

Measures

Daily acts

In addition to the 10 dichotomous items assessing daily CL acts displayed in Table 1, participants also completed items assessing more global positive (but not CL) behavior and negative behavior. In reporting whether or not each behavior occurred, participants were instructed to “choose ‘yes’ only if you can recall a specific instance that occurred today that matches the description.” Three items assessed general positive behavior, “Today, I did something to have fun with my partner”; “Today, I did something so my partner and I would enjoy spending time together”; “Today, I showed physical affection to my partner.” Four items assessed general negative behavior, “Today, I was mean to my partner”; “Today, I was inattentive or insensitive to my partner”; “Today, I did something that made (or might make) my partner worry about our relationship”; “Today, I was moody or critical with my partner.” These 17 items were intermingled on a single page of the online diary survey. As described previously, each act included parallel items asking whether or not the participant’s partner had enacted the behavior. These 17 items assessing perceived partner behavior were intermingled on a separate page of the survey. Each set of items was averaged within each day, yielding six composites representing both self-reported and perceived partner reports of CL, positive acts, and negative acts.

Daily relationship satisfaction

We created a single composite measure by summing three questions assessed daily. The first asked, “Today our relationship was … terrible/terrific,” and participants indicated their responses on a 7-point scale anchored by terrible and terrific on opposing ends. The other two items were, “Today, I felt close and connected to my partner” and “Today, I enjoyed our time together,” and were rated on 1 (not at all) to 7 (a great deal) scales. Cronbach’s α computed separately for each day ranged from .83 to .92 for wives (M = .88) and from .86 to .92 for husbands (M = .88).

Results

We began by examining the relative prevalence of CL acts during the diary record keeping period. In order to determine whether a single composite composed of the 10 CL acts would be justified, we conducted a series of factor analyses (using principal axis factoring with oblique rotation and extracting factors with eigenvalues greater than one). For these analyses, we computed the proportion of days on which a given behavior occurred, separately for husbands and wives and for self-reported and perceived partner behavior. In each analysis, a dominant single factor emerged, as indicated by a scree plot, on which all 10 items loaded. This primary factor accounted for between 53% and 67% of the variance (eigenvalues of 5.68 or greater). In two out of these four analyses, a second factor appeared, accounting for less than 8% of the variance (initial eigenvalues were 1.11 or smaller). In both instances, the items that defined this factor also cross loaded on the dominant factor. We therefore decided to retain only a single composite CL factor that included all 10 items. For own actions, Cronbach’s α (computed separately for each day and then averaged) was .82 for wives and .85 for husbands; for perceived partner actions, α averaged .87 for wives and .89 for husbands.

Table 1 presents means for the proportion of days on which CL acts were enacted or perceived, both for the CL composite and the various individual actions. A repeated measures analysis of variance (ANOVA) on the CL composite indicated that spouses reported enacting CL more often than they perceived it, a result that was qualified by a significant interaction with gender. The difference between husbands and wives was not significant. Means for this analysis are shown in the top row of Table 1. We used simple effects tests to compare each spouse’s report of the behaviors enacted to the other spouse’s report of the behaviors received. These analyses showed that husbands reported enacting significantly more CL than their wives perceived, F(1, 174) = 25.32, p < .001, whereas the difference between wives’ reports of the behaviors they enacted and husbands’ reports of the behaviors they received did not differ significantly, F(1, 174) = .54, p = .46.

Table 1 also displays findings from comparable analyses for the 10 individual acts. In general, these analyses supported results from the CL composite. For 6 of the 10 acts, spouses reported enacting the behavior significantly more often than their partners perceived receiving them (the largest effect occurred for “tried to understand my partner’s thoughts and feelings,” which may reflect the more cognitive focus of this item.) The sex main effect was significant in only two instances, and in both cases husbands’ reports were larger than wives’ reports. More importantly, the enacted–perceived × sex interaction was significant for 8 of 10 behaviors. Simple effect tests revealed that for all but one of these variables, husbands reported enacting the behavior significantly more often than their wives reported receiving them. On the other hand, wives reported enacting behaviors significantly more often than their husbands reported receiving them in only three instances. For two behaviors, wives reported enacting that behavior significantly less often than their husbands reported receiving it. In short, these analyses indicate that whereas husbands reported enacting CL more often than their wives reported receiving those actions, wives’ reports of the behavior they enacted generally did not differ from what their husbands reported receiving.

Table 2 provides correlations among these across-day aggregates. In general, spouses who enacted more CL across the diary record keeping period also reported more positive acts and higher average daily satisfaction. Interestingly, aggregated negative acts were positively albeit weakly correlated with CL acts among wives, suggesting that engaging in more negativity can co-occur over time with more compassionate acts.¹

Table 2.

Correlations among daily measures (aggregated across days).

Variable	1	2	3	4	5	6	7
1. Compassionate love acts	.47***	.58***	.18*	.84***	.57***	.04	.42***
2. Positive acts	.71***	.46***	.04	.56***	.92***	−.17*	.56***
3. Negative acts	.05	−.10	.58***	.18*	.09	.71***	−.22**
4. Perceived partner compassionate love acts	.90***	.69***	.03	.39***	.65***	−.10	.54***
5. Perceived partner positive acts	.67***	.92***	−.05	.72***	.54***	−.16*	.60***
6. Perceived partner negative acts	.17*	.03	.73***	.01	−.003	.50***	−.41***
7. Daily relationship satisfaction	.50***	.57***	−.32***	.59***	.59***	−.30***	.49***

Note. Correlations for wives are presented above the diagonal; correlations for husbands are presented below the diagonal. The numbers in boldface on the diagonal indicate the correlation between husbands and wives.

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

Do CL acts predict relationship satisfaction?

