Emotions play a crucial role in relationship dynamics, with their elicitation and management being essential factors that can significantly influence conflict outcomes (Kopystynska et al., 2017; Kieslich and Steins, 2022). Emotions are complex human states that involve cognitive, physiological, and behavioral components (Kolb et al., 2019). These components encompass subjective states and physiological responses that prepare the organism to respond to stimuli. Evidence in the literature suggests that people maintain a certain degree of control over their emotions, using various strategies to regulate how and when these emotions are experienced. In doing so, individuals can monitor, evaluate, and modify the processes involved in the genesis of their emotions, allowing them to modulate the intensity and expression of these emotions (Salovey and Mayer, 1990; Mayer et al., 1999; Pineda et al., 2018). This ability to control and use emotions to guide thought and behavior is what Salovey and Mayer (1990) defined as Emotional Intelligence (EI), distinguishing between the processing of affective information at intrapersonal and interpersonal levels. EI has been shown to enhance people’s ability to cope with stressors and conflict situations of high emotional intensity (Ruiz-Mamani et al., 2022) and is related to important social functioning variables such as attachment style and family cohesion (Cócola, 2022). Moreover, recent reviews have highlighted the role of EI in several health and social scenarios; improving the quality of life of breast cancer survivors (Durosini et al., 2022), enhancing EI in patients with panic disorder to improve their flexibility in emotion regulation (Oussi et al., 2023), or in individuals with social anxiety disorder (Rozen and Aderka, 2023). In the specific context of couple relationships, EI has been shown to be related to several variables. In a recent systematic review and meta-analysis, Jardine et al. (2022) found a positive and significant correlation between EI and romantic relationship satisfaction. In addition, and focusing on women, interventions aimed at working on and increasing EI in married couples are associated with improved marital satisfaction, sexual quality of life, and psychological well-being. (Milani et al., 2020). In a multiple study, Alonso-Ferres et al. (2019) found gender differences in coping with couple conflict and a mediating role of EI in the adoption of adaptive responses, leading to greater psychological well-being and higher relationship satisfaction.

As shown, EI has a significant impact on individuals’ lives and reactions (Zeidner and Kaluda, 2008), and an intelligent use of emotions leads to an appropriate handling of both emotions and relationships (Goleman et al., 2002). Beyond individual differences, traditional socialization of boys and girls according to gender produces different emotional experiences in terms of emotional intensity, strength, clarity, and instability (Bailen et al., 2019). The need for researchers to consider the gender perspective to break down some of the traditional stereotypes and broaden the field of study regarding women and emotions has been widely accepted by the scientific community, especially in social and couple situations where emotional identification, expression, and regulation are essential, such as in the psychological context of domestic violence (Ferrer and Bosch, 2005; Pineda et al., 2023). Indeed, research has shown that in the context of couples, women’s emotional intelligence is negatively related to the propensity to initiate and maintain a relationship with an abusive partner (Ambler et al., 2023). Therefore, in close relationships, the deepest emotions and feelings are shared; therefore, couples must be able to understand, talk about and manage them to ensure a healthy relationship (Batool and Khalid, 2012). Perceived Emotional Intelligence (PEI) is the perception of one’s emotional abilities as measured by self-report (Salovey et al., 2002). Following the model developed by Mayer and Salovey (1997), (for a more in-depth theoretical review see Fernández-Berrocal and Extremera, 2006) and the ensuing debate on how to measure EI, the term PEI has been proposed to refer to individual meta-knowledge about one’s emotional abilities rather than actual capacity, measuring fundamental aspects of emotional self-awareness such as perceived attention to one’s own emotional states (Emotional Attention), perceived understanding of these states (Emotional Clarity), and perceived ability to regulate these states (Emotional Repair; Extremera and Fernández-Berrocal, 2005; Salguero et al., 2012). In romantic relationships, emotional awareness and the ability to identify and regulate emotions have a significant impact on the well-being of couples (Fishbane, 2011). PEI has been positively related to adaptive coping and problem solving, especially when considering the Emotional Clarity factor (Augusto et al., 2008). Furthermore, those capable of perceiving, assimilating, and regulating their emotions, as well as appropriately interpreting those of others, are more resilient and more satisfied with life (Veloso-Besio et al., 2013). Therefore, PEI can be crucial in preventing conflict situations in couples by effectively identifying and regulating emotions that can cause problems in the relationship. Positive conflict resolution styles have been associated with effective anger regulation, which helps avoid negative unregulated emotions that drive conflict (Rivers et al., 2007).

