You Feel Better When Your Partner is Emotionally Intelligent: Self-Rated Emotional Intelligence Shows Partner Effects on Subjective Wellbeing

There are many conceptualizations of wellbeing, with perhaps the most prominent being the distinction between feeling good versus feeling fulfilled (Keyes et al., 2003; Ryan & Deci, 2001). Feeling good (subjective wellbeing) represents a hedonistic view of wellbeing focused on pleasure, variously referred to as subjective wellbeing, hedonic wellbeing or emotional wellbeing. It includes both affective and cognitive components. The affective component involves the experience of pleasant feelings (such as joy) and the absence of unpleasant feelings (such as sadness or fear; Diener et al., 1999). The cognitive component involves the positive evaluation of life overall (life satisfaction) or specific life experiences (such as job satisfaction or relationship satisfaction (Pavot & Diener, 2008). In contrast, feeling fulfilled (psychological wellbeing) represents the Aristotelian tradition of eudaimonia, and is referred to as flourishing, psychological wellbeing, or eudaimonic wellbeing (Ryan & Deci, 2001). In the current study, we use wellbeing indices from each of the three major elements of wellbeing: (1) eudaimonic wellbeing (assessed with Ryff’s [1989] psychological wellbeing scale); (2) the cognitive element of subjective wellbeing (life satisfaction); and (3) the affective component of subjective wellbeing (positive and negative affective states).

There is considerable evidence (mainly from meta-analyses) that both streams of EI have a positive association to all three major elements of wellbeing (Fernández-Berrocal & Extremera, 2016; Helmi, 2021; MacCann et al., 2020; Sánchez-Álvarez et al., 2016). However, there are differences across the streams and across the wellbeing indicators. First, ability EI shows a weaker relationship to all aspects of wellbeing as compared to self-rated EI (Sánchez-Álvarez et al., 2016; Xu et al., 2021). Second, both streams of EI show stronger associations with eudaimonic wellbeing (psychological wellbeing) than subjective wellbeing (life satisfaction or affect; Bhullar et al., 2013; Brackett & Mayer, 2003). Third, ability EI shows stronger relationships to cognitive than to affective subjective wellbeing (Fernández-Berrocal & Extremera, 2016; Xu et al., 2021), whereas self-rated EI shows similar relationships to cognitive versus affective wellbeing (Sánchez-Álvarez et al., 2016). Fourth, ability EI predicts lower negative affect but not higher positive affect (MacCann et al., 2020).

The above relationships all refer to actor effects—whether individual differences in people’s EI affects their own well-being. However, a person’s level of EI does not just affect their own life, but the lives of those around them. We also expect there to be partner effects of EI on wellbeing in romantic relationships. There are clear mechanisms linking EI to better partner wellbeing. In the case of ability EI, an emotionally intelligent person would be able to see when their partner is upset (because they perceive emotions well), understand what has upset them (because they understand the causes of emotions) and know what to do to manage their emotion (because they know what to do to manage emotions). As such, all else being equal, people with high EI partners may feel less negative emotion (as their partner is able to help them manage their emotions effectively). Moreover, someone whose partner consistently responds well to them and understands them would evaluate that their life is good and would feel more fulfilled—that is, feel greater satisfaction with their life and greater psychological wellbeing.

In the case of self-rated EI, self-perceptions of emotional skills are likely to drive emotional and interpersonal behaviour. This is because a person’s beliefs about their capabilities motivate their actions (e.g., Bandura, 1982). For example, if a person believes they can detect their partner’s feelings (emotion perception beliefs) and can control their temper (emotion management beliefs), they are likely to invest the time and effort into doing these things. Most people would feel happier and more satisfied with their life with a partner who attends to their emotional state and refrains from temper tantrums.

The current study aimed to evaluate the effects of one’s romantic partner’s EI on an individual’s wellbeing. Drawing from a large sample of romantic couples, we evaluated whether self-rated and ability EI streams predict various wellbeing components (life satisfaction, psychological wellbeing, positive/negative affect) in both members of the couple. Using the Actor-Partner Interdependence Model (Kenny et al., 2006), we analyzed the dyadic EI-wellbeing relationships. Given the strong positive relationship between EI and one’s wellbeing as well as the importance of one’s romantic partner’s characteristics for wellbeing, we hypothesized that there would be both positive actor and partner effects of EI on wellbeing.