To evaluate our hypothesis that CL acts would be associated with increases in relationship satisfaction, we conducted actor–partner interdependence modeling within a multilevel model (Kenny et al., 2006). This analysis controls for within-person and within-couple dependencies, while at the same time allowing us to compare actor effects (effect of one’s actions on one’s own satisfaction) and partner effects (effect of a partner’s actions on one’s satisfaction). We used multilevel modeling, with a two-level model for matched pairs in which effects for husbands and wives were assessed simultaneously at level 1 (Bolger & Laurenceau, 2013; Raudenbush, Brennan, & Barnett, 1995). All predictors were person-mean centered prior to analysis. Intercepts were allowed to vary randomly for both husbands and wives; other effects were estimated as fixed effects. All analyses also included two types of control variables. First, all analyses controlled for relationship satisfaction on the prior day (including both actor and partner effects). Thus, the analyses examine the degree to which CL acts predict changes in relationship satisfaction from one day to the next. Second, as recommended by Bolger and Laurenceau (2013), we controlled for linear changes in satisfaction over the 14 days of the diary for both males and females.

In the first of these analyses, we used composite self-reported CL scores to predict daily relationship satisfaction (controlling for relationship satisfaction on the prior day). This analysis also included indices of the frequency of daily positive and daily negative interactions. We included these control variables because we wanted to identify the effect of CL acts over and above the general affective valence of interaction. (As shown in Table 3, compassionate acts occurred more often on days in which positive interaction was more common and negative interaction was less common.) Thus, these analyses examine the effect of CL acts over and above any effects that might be due to having interacted in a positive or negative manner on that day. Table 4 reports all coefficients, standard errors, and p values for this analysis. A simplified representation of the key results is depicted in Figure 1. The top section of this figure shows that both actor and both partner effects were significant. Husbands’ and wives’ reports of their CL activities predicted changes in their own relationship satisfaction, Bs = .86 and .84, ts(3, 833) = 10.46 and 10.15, ps < .001, respectively, and in their partners’ relationship satisfaction, Bs = .44 and.50, ts(3, 833) = 5.35 and 6.11, ps < .001, respectively.

Figure 1.

Predicting daily satisfaction from own and perceived partner compassionate acts. *p < .05; **p < .01; ***p < .001. (a) Based on participants’ reports of their own behavior. (b) Based on perceived partner behaviors. Control variables are omitted from the figure for simplicity.

Table 3.

Predicting daily compassionate acts from daily positive and negative acts.

	Compassionate acts
	B	SE	p
Positive acts
Husband acts → own compassionate love acts	.35	.02	<.001
Wife acts → own compassionate love acts	.31	.02	<.001
Husband acts → wife’s compassionate love acts	.05	.02	.004
Wife acts → husband’s compassionate love acts	.01	.02	.585
Negative acts
Husband acts → own compassionate love acts	−.08	.02	.002
Wife acts → own compassionate love acts	−.21	.02	<.001
Husband acts → wife’s compassionate love acts	.004	.02	.861
Wife acts → husband’s compassionate love acts	−.04	.02	.101
Linear trend of time
In husbands	.001	.001	.518
In wives	.001	.001	.366

Table 4.

Predicting daily relationship satisfaction from daily behavior.

	Daily relationship satisfaction
	Self-reported behavior			Perceived partner behavior
	B	SE	p	B	SE	p
Compassionate love acts
Husband reports → own daily satisfaction	.86	.08	<.001	1.09	.07	<.001
Wife reports → own daily satisfaction	.84	.08	<.001	1.10	.08	<.001
Husband reports → wife’s daily satisfaction	.44	.08	<.001	.26	.07	.001
Wife reports → husband’s daily satisfaction	.50	.08	<.001	.18	.08	.017
Control variables
Positive acts
Husband reports → own daily satisfaction	.67	.07	<.001	.72	.06	<.001
Wife reports → own daily satisfaction	.84	.06	<.001	.77	.06	<.001
Husband reports → wife’s daily satisfaction	.30	.07	<.001	.36	.06	<.001
Wife reports → husband’s daily satisfaction	.40	.06	<.001	.37	.06	<.001
Negative acts
Husband reports → own daily satisfaction	−1.29	.09	<.001	−1.01	.09	<.001
Wife reports → own daily satisfaction	−1.05	.09	<.001	−1.15	.09	<.001
Husband reports → wife’s daily satisfaction	−.75	.09	<.001	−.49	.09	<.001
Wife reports → husband’s daily satisfaction	−.44	.09	<.001	−.43	.09	<.001
Yesterday’s relationship satisfaction
Husband’s prior sat → own daily satisfaction	.04	.02	.049	.04	.02	.062
Wife’s prior sat → own daily satisfaction	.08	.02	<.001	.07	.02	<.001
Husband’s prior sat → wife’s daily satisfaction	.03	.02	.156	.02	.02	.397
Wife’s prior sat → husband’s daily satisfaction	.02	.02	.300	.02	.02	.313
Linear trend of time
In husbands	.00	.004	.996	.002	.004	.557
In wives	.01	.004	.085	.005	.004	.190

To determine whether the actor effects differed by gender, we constrained these two paths to be equal and compared the deviance statistics for the two models using full maximum likelihood estimation. (All model comparisons reported in this article used this procedure.) This did not produce a significant drop in model fit, χ²(1) = .02, p = .88, indicating that the husband and wife actor effects do not differ significantly. Similarly, when we constrained the two partner effects to be equal, the model fit again was not significantly worse, χ²(1) = .34, p = .56, indicating that the husband and wife partner effects do not differ significantly. Of somewhat greater theoretical interest is the result of constraining the actor and partner effects to be equal. In this case, model fit was significantly worse, χ²(2) = 18.57, p < .001, indicating that satisfaction was more strongly associated with one’s own self-reported compassionate behavior than with the behavior reported by one’s partner.