High emotional arousal in interpersonal conflict situations can lead to “flooding,” a state of overstimulation, overwhelm and cognitive disorganization (Malik et al., 2020). This term was first introduced by behavior psychologists analyzing the treatment of anxiety due to immediate and prolonged exposure to feared stimuli (Rachman, 1966). However, it was (Gottman, 1993a,b) who first used the term Emotional Flooding (EF) to refer to couple relationships in which hard-to-control feelings of emotional unrest and overload affect sensory perception and proper cognitive functioning. Emotions guide our decisions along with the rational mind, and the thinking brain moderates their expression, except in situations of EF, when emotions overwhelm reason and the emotional brain takes complete control of the situation (Goleman, 1996). EF focusses on how an individual experiences a situation of anger and negative emotions, where their partner loses control, and how this affects the outcome. In most cases, it leads to aggressive behavioral responses, as described by Braiker and Kelley (1979) in their model described below. When on a regular basis either member of the couple feels emotionally flooded due to the attitude of his/her partner, any action carried out by said partner may be negatively interpreted, leading to intensely inappropriate reactions and negative feelings that will inhibit problem-solving (Malik et al., 2020). As time passes, the individual may feel that talking and trying to solve the problems are useless, resulting in the potential search for relief from these negative feelings in a parallel or isolated manner. This may result in an increased distancing between partners, who may ultimately decide to separate (Gottman, 1993b).

As Sotskova et al. (2015) point out, flooding is a relatively new concept in the emotion literature, but we can find theoretical similarities with aspects of emotional sensitivity. Furthermore, EF understood as the loss of rational control following a flood of intense negative emotions that leads the person to engage in irrational behaviors (Penberthy et al., 2018) can trigger unexpected and emotionally charged responses such as aggressive behaviors (Braiker and Kelley, 1979; Gómez-Leal et al., 2022) escape or withdrawal behaviors (Gottman, 1993b; Malik et al., 2020), which would ultimately have a negative impact on couple satisfaction. In this sense, aggression and impulsivity have shown indirect and significant effects with EI (Coccaro et al., 2016). Consequently, people with higher EI would be able to adopt more positive behavioral styles in stressful situations like PCR, avoiding both aggressive and dysregulated negative emotional behaviors and conflict. Therefore, this work aims to deepen the analysis between these variables, assuming that people with high emotional self-awareness, who are better able to regulate their negative emotions, will end up using more PCR when arguing with their couples and, as a result, will be less involved in conflict situations.

Couple relationships differ from other relationships given the presence of certain unique processes and expectations such as fidelity and emotional and romantic exclusivity (Vidal et al., 2012). One of the most damaging aspects, both for people’s well-being and for the relationship, is conflict (García and Umberson, 2019). Conflicts are disagreements that can be constructive for the relationship, depending on the strategies employed to resolve them (Cascón, 2006). In the context of Social Exchange Theory (Thibaut and Kelley, 1959; Blau, 1964; Burns, 1973; Homans, 1974), Braiker and Kelley (1979) develop their model of dyadic close relationship in which they underline the emotion-conflict relation according to the intensity. Thus, with a high intensity of interpersonal conflict, we will find a high level of emotional activation, leading the subject to resolve the conflict in the short term through the use of two strategies; escape or exaggerated behavioral response (e.g., aggressive behavior against the partner). Researchers have found significant differences when examining conflict resolution strategies used by men and women in their relationships (Black, 2000; Alexander, 2001; Laca et al., 2006) as well as in other areas such as education (Skordoulis et al., 2020; Galindo-Domínguez et al., 2022). In general, women tend to use more positive strategies, such as cooperation-based conflict resolution styles, compared to men who typically rely on more aggressive styles (Ma, 2005; Garaigordobil and Maganto, 2011; Garaigordobil et al., 2016). The ability to resolve conflicts is therefore a key element in the development of social relationships, and is particularly important in sentimental relationships, where disagreements and arguments often arise which, if not properly resolved, can lead to insecurity, deception or dissatisfaction, both sexual and with the partner (de Wied et al., 2007; Quesada, 2020). In addition, certain couple characteristics such as age or length of relationship, which may influence both the frequency and intensity of conflict, have been shown to be associated with increased levels of emotional partner violence (Juarros-Basterretxea et al., 2022; Rojas-Solís and Romero-Méndez, 2022). Throughout his research, Gottman has also been interested in analyzing the behavior that occurs in conflict situations between couples (Gottman, 1993b, 1994), conflicts identified as the “four horsemen of the apocalypse” (criticism, defensiveness, contempt, and stonewalling), which are essentially negative conflict resolution styles given their impact on personal well-being and relationship satisfaction. Therefore, in line with this approach, in the present research, we expect to find that the opposite strategies, i.e., Positive Conflict Resolution strategies (PCR), lead to fewer and less intense conflicts with the partner.