Participants were recruited from English speaking countries (UK, US, Australia, New Zealand) on Prolific (www.prolific.com) as part of a large-scale ongoing longitudinal study of romantic couples. The measures for this study were completed over two online sessions. 422 couples in heterosexual relationships completed both sessions. The sample size was based on the maximum number of romantic couples that we could recruit from Prolific in a 3-week period. 8 couples failed both attention check items in one or both sessions and were excluded from the analysis. In addition, 3 couples completed a session in less than a third of the median time and were also excluded. 4 couples had missing data on one or more of the key study variables and were excluded from the subsequent analyses, leaving a final sample of 407 couples (M_age = 36.50, SD_age = 10.52). The average relationship duration was 10.80 years (SD = 8.47 years).

We note that Prolific is known to produce high quality data and has multiple mechanisms in place for the detection and reduction of bot responding (e.g., Douglas et al., 2023; Eyal et al., 2021). Data was collected in 2021.

The 18-item version of this instrument assesses aspects of wellbeing associated with autonomy, personal growth, relationships, meaning, self-acceptance, and environmental mastery (e.g., “For me, life has been a continuous process of learning, changing, and growth”). Participants responded to each item on a 6-point scale (1 = “Strongly Disagree”, 6 = “Strongly Agree”).

This 5-item scale measures global life satisfaction. Items are rated on a 7-point scale from 1 = “Strongly Disagree” to 7 = “Strongly Agree” (e.g., “I am satisfied with my life”).

Affect was assessed using 14 emotion adjectives. 12 items were from the PANAS-X (Watson & Clark, 1994), five of which were negatively valenced (afraid, nervous, upset, ashamed, hostile) and seven of which were positively valenced (determined, attentive, alert, inspired, active, calm, relaxed). Two additional positively-valenced items (peaceful, content) were also included in order to include more low-arousal positive items. Each item was phrased to refer to the participant’s feelings over the last week (i.e., Please indicate the extent you have felt this way over the LAST WEEK) and were rated on a 5-point scale from 1 = “very slightly or not at all” to 5 = “extremely.” The set of affect items was selected with the goal of covering varying degrees of arousal and valence, as well as being affective states that participants would most likely experience over the course of a typical week.

Three measures of ability EI were administered to capture the major branches of emotion perception, emotion understanding, and emotion management. The understanding and management branches of emotional intelligence were assessed with the 42-item Situational Test of Emotion Understanding (STEU; MacCann & Roberts, 2008) and the 18-item Situational Test of Emotion Management– Brief (STEM-B; Allen et al., 2015). The emotion perception branch was assessed using the 56-item Japanese and Caucasian Brief Affect Recognition Test (JACBART; Matsumoto et al., 2000).

In the STEM-B, participants are presented with brief vignettes detailing an emotional situation and asked to choose the most effective response. E.g., ‘Andre moves away from the city his friends and family are in. He finds his friends make less effort to keep in contact than he thought they would. What action would be the most effective for Andre? (a) Try to adjust to life in the new city by joining clubs and activities there; (b) He should make the effort to contact them, but also try to meet people in his new city; (c) Let go of his old friends, who have shown themselves to be unreliable; (d) Tell his friends he is disappointed in them for not contacting him.’ Items were scored based on expert consensus using partial scoring, with scoring weights determined by the proportion of experts who select each option as the best answer.

In the JACBART, participants are briefly presented with Japanese and Caucasian faces for 1s and then asked to identify the emotion that was being displayed from a list of seven potential options. Participants were only able to watch each expression one time. The 56 items measure a variety of emotions using faces of differing sex and race, with an equal distribution of items across sex and race. The emotions captured are balanced across the seven ‘universal’ emotions (Ekman & Friesen, 1986).

Self-rated EI was measured using the 16-item Wong and Law Emotional Intelligence Scale (Wong & Law, 2002). The items assess participants’ beliefs about their emotional abilities, including the ability to perceive their own emotions (e.g., “I have good understanding of my own emotions”), perceive others’ emotions (e.g., “I am a good observer of others’ emotions”), self-efficacy, (e.g., “I always tell myself I am a competent person.”) and regulate their own emotions (e.g., “I have good control of my own emotions”). Participants respond to each item on a 5-point scale from 1 = “Strongly Disagree” to 5 = “Strongly Agree”.

Education was self-reported on a 7-point scale from 1 = “No formal qualifications” to 7 = “Doctorate degree.”