The lower section of Figure 1 (and the second set of coefficients displayed in Table 4) displays results for similar analyses of perceived partner CL acts. Both actor and both partner effects were again significant. Husbands’ and wives’ reports of their partner’s CL activities predicted changes in their own relationship satisfaction, Bs = 1.09 and 1.10, ts(3, 833) = 14.74 and 14.36, ps < .001, respectively, and in their partners’ relationship satisfaction, Bs = .26 and .18, ts(3, 833) = 3.49 and 2.38, ps ≤ .001 and .017, respectively. Constraining the actor effects to be equal for husbands and wives did not produce a significant drop in model fit, χ²(1) = .002, p = .97, indicating that the husband and wife actor effects do not differ significantly. Similarly, constraining the partner effects to be equal across husbands and wives did not significantly reduce model fit, χ²(1) = .52, p = .47, indicating that the husband and wife partner effects do not differ significantly. Constraining the actor and partner effects to be equal did produce a significant drop in model fit, χ²(2) = 117.98, p < .001, indicating that perceived partner CL had a significantly larger effect on own relationship satisfaction than on the partner’s satisfaction.

Quasi-signal detection analysis

After establishing that self-reported and perceived partner CL predicted change in daily satisfaction, we conducted further analyses to examine the importance of agreement or disagreement across spouses in these effects. This analysis was designed to determine the relative frequency of hits, misses, false alarms, and correct rejections in couples’ everyday interaction, as well as the impact of each of these categories on relationship satisfaction. Table 5 provides a definition of each of these effects for both actor and partner effects. We first examined the relative frequency of each category by calculating separately for each behavior on each day whether it was a hit, miss, false alarm, or correct rejection. One such analysis was performed for behaviors enacted (or not) by husbands and another for behaviors enacted (or not) by wives. The relative proportions of each of these categories across the 14-day record keeping period are reported in Table 6. As in the earlier analysis, a composite was also created by averaging across the 10 behaviors within each signal detection category.

Table 5.

Description of effects included in quasi-signal detection analysis using APIM.

Type of action	Focus of effect	Description
Hits	Actor	Impact of behaviors enacted by Jack and perceived by Jill on Jack’s satisfaction.
Hits	Partner	Impact of behaviors enacted by Jack and perceived by Jill on Jill’s satisfaction (visible compassionate love).
Misses	Actor	Impact of behaviors enacted by Jack but not perceived by Jill on Jack’s satisfaction.
Misses	Partner	Impact of behaviors enacted by Jack but not perceived by Jill on Jill’s satisfaction (invisible compassionate love).
False alarms	Actor	Impact of behaviors not enacted by Jack but nevertheless perceived by Jill on Jack’s satisfaction.
False alarms	Partner	Impact of behaviors not enacted by Jack but nevertheless perceived by Jill on Jill’s satisfaction.
Correct rejections	Actor	Impact of behaviors not enacted by Jack and not perceived by Jill on Jack’s satisfaction.
	Partner	Impact of behaviors not enacted by Jack and not perceived by Jill on Jill’s satisfaction.

APIM: actor–partner interdependence model.

Table 6.

Proportion of days classified as hits, misses, and false alarms using quasi-signal detection.

	Wives’/husbands’ enacted behavior				Hit/miss/false alarm main effect	Gender main effect	Interaction
	Hit	Miss	False alarm	Correct rejection	Hit/miss/false alarm main effect	Gender main effect	Interaction
Compassionate acts composite	.42/.46	.17/.19	.17/.15	.24/.20	172.46***	23.72***	2.62^†
Put partner’s goals first	.36/.43	.17/.22	.17/.13	.30/.22	88.16***	27.28***	6.82**
Show I value partner	.62/.64	.16/.16	.13/.13	.09/.07	311.54***	5.93*	.28
Tried to understand partner	.43/.47	.23/.24	.15/.12	.19/.17	110.68***	3.86†	1.77
Expressed tenderness	.56/.58	.16/.12	.13/.16	.15/.14	257.47***	.39	3.26*
Tried to “be there” for partner	.23/.30	.15/.20	.20/.16	.42/.34	17.16***	27.41***	4.80**
Fully accepting of partner	.38/.40	.20/.24	.21/.17	.21/.19	42.77***	5.44*	1.93
Modified my own plans	.29/.35	.18/.22	.17/.15	.36/.28	54.47***	36.65***	4.04*
Did something special	.35/.38	.15/.18	.20/.17	.30/.27	68.16***	5.01*	2.69^†
Show I respect and admire	.44/.43	.17/.19	.20/.17	.19/.21	66.87***	.33	.90
Open and receptive	.54/.57	.18/.18	.15/.13	.13/.12	192.47***	.55	1.23

Note. The hit/miss/false alarm main effect assesses whether these three categories differ significantly from one another.

*p < .05; **p < .01; ***p < .01; ^† p < .10.

A 2 (husband/wife) × 3 (hit/miss/false alarm) repeated measures ANOVA revealed a significant main effect of type of report, F(2, 348) = 172.46, p < .001.² A post hoc Bonferroni test indicated that all three types of reports differed significantly from each other, such that hits were reported significantly more often than both misses and false alarms (ps < .001), and misses were reported significantly more often than false alarms (p < .001). There was also a significant gender effect, such that husbands’ CL behavior tended to be reported more frequently than wives’ behavior, F(1, 174) = 23.72, p < .001. However, this main effect was qualified by a marginal interaction, F(2, 348) = 2.62, p = .074. Simple effects analyses revealed that hits were more common for husbands’ behavior than for wives’ behavior, F(1, 348) = 3.75, p = .05, but misses or false alarms did not differ, Fs(1, 348) < 1.43, ps > .23. Table 6 also displays the results of similar analyses for the 10 individual acts.