On the basis of the evidence described above, we conducted the present research in an all-female sample with a 2-fold aim. On the one hand, we sought to extend the literature on women’s relationships by exploring group differences by age and relationship length on the endogenous variables of PCR, Conflict, EF, and Satisfaction. We also wanted to see if having children had any impact on Conflict and Satisfaction. On the other hand, we want to analyze the relationship between PEI, PCR, Conflict, EF, and Satisfaction by building an explanatory model using structural equations in this sample of women. To this end, the following hypotheses were formulated:

This study followed a cross-sectional design, with participants being recruited through an online questionnaire and relying on non-probabilistic, snowball sampling that was distributed across distinct social media directed at Spanish residents. The hypothetical model based on the theory, together with the proposed hypotheses, is shown in Figure 1.

Participation in this study was voluntary and no compensation or payment was given to the participants. After an initial literature review and selection of instruments, an online questionnaire was created in Google Docs and the link was distributed through several social networks such a Twitter, Facebook. and Instagram. Participants were asked to share the link with friends and acquaintances. The study objectives were explained along with the potential benefits of participation. Informed consent was requested from the participants to process their data. A total of 825 responses were received. As the purpose of the study was to analyze the functioning of the variables from a gender perspective, only responses from female participants who were currently or previously in a relationship with a heterosexual partner were selected. The sample size exceeded the minimum required considering the number of latent constructs and items (Hair et al., 2010) This study is part of a larger research project and was conducted following the Declaration of Helsinki guidelines. Research ethical approval for the study was granted by the Research Ethics and Integrity Committee of Miguel Hernández University, Spain (DPS.AGL.01.21).

The final sample consisted of 692 women with a mean age of 31.76 years (SD = 7.99). Of these, 64.6% were involved in a couple relationship, 21.7% were married, 11.0% single, 2.3% divorced, and 0.4% widowed. Of the selected sample, 28.0% had children. To make comparisons between groups, the sample was divided into three age groups and four groups according to the length of the relationship. Age groups were created following the recommendations of Martín (2005): young adults (up to 39 years), middle-aged (40–49 years), and older adults (>50 years). The groups by relationship length were divided into four categories according to a statistical distribution criterion of equal percentiles based on scanned cases with three cut-off points: less than 3 years, between 3 and 7 years, between 7 and 11 years, and more than 11 years. A total of 32.1% of the sample had been in a relationship for 3 years or less, 29.9% between 3.1 and 7 years, 18.4% between 7.1 and 11 years, and 19.7%, for over 11.1 years. Most of the participants had studies at the university or master’s level studies (83.8%).

A sociodemographic data questionnaire was used in this study. It was an ad hoc questionnaire that recorded age, civil status, sexual orientation, number of children, and level of education. In addition, a Spanish version of the Trait Meta-Mood Scale (TMMS-24) by Salovey et al. (1995), the Scale of Emotional Flooding in Couple Relationships (Berenguer, 2019), the Spanish version of the Conflict Resolution Style Inventory (CRSI; Kurdek, 1994), the Spanish version of the Kansas Marital Satisfaction Scale (KMS; Schumm et al., 1983), and an ad hoc questionnaire measuring conflict in couples were used.

The SPSS 18.0, AMOS 23.0, and JASP 0.17.1 statistical packages were used for descriptive analysis, correlations, factorial analyses, and between-groups comparisons. Data were preliminarily screened for outliers and errors (not detected). For correlations, Spearman’s rho coefficient was used. The normality of the variables was verified with the Kolmogorov–Smirnov test and homoscedasticity was examined using Levene’s test for equality of variance. Due to the lack of assumptions required for parametric tests such as normality and symmetry, Kruskal-Wallis H and Mann–Whitney U tests were used to compare groups. For all of these tests, an alpha level of 0.05 was used. To correct the possible error resulting from the 5% threshold, the Bonferroni correction was used. The post hoc Dunn’s adjustment test was performed to examine significant differences between groups. Effect sizes were calculated using r and Cohen’s d.

Confirmatory factorial analyses (CFA) were used to assess the psychometric properties of the instruments, as well as their factorial structure. Varimax rotation was used for the extraction of the main components. Estimation, identification, assessment, and re-specification of the structural equation model (SEM) were performed in AMOS. A variance–covariance matrix was used as well as maximum likelihood estimation. When estimating the parameters that define the factors, as well as the relations between variables, recommendations of Bentler and Chou (1987) were considered with respect to sample size.