Participants completed the aforementioned measures over two online sessions. In the first session, participants completed basic demographic variables (age, gender, education, relationship duration) before completing the ability EI measures as well as the affect and psychological wellbeing scales. In the second session, participants completed the remaining EI measure along with the life satisfaction scale. The two sessions were completed approximately a week apart, with both members of a couple needing to complete the first session before being invited to the subsequent session. All measures were administered using Qualtrics. All procedures were approved by the Human Ethics Committee.

Bivariate correlations are presented in Table  1 separately for men (upper triangle) and women (lower triangle). Wellbeing outcomes were all moderately related to each other in the expected directions. Self-rated EI was significantly positively related to all wellbeing outcomes (except negative affect where the relationship was, as expected, negative). Ability EI had an unexpectedly negative correlation with positive affect for both men and women and was not significantly related to any of the other wellbeing outcomes.

Table 2 presents within-dyad correlations (cross-partner). All variables showed positive covariation between partners (on the same variable). The cross-partner EI-wellbeing effects replicated the within-person correlations described above, but with overall smaller magnitudes. Self-rated EI scores were positively related to psychological wellbeing, life satisfaction, and positive affect while being negatively correlated with negative affect. Ability EI, however, again showed unexpected relationships with the wellbeing variables– significantly predicting lower positive affect in both cases (i.e., both men’s EI to women’s affect and women’s EI to men’s affect) and lower satisfaction with life in the case of women to men. There were no significant cross-partner effects of ability EI on negative affect or psychological wellbeing.

A structural equation modelling (SEM) approach applying the actor–partner interdependence model (APIM) was used to examine within-dyad effects. All models included age, education, and relationship length as covariates. Age and education were within-dyad variables, while relationship duration was entered as a between-dyad variable¹. We present the APIM results in Figs. 1 and 2. Note that for simplicity covariates are not included in the figures, with full model results including all regression paths available in the online supplementary materials.

While self-rated EI significantly predicted all wellbeing variables, effects were larger for psychological wellbeing than subjective wellbeing, consistent with previous findings (Bhullar et al., 2013; Urquijo et al., 2016). Ability EI showed inconsistent actor effects: (a) there was no significant relationship for life satisfaction, (b) only men showed significant effects for higher psychological wellbeing and lower negative affect, and (c) only women showed a relationship with positive affect, and this was negative (i.e., higher EI was related to lower positive affect). Stronger effects of ability EI for men than women EI is consistent with prior research. For example, Brackett et al. (2004) found that ability EI in college students was significantly related to lower drug use, alcohol use, deviant behaviour and negative social interactions for men but not woman. One reason proposed for this difference related to gendered social norms for positive interpersonal behaviour. That is, women are under greater social pressure to perform positive interpersonal behaviour, such that even low-EI women are compelled to learn to do this adequately (whereas low-EI men are not). Compared to the unstable actor effects of ability EI, the current study shows that self-rated EI consistently relates to one’s own well-being.

Overall, partner effects of EI were smaller than actor effects (unsurprisingly, given different-source ratings for actor and partner; Orth, 2013). Results differed across EI streams.

For self-rated EI, there were significant partner effects, predicting higher positive affect, lower negative affect (women to men), and higher life satisfaction (but not higher psychological wellbeing). This suggests that a person’s self-rated EI can enhance the positive feelings and life satisfaction of their romantic partner, although these findings may not extend to psychological wellbeing. This may be because PWB items capture eudaimonic themes (autonomy, growth) which may be less immediately influenced by daily dyadic emotion regulation than hedonic indices.

For ability EI, women’s EI predicted lower negative affect in men (as expected) but also lower positive affect in men (i.e., the opposite direction to our expectations). Partner effects from men to women for positive affect were also negative (i.e., higher ability EI in men related to lower positive affect in women). This result is consistent with theories that ability EI magnifies rather than improves emotional experiences (Fiori & Ortony, 2021). Under this hypersensitivity hypothesis, higher EI abilities lead to greater sensitivities to emotional cues, which can have negative consequences as well as positive consequences. For example, Fiori and Ortony (2021) found that higher ability EI individuals provide more extreme (positive or negative, depending on the situation) impressions of others, rather than simply providing more positive impressions of others. While high EI people felt more positively towards people with positive qualities, they felt more negatively towards people with negative qualities. In close personal relationships, this hypersensitivity of a high-EI partner is not necessarily good for you. If your partner can perceive and understand your emotions, then you cannot conceal your negative feelings towards them—a high-EI partner will detect contempt, anger, and disgust, which can kickstart processes such as negative affect reciprocity that are harmful to harmonious relationship functioning (e.g., Gottman et al., 1998). We note that the hypersensitivity hypothesis applies only to emotional abilities (ability EI) and not to emotional beliefs (self-rated EI; Fiori et al., 2023). This is because ability emotional intelligence entails genuine, objectively measurable emotional sensitivity, whereas self-rated emotional intelligence represents subjective self-assessment and does not inherently produce the amplified emotional experience. Moreover, because positive and negative affect were modelled in separate equations, the hypothesized magnification process could appear as a dampening of partners’ positive affect without a parallel rise in negative affect when baseline negativity is already near floor level; future work should test this floor-effect interpretation directly.