Although the gender differences showing that hits were more frequently reported for husbands than for wives may suggest that husbands’ behavior was viewed more accurately than wives’ behavior, this conclusion is misleading, as these analyses excluded the “correct rejection” category. To examine gender differences in accuracy, we created a new variable that was the sum of “hits” and “correct rejections” (indicating both forms of agreement about a single partner’s behavior). This measure of agreement did not differ significantly across husbands and wives on the CL composite, t(174) = .25, p = .80, or for any individual behavior, all ts(173) < 1.87, all ps > .06, indicating similar levels of agreement in reports of husbands’ and wives’ behavior.

APIM analyses

Next, we conducted APIM analyses to examine the relative impact of each of these categories on relationship satisfaction. To avoid capitalizing on chance, this analysis was conducted only on the composite that averages across all 10 behaviors. As before, we relied on multilevel modeling, using a two-level model for matched pairs in which effects for husbands and wives are assessed simultaneously at level 1 (Bolger & Laurenceau, 2013; Raudenbush et al., 1995). All predictors were person-mean centered prior to analysis and all analyses again included prior day relationship satisfaction (actor and partner effects) and the linear effect of time as control variables. Thus, the analyses describe the degree to which hits, misses, and false alarms in compassionate acts predict changes in relationship satisfaction from one day to the next.

Quasi-signal detection analyses evaluate the impact of each category relative to a reference standard, correct rejections (Finkenauer et al., 2010; Gable et al., 2003). In other words, each effect represents the relative impact of a particular kind of event occurrence—a hit, miss, or false alarm—compared with days on which both partners report no event (correct rejections). Results of this analysis are detailed in Table 7 and depicted in a simplified manner, omitting control variables in Figure 2. Both actor and partner effects for hits were significant for both spouses, indicating that agreed-upon acts of CL were associated with significant increases in relationship satisfaction from the prior day for both partners. Both actor effects for misses were also significant, indicating that spouses’ own relationship satisfaction was higher on days when they behaved compassionately, even if their partner did not notice the act. In contrast, the partner effect for misses was significant only for husbands. This means that husbands’ relationship satisfaction improved when their wives behaved compassionately, even if they did not notice that act. This was not the case for wives, however, who did not show significant benefits from acts their husbands reported doing that they did not notice. Both actor and partner effects for false alarms were significant (marginally so, in one instance). These effects mean that both partners benefit when one perceives the other to have behaved compassionately, even if the provider did not report enacting that behavior.³

Figure 2.

Quasi-signal detection analysis predicting daily satisfaction from compassionate acts. *p < .05; **p < .01; ***p < .001. Control variables are omitted from the figure for simplicity.

Table 7.

Predicting daily relationship satisfaction from hits, misses, and false alarms.

	Daily relationship satisfaction
	B	SE	p
Daily compassionate acts
Hits
Husband acts → own satisfaction	1.02	.11	<.001
Wife acts → own satisfaction	1.39	.10	<.001
Husband acts → wife’s satisfaction	1.67	.11	<.001
Wife acts → husband’s satisfaction	1.92	.10	<.001
Misses
Husband acts → own satisfaction	.70	.13	<.001
Wife acts → own satisfaction	1.17	.14	<.001
Husband acts → wife’s satisfaction	.16	.13	.223
Wife acts → husband’s satisfaction	.59	.14	<.001
False alarms
Husband acts → own satisfaction	.41	.14	.005
Wife acts → own satisfaction	.81	.13	<.001
Husband acts → wife’s satisfaction	1.44	.14	<.001
Wife acts → husband’s satisfaction	1.50	.13	<.001
Control variables
Yesterday’s relationship satisfaction
Husband’s prior sat → own satisfaction	.03	.02	.145
Wife’s prior sat → own satisfaction	.06	.02	.003
Husband’s prior sat → wife’s satisfaction	.03	.02	.221
Wife’s prior sat → husband’s satisfaction	.03	.02	.135
Linear trend of time
In husbands	.00	.004	.862
In wives	.003	.004	.418

In order to further explore these effects, we conducted focused contrasts by constraining pairs of paths to be equal and then determining whether model fit was significantly reduced. The first set of tests compared the actor effect and the corresponding partner effect for each spouses’ satisfaction (in other words, comparing effects on the same horizontal line in Figure 2, separately for each partner’s satisfaction). For hits, this test asks whether a spouse’s satisfaction is more influenced by acts they perform (the actor effect) or acts their partner performs (the partner effect). Constraining these paths to be equal produced a significant decrease in model fit for husbands’ satisfaction, χ²s(1) = 23.38, p < .001, but not for wives’ satisfaction, χ²(1) = 2.28, p = .13. Thus, for husbands but not wives, acts performed by their partners had greater impact on their satisfaction than their own actions.

For misses, constraining the actor and partner effects to be equal led to a significant decrease in model fit for wives’ satisfaction, χ²(1) = 24.64, p < .001, but not for husbands’ satisfaction, χ²(1) = .31, p = .58. This means that wives’ satisfaction was more strongly influenced by compassionate acts that they performed that their husbands did not see than by compassionate acts that their husbands performed that they did not see. For husbands, however, these two types of misses had the same relative effect.

For false alarms, constraining the actor and partner effects to be equal led to a significant decrease in model fit for both husbands’ satisfaction, χ²(1) = 27.91, p < .001, and wives’ satisfaction, χ²(1) = 9.13, p = .003. In this case, partner effects were greater than actor effects, indicating that spouses benefitted more from actions they perceived that their spouse did not report enacting than from actions they did not report enacting that their partners nevertheless perceived.

We also conducted three additional contrasts that were of theoretical interest. The first test constrained the partner effects for hits and misses to be equal. This test compares the effect of visible and invisible acts of CL on the recipient. For both husbands and wives, this constraint led to significantly poorer model fit, χ²s(1) = 124.77 and 182.04, respectively, both ps < .001. As hypothesized, visible acts were more influential than invisible acts.

The second test compared actor effects for hits and misses. This test asks whether the impact of acts one performs depends on whether one’s partner has noticed that act. When these two paths were constrained to be equal, the model fit was significantly worse for husbands’ satisfaction, χ²(1) = 8.29, p = .004, and marginally worse for wives’ satisfaction, χ²(1) = 3.44, p = .06. In both cases, greater benefit was derived when one’s compassionate acts were noticed.

Finally, we compared the partner effects for hits and false alarms. This comparison asks whether “shared reality” has the same effect as the perception of having been treated compassionately. This constraint led to a significant decrease in model fit for husbands’ satisfaction, χ²(1) = 11.82, p = .001, and a marginal decrease for wives’ satisfaction, χ²(1) = 3.14, p = .073. In both instances, hits had a stronger effect than false alarms, suggesting a beneficial role for shared perceptions of an action.

Discussion

In the most general sense, these results demonstrate that the enactment and perception of acts of CL are associated with feelings of relationship satisfaction in everyday life. Across the 10 different acts included in our composite, newlywed spouses reported enacting these behaviors on about 60% of days. It would be interesting to see whether this figure declines over time, as has been shown with many other indicators of marital quality and satisfaction (e.g., Kurdek, 1998, 1999). We did not expect that husbands would report enacting CL more often than wives did, nor did we expect that the difference between reports of enacted and perceived behavior would be significant for husbands’ actions but not for wives’ actions. These differences may mean that husbands’ reports of their own behavior are more strongly influenced than wives’ reports by motivated construal—that is, by self-regulatory concerns to see their own behavior in a positive light. Some prior research indicates that wives have higher levels of empathic accuracy than husbands do, although other studies find little or no sex difference (see Graham & Ickes, 1997, for a review). This result may also indicate that husbands actually do enact CL more often than wives do, but in ways that may be harder for their wives to recognize or acknowledge.

Of somewhat greater theoretical relevance, as shown in APIM analyses, reports of enacted CL predicted increases in relationship satisfaction equally for both sexes. Because these analyses controlled for daily positive and negative acts, the observed boosts in daily relationship quality represent a type of interaction that goes beyond simple changes in the valence of couple socializing—although, to be sure, CL was associated to some extent with the presence of positive interaction and the absence of negative interaction. Notably, both actor and partner effects were significant for reports of enacted behavior—that is, both the provider and recipient benefit when they believe that CL has been enacted, a result that supports earlier theorizing about the mutual benefits of CL (Underwood, 2009). The fact that participants benefitted from providing CL is consistent with results from studies of social support (e.g., Brown, Nesse, Vinokur, & Smith, 2003) and may reflect the increased sense of meaning and belonging that such acts entail. Alternatively, these benefits may stem from anticipated or actual gratitude by their partners (Algoe, Haidt, & Gable, 2008).

Relatedly, and not surprisingly, both actor effects were significant for perceived acts of CL—relationship satisfaction is higher when spouses believe that their partner has been compassionate. Furthermore, the corresponding partner effects were also significant among both husbands and wives—in other words, spouses benefited when their partners perceived that they had behaved compassionately, supporting a potential role for gratitude, as mentioned above. Because these effects were modeled within dyads, they do not simply demonstrate the benefit of having a compassionate partner. Rather, these results show how compassionate behavior predicts shifts in satisfaction from day to day. Crocker and Canevello (2008) found a similar result for college roommates whose goals reflected compassionate caring for each other (although they did not examine CL acts directly).

It is likely that these results for perceived partner CL reflect the influence of both projection and reality. Lemay and Clark (2008) have shown that people project their own responsiveness onto partners, perceiving those partners to be just as responsive as they themselves are, and several findings from our research suggest that this occurs for CL as well—for example, the high correlation between self-rated compassionate acts and perceptions of the partner’s CL acts (Table 2). On the other hand, our results also clearly indicate that each partner’s report of perceived partner compassionate behavior was to a large extent influenced by the other partner’s report of enacted compassion. Together, then, these two results indicate that CL is influenced by actual interaction and projection, similar to Reis et al.’s (2004) conclusion that both sets of processes contribute to responsiveness (see Fletcher & Kerr, 2010, for a similar conclusion). In future research, it will be important to determine the motives, conditions, and individual differences for which projection and accuracy play relatively larger or smaller roles.

These effects are clarified by the results of our quasi-signal detection analyses. Partner effects were significantly stronger than actor effects for hits (among husbands) and for false alarms (among husbands and wives), indicating that recipients of CL acts may benefit relatively more than providers do. This result is also interesting because partner effects are typically weaker in relationship research than actor effects (Kenny & Ledermann, 2010), although a growing literature shows how partners can facilitate (or hinder) each other’s emotional self-regulation (see Overall & Simpson, 2013, for a review). It may be that the other-focused nature of CL heightens the relative benefit for partners compared to actors. In contrast, for misses, the actor effects were stronger than the partner effects (albeit nonsignificantly so among husbands), a result that is consistent with the idea that recognition of an action contributes to its impact, as discussed next.