There were significant partner effects of self-rated EI on positive affect, negative affect, and life satisfaction, but not psychological wellbeing. This suggests that a person’s self-rated EI can enhance the positive feelings and life satisfaction of their romantic partner, but not their partner’s psychological wellbeing. These mixed results are generally consistent with prior research examining partner effects of EI on partner outcomes. Although there were significant partner effects of self-rated EI on relationship satisfaction, conflict resolution, and reduced negative affect (Schröder–Abé & Schütz, 2011), there were no significant partner effects of ability EI on partner outcomes (Zeidner & Kloda, 2013). In contrast to prior findings, we found significant partner effects of ability EI on positive affect, but these were negative such that a person’s ability EI negatively predicted their partner’s positive feelings. Overall, we would conceptually expect a person’s EI to impact their partner’s wellbeing, however, the data only partially supports this expectation across self-rated EI, and not as we predicted for ability EI.

Our findings indicate that an individual’s beliefs about their emotional intelligence (self-rated EI) rather than their objectively measured emotional skills (ability EI) are most strongly linked to their partner’s emotional wellbeing, especially in terms of enhancing positive affect and life satisfaction and reducing negative affect. This highlights important implications for counselling and therapeutic interventions targeting relationship satisfaction and emotional wellbeing within couples.

Specifically, our results suggest that therapeutic approaches aimed at enhancing people’s confidence and self-efficacy in their emotional skills might yield greater benefits for partner wellbeing than those primarily focused on training such skills. Couples counselling sessions that emphasize cognitive reframing, self-efficacy building, and positive affirmation of each partner’s emotional competence (e.g., reinforcing beliefs such as “you are capable of understanding your partner’s feelings and making them feel valued and supported”) could potentially lead to more substantial improvements in the partner’s emotional experience.

Conversely, traditional interventions that rely solely on explicit skill training—such as workshops or exercises designed to teach emotion recognition or emotional management strategies—might be less effective or even inadvertently counterproductive if they undermine an individual’s perceived emotional competence. Counsellors and therapists might therefore consider prioritizing interventions that bolster partners’ perceptions of their emotional effectiveness through reflective practices, positive feedback, and supportive counselling dialogues. Approaches such as cognitive behavioural couples therapy (CBCT), strengths-based counselling, or solution-focused therapy, which actively cultivate positive self-beliefs and affirmations regarding emotional abilities, could be particularly beneficial in improving partners’ mutual emotional wellbeing and overall relationship satisfaction.

We showed that there are small but significant partner effects linking EI and wellbeing. However, the partner effects were not consistent across EI streams nor wellbeing constructs. As mentioned above, there may be specific theoretical reasons for these inconsistencies, but they may also be due to inadequate power in the current sample to detect very small effects. Although relative to most dyadic studies the sample presented here is relatively large, most of the partner effects were in the small to very small effects size range and both our own study and future studies may have difficulty accurately detecting such effects (particularly given the logistical challenges of recruiting dyads). Meta-analytic techniques may therefore be needed to provide a robust estimate of partner EI-wellbeing effects. Furthermore, given the lack of support for partner effects from ability EI to subjective wellbeing, it might be that there are, in fact, little benefits of higher EI to one’s partner’s wellbeing and the relationships we observed between self-rated EI and subjective wellbeing are due to the conceptual issues associated with the construct of self-rated EI discussed previously.

Moreover, future work should incorporate theoretically adjacent constructs—most notably perceived partner-responsiveness and self-esteem—into the dyadic models. Including these variables will allow researchers to quantify the incremental predictive power of self-rated and ability EI over and above well-established interpersonal (responsiveness) and intrapersonal (self-esteem) determinants of partner wellbeing, clarifying whether EI confers unique benefits or simply mirrors these broader traits.