A key element of our theorizing concerned the importance of mutual recognition of CL acts. This theorizing was supported in several respects. First, hits, which were more common than either misses or false alarms, were strongly predictive of relationship satisfaction. Second, hits had stronger effects on relationship satisfaction than misses did, both for partner and for actor effects. Third, hits had stronger effects than false alarms, indicating that even when the recipient’s perception of a compassionate act is the same, those acts that were also described as enacted were more influential than those that were not so described. Thus, as has previously been shown with regard to processes such as intimacy, trust, and responsiveness (see Reis & Clark, 2013, for a review), relationships benefit more when both partners recognize their compassionate actions than when they do not. Nevertheless, we also found evidence for the impact of less visible acts of CL, consistent with research on invisible support (Bolger et al., 2000). For example, the partner effect for misses on husbands’ satisfaction was significant, indicating that husbands benefited from their wives’ reported actions, even when unaware of them. Perhaps such actions are sometimes too subtle to be noticed but still have beneficial effects outside of awareness. Wives, in contrast, may be more likely to notice such actions.

Taken together, these results contribute to the existing literature on CL in three general respects. First, acts of CL are directly linked to relationship satisfaction for both providers and recipients. Whereas previous research has emphasized feelings and cognitions, these findings provide the first evidence about the role of compassionate behavior in newlyweds and its beneficial impact on relationships. Second, CL acts have greater influence on relationship well-being when both partners acknowledge their occurrence, rather than when those actions are evident only from one partner’s vantage point. This finding supports theorizing by Underwood (2009) and others about the importance of mutual recognition. More generally, as the first study of CL to adopt actor–partner interdependence modeling, our work corroborates the value—indeed, the necessity—of a fully dyadic perspective in relationship process research.

A third general contribution concerns sex differences, which have not been studied extensively by researchers interested in CL. Although for the most part our findings for husbands and wives were similar, several noteworthy differences appeared. Husbands reported higher levels of compassionate acts relative to wives’ reports of these behaviors or of their own actions.⁴ Also, wives’ compassionate acts were beneficial for husbands even when husbands did not explicitly notice them. Our data are silent when it comes to interpreting these effects. Perhaps husbands are more focused on their own actions, seeing more CL in their own actions and overlooking some of their wives’ acts. Another possibility is that newlywed husbands go out of their way to be compassionate more often than wives do, but wives are better at recognizing these actions. Future research is needed to examine these and other possible explanations.

Limitations and future directions

Several important limitations to our work should be noted. First, we did not obtain independent corroboration of these behaviors, rendering ambiguous interpretations of perspective-based differences. That is, when husbands and wives disagreed, we could not determine which perspective was the more accurate one. Laboratory observations will be especially helpful in this regard, once appropriate situations for the appearance of CL have been identified. Laboratory observation would also help determine whether agreement depends on the strength of the act. It is possible that agreement (as in the case of hits) is more common when an act is strongly and unambiguously compassionate, whereas disagreement (as in the case of misses and false alarms) follows from ambiguously compassionate acts. Furthermore, because our method asks participants to recall a specific event but not to describe that event, it is possible that spouses were thinking about different events from a given day. Future research is needed to examine these various possibilities.

A second limitation concerns the particular compassionate acts that we chose to study. We were to some extent limited by the lack of existing literature about how compassionate feelings are expressed in everyday behavior. Although the procedures we used to generate the CL acts diary tapped into both theory and lay perspectives, it is possible that we overlooked important alternatives. Moreover, individuals may differ systematically in their ways of expressing CL, and it may be useful to directly examine these differences.

Another limitation is the nature of our sample: relatively educated, affluent, and tech-savvy newlyweds. It will be important in future research to determine whether CL is similarly prevalent and influential in less well-to-do samples. Inasmuch as relationship quality tends to be somewhat lower on average among such couples (Karney & Bradbury, 1995), it is possible that differences in CL may be observed.

As for future directions, we think it will be critical to examine CL after the first year of marriage. Newlyweds are still thick in the bloom of new love; it is plausible that as couples adapt to everyday interdependence, the frequency of thoughtful, compassionate acts toward the other may decrease. This decrease may be responsible for the well-established decrease in relationship satisfaction that occurs in the early years of marriage (Kurdek, 1998, 1999), an important possibility that merits investigation. Also needed are studies comparing daily acts of CL with daily acts of passionate and companionate love. Although theorized to be distinct (Berscheid, 2010; Fehr et al., 2009), it remains to be established just how these varied forms of love are differentially expressed in everyday life.

Conclusion

As relationship science approaches a more mature stage of development, new concepts tend to be identified with decreasing frequency. The recent emergence of CL as a research topic is therefore exciting, providing the field with a stimulus toward new findings and explanatory constructs. Our results indicate that compassionate actions are a beneficial component of everyday marital interaction among newlyweds. We hope this report stimulates more detailed investigations about the manner in which CL is experienced and expressed and about its consequences not just for new marriages but for all human relationships.

Footnotes

Authors’ note

We thank the Fetzer Institute for its generous support of this research. We also thank Wayne Ramsey for graciously facilitating our work, Marc McIntosh for assistance in recruiting our sample, David de Jong, Stephanie O’Keefe, and Christine Walsh for help in maintaining contact with our couples, and Sue Sprecher for her helpful comments on an earlier draft.

Funding

This research received support from the Fetzer Institute.

Notes

References

Algoe

S. B.

Haidt

Gable

S. L.

(2008). Beyond reciprocity: Gratitude and relationships in everyday life. Emotion, 8, 425–429.

Aron

Davies

K. M.

(2009). Aristotle and Plato on relationships. In Reis

H. T.

Sprecher

(Eds.), Encyclopedia of Human Relationships (Vol. 1, pp. 108–110). Thousand Oaks, CA: Sage.

Asch

S. E.

(1952). Social psychology. New York, NY: Prentice Hall.

Baldwin

M. W.

(1992). Relational schemas and the processing of social information. Psychological Bulletin, 112, 461–484.

Baxter

L. A.

Bullis

(1986). Turning points in developing romantic relationships. Human Communication Research, 12, 469–493.

Berscheid

(2010). Love in the fourth dimension. Annual Review of Psychology, 61, 1–25.

Bolger

Amarel

. (2007). Effects of social support visibility on adjustment to stress: Experimental evidence. Journal of Personality and Social Psychology, 92, 458–475.

Bolger

Davis

Rafaeli

(2003). Diary methods: Capturing life as it is lived. Annual Review of Psychology, 54, 579–616.

Bolger

Laurenceau

J.-P.

(2013). Intensive longitudinal designs. New York, NY: Guilford Press.

10.

Bolger

Zuckerman

Kessler

R. C.

(2000). Invisible support and adjustment to stress. Journal of Personality and Social Psychology, 79, 953–961.

11.

Bradbury

T. N.

Fincham

F. D.

Beach

S. R.

(2004). Research on the nature and determinants of marital satisfaction: A decade in review. Journal of Marriage and Family, 62, 964–980.

12.

Brown

S. L.

Nesse

R. M.

Vinokur

A. D.

Smith

D. M.

(2003). Providing social support may be more beneficial than receiving it results from a prospective study of mortality. Psychological Science, 14, 320–327.

13.

Buss

D. M.

Craik

K. H.

(1983). The act frequency approach to personality. Psychological Review, 90, 105–126.

14.

Clark

M. S.

Lemay

Jr., E. P.

Graham

S. M.

Pataki

S. P.

Finkel

E. J.

(2010). Ways of giving benefits in marriage: Norm use, relationship satisfaction, and attachment-related variability. Psychological Science, 21, 944–951.

15.

Clark

M. S.

Mills

(2012). Communal (and exchange) relationships. In Van Lange

P. A. M.

Kruglanski

A. W.

Higgins

E. T.

(Eds.), Handbook of theories of social psychology (pp. 232–250). Thousand Oaks, CA: Sage Publications.

16.

Coontz

(2005). Marriage: A history. New York, NY: Penguin.

17.

Crocker

Canevello

(2008). Creating and undermining social support in communal relationships: The role of compassionate and self-image goals. Journal of Personality and Social Psychology, 95, 555–575.

18.

Fehr

Sprecher

(2009). Compassionate love: Conceptual, measurement, and relational issues. In Fehr

Sprecher

Underwood

L. G.

(Eds.), The science of compassionate love (pp. 27–52). Malden, MA: Wiley-Blackwell.

19.

Fehr

Sprecher

Underwood

L. G.

(Eds.) (2009). The science of compassionate love. Malden, MA: Wiley-Blackwell.

20.

Finkenauer

Righetti

(2011). Understanding in close relationships: An interpersonal approach. In Stroebe

Hewstone

(Eds.), European review of social psychology (pp. 316–363). Brighton, UK: Psychology Press.

21.

Finkenauer

Wijngaards-De

Meij

Reis

H. T.

Rusbult

C. E.

(2010). The importance of seeing what is not there: A quasi-signal detection analysis of positive and negative behavior in newlywed couples. Personal Relationships, 17, 615–633.

22.

Fletcher

G. J. O.

Kerr

P. S. G.

(2010). Through the eyes of love: Reality and illusion in intimate relationships. Psychological Bulletin, 136, 627–658.

23.

Floyd

(2006). Communicating affection. New York, NY: Cambridge University Press.

24.

Gable

S. L.

Reis

H. T.

Downey

(2003). He said, she said: A quasi-signal detection analysis of daily interactions between close relationship partners. Psychological Science, 14, 100–105.

25.

Gleason

M. E. J.

Iida

Bolger

Shrout

P. E.

(2003). Daily supportive equity in close relationships. Personality and Social Psychology Bulletin, 29, 1036–1045.

26.

Gleason

M. E. J.

Iida

Shrout

P. E.

Bolger

(2008). Receiving support as a mixed blessing: Evidence for dual effects of support on psychological outcomes. Journal of Personality and Social Psychology, 94, 824–838.

27.

Graham

Ickes

(Eds.) (1997) When women’s intuition isn’t greater than men’s. In Ickes

(Ed.), Empathic accuracy (pp. 117–143). New York, NY: Guilford Press.

28.

Hendrick

S. S.

(2006). Styles of romantic love. In Sternberg

R. J.

Weis

(Eds.), The new psychology of love (pp. 149–170). New Haven, CT: Yale University Press.

29.

Hinde

R. A.

(1995). A suggested structure for a science of relationships. Personal Relationships, 2, 1–15.

30.

Karney

B. R.

Bradbury

T. N.

(1995). The longitudinal course of marital quality and stability: A review of theory, methods, and research. Psychological Bulletin, 118, 3–34.

31.

Kenny

D. A.

Kashy

D. A.

Cook

W. L.

(2006). Dyadic data analysis. New York, NY: Guilford Press.

32.

Kenny

D. A.

Ledermann

(2010). Detecting, measuring, and testing dyadic patterns in the actor–partner interdependence model. Journal of Family Psychology, 24, 359–366.

33.

Kurdek

L. A

. (1998). Developmental changes in marital satisfaction: A 6-year prospective longitudinal study of newlywed couples. In Bradbury

T. N.

(Ed.), The developmental course of marital dysfunction (pp. 180–204). New York, NY: Cambridge University Press.

34.

Kurdek

L. A.

(1999). The nature and predictors of the trajectory of change in marital quality for husbands and wives over the first 10 years of marriage. Developmental Psychology, 35, 1283–1296.

35.

Lemay

Jr., E. P.

Clark

M. S.

(2008). How the head liberates the heart: Projection of communal responsiveness guides relationship promotion. Journal of Personality and Social Psychology, 94, 647–671.

36.

MacGeorge

E. L.

Feng

Burleson

B. R

. (2011). Supportive communication. In M. A. Knapp & J. A. Daly (Eds.), Handbook of Interpersonal Communication (pp. 317–354). Sage.

37.

Mikulincer

Shaver

(2007). Attachment patterns in adulthood: Structure, dynamics, and change. New York, NY: Guilford Press.

38.

Murray

S. L.

(1999). The quest for conviction: Motivated cognition in romantic relationships. Psychological Inquiry, 10, 23–34.

39.

Murray

S. L.

Holmes

J. G.

(2011). Interdependent minds. New York, NY: Guilford Press.

40.

Murray

S. L.

Holmes

J. G.

Bellavia

Griffin

D. W.

Dolderman

(2002). Kindred spirits? The benefits of egocentrism in close relationships. Journal of Personality and Social Psychology, 82, 563–581.

41.

Murray

S. L.

Holmes

J. G.

Collins

N. L.

(2006). Optimizing assurance: The risk regulation system in relationships. Psychological Bulletin, 132, 641–666.

42.

Neff

L. A.

Karney

B. R

. (2005). To know you is to love you: The implications of global adoration and specific accuracy for marital relationships. Journal of Personality and Social Psychology, 88, 480–497.

43.

Neff

L. A.

Karney

B. R.

(2009). Compassionate love in early marriage. In Fehr

Sprecher

Underwood

L. G.

(Eds.), The science of compassionate love (pp. 201–221). Malden, MA: Wiley-Blackwell.

44.

Overall

N. C.

Simpson

J. A

, (2013). Regulation processes in close relationships. In Simpson

J. A.

Campbell

(Eds.), The Oxford handbook of close relationships (pp. 427–451). New York, NY: Oxford University Press.

45.

Rafaeli

Gleason

M. E. J.

(2009). Skilled support within intimate relationships. Journal of Family Theory and Review, 1, 20–37.

46.

Raudenbush

S. W.

Brennan

R. T.

Barnett

R. C.

(1995). A multivariate hierarchical model for studying psychological change within married couples. Journal of Family Psychology, 9, 161–174.

47.

Reis

H. T.

Aron

(2008). Love: What is it, why does it matter, and how does it operate? Perspectives on Psychological Science, 3, 80–86.

48.

Reis

H. T.

Carmichael

(2006). Perceived partner regard and interaction in close relationships. Unpublished manuscript, University of Rochester.

49.

Reis

H. T.

Clark

M. S.

(2013). Responsiveness. In Simpson

J. A.

Campbell

(Eds.), The Oxford handbook of close relationships (pp. 400–423). New York, NY: Oxford University Press.

50.

Reis

H. T.

Clark

M. S.

Holmes

J. G.

(2004). Perceived partner responsiveness as an organizing construct in the study of intimacy and closeness. In Mashek

Aron

(Eds.), The handbook of closeness and intimacy (pp. 201–225). Mahwah, NJ: Erlbaum.

51.

Reis

H. T.

Gable

S. L.

Maniaci

M. R

. (in press). Methods for studying everyday experience in its natural context. In Reis

H. T.

Judd

(Eds.) Handbook of research methods in social and personality psychology (2nd ed.). New York, NY: Cambridge University Press.

52.

Reis

H. T.

Patrick

B. C.

(1996). Attachment and intimacy: Component processes. In Kruglanski

Higgins

E. T.

(Eds.), Social psychology: Handbook of basic principles (pp. 523–563). New York, NY: Guilford Press.

53.

Reis

H. T.

Shaver

P. R.

(1988). Intimacy as an interpersonal process. In Duck

(Ed.), Handbook of personal relationships (pp. 367–389). Chichester, UK: John Wiley & Sons.

54.

Rusbult

C. E.

Kumashiro

Stocker

S. L.

Wolf

S. T.

(2005). The Michelangelo phenomenon in close relationships. In Tesser

Wood

J. V.

Stapel

D. A.

(Eds.), On building, defending, and regulating the self: A psychological perspective (pp. 1–29). New York, NY: Psychology Press.

55.

Rusbult

C. E.

Olsen

Davis

J. L.

Hannon

P. A.

(2001). Commitment and relationship maintenance mechanisms. In Harvey

J. H.

Wenzel

(Eds.), Close romantic relationships: Maintenance and enhancement (pp. 87–113). Mahwah, NJ: Erlbaum.

56.

Schwarz

(2012). Why researchers should think “real-time”: A cognitive rationale. In Mehl

M. R.

Conner

T. S.

(Eds.), Handbook of research methods for studying daily life (pp. 22–42). New York, NY: Guilford Press.

57.

Sprecher

Toro-Morn

(2002). A study of men and women from different sides of earth to determine if men are from Mars and women are from Venus in their beliefs about love and romantic relationships. Sex Roles, 46, 131–147.

58.

Thomas

Fletcher

G. J. O.

(1997). Empathic accuracy in close relationships. In Ickes

(Ed.), Empathic accuracy (pp. 218–250). New York, NY: Guilford Press.

59.

Underwood

L. G.

(2009). Compassionate love: A framework for research. In Fehr

Sprecher

Underwood

L. G.

(Eds.), The science of compassionate love (pp. 3–25). Malden, MA: Wiley-Blackwell.

60.

Weiselquist

Rusbult

C. E.

Foster

C. A.

Agnew

C. R.

(1999). Commitment, pro-relationship behavior, and trust in close relationships. Journal of Personality and Social Psychology, 77, 942–966.

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