For path analysis, SEM was used. Direct and indirect effects were calculated using a bootstrap technique to determine significance. A bootstrap sample of 5,000 was used. Significance was assessed using the bias-corrected percentile method with a 95% confidence interval. A multicollinearity test was performed with negative results. The goodness of fit parameters were estimated using a maximum likelihood estimation model. Given the lack of consensus regarding the concept of fit, various indices were considered simultaneously (Tanaka, 1993) to analyze the model’s fit: the discrepancy between the Chi-squared distribution and the degrees of freedom (χ2/df, values lower than 5 are considered correct), the root mean square error of approximation (RMSEA, a value equal to or less than 0.05 indicates a good fit). Values equal to or less than 0.08 can also be considered to correspond to reasonable errors of approximation (Jöreskog, 1993), the normed fit index (NFI, values above 0.90 are interpreted as an acceptable fit; Byrne, 1994, 2010; Marsh and Grayson, 1995), the Tucker-Lewis index (TLI, values above 0.90 indicate an acceptable fit (Bentler and Bonett, 1980), the comparative fit index (CFI, values equal to or greater than 0.90 indicate a good fit; Byrne, 1994, 2010; Bentler, 2005), the goodness-of-fit index (GFI, values above 0.90 are considered an acceptable fit; Marsh and Grayson, 1995), and the adjusted goodness-of-fit index (AGFI, values above 0.90 are considered an acceptable fit; Jöreskog and Sörbom, 1981, 1984, 1986; Hoyle, 1995).

Table 1 shows the descriptive statistics of the scales completed by participants according to relationship length and age.

Table 2 shows the correlations between the variables.

Significant differences in Conflict by age were found (Figure 2) between the Young Adults (YA, M = 6.740, SD = 1.819) and Middle-Aged adults (MA, M = 7.376, SD = 1.939) groups H (2) = 11.914, p < 0.05, with higher conflict scores in the MA group, with a moderate effect size d = 0.350, 95% CI = 0.071–0.630. By relationship length, significant differences were found between <3 (M = 6.595, SD = 1.756) and + 11 (M = 7.346, SD = 1.823) groups as well as between 3–7 (M = 6.700, SD = 1.835) and + 11 groups H = 15.32 (3), p < 0.001, with moderate effects sizes d = 0.415, 95% CI = 0.125–0.704 and d = 0.356, 95% CI = 0.063–0.649, respectively.

In terms of PCR, significant age differences were found (Figure 3) between the YA (M = 15.740, SD = 2.656) and Older Adults (OA, M = 14.556, SD = 2.501) groups H (2) = 7.264, p < 0.05, with moderate effect size d = 0.443, 95% CI = 0.030–0.916. By relationship length, significant differences were found H (3) = 20.080, p < 0.001 between <3 (M = 15.869, SD = 2.758) and + 11 (M = 14.757, SD = 2.454) groups, 3–7 (M = 16.039, SD = 2.666) and + 11 groups, as well as 7–11 (M = 15.551, SD = 2.609) and + 11 groups, with moderate effects sizes d = 0.420, 95% CI = 0.131–0.710, d = 0.484, 95% CI = 0.190–0.778 and d = 0.300, 95% CI = 0.027–0.627, respectively.

Regarding EF (Figure 4) significant differences were found H (2) = 13.042, p < 0.001 between YA (M = 42.124, SD = 12.762) and MA (M = 45.859, SD = 13.238) groups as well as YA and OA (M = 48.778, SD = 10.868) groups, with a moderate effect size d = 0.293, 95% CI = 0.013–0.572 and medium effect size d = 0.522, 95% CI = 0.048–0.995, respectively. By relationship length, significant differences were found H (3) = 16.407, p < 0.001 between <3 (M = 40.959, SD = 13.748) and + 11 (M = 45.838, SD = 11.157) groups, 3–7 (M = 43.014, SD = 12.785) and + 11 groups as well as 7–11 (M = 42.646, SD = 12.554) and + 11 groups, with moderate effect sizes d = 0.382, 95% CI = 0.093–0.672, d = 0.221, 95% CI = 0.071–0.514 and d = 0.250, 95% CI = 0.077–0.577, respectively.

Satisfaction also was significant when comparing groups (Figure 5). By age, significant differences were found between YA (M = 12.967, SD = 1.902) and MA (M = 12.282, SD = 1.937) groups H (2) = 18.163, p < 0.001 with moderate effect size d = 0.357, 95% CI = 0.077–0.636 as well as between YA and OA (M = 11.630, SD = 2.239) with a medium effect size d = 0.697, 95% CI = 0.222–1.171. By relationship length, significant differences were found H (3) = 20.230, p < 0.001 between <3 (M = 12.946, SD = 2.094) and + 11 (M = 12.191, SD = 1.994) groups with moderate effect size d = 0.392, 95% CI = 0.103–0.682, 3–7 (M = 12.976, SD = 1.711) and + 11 groups with moderate effect size d = 0.408, 95% CI = 0.115–0.702, as well as 7–11 (M = 13.079, SD = 1.859) and + 11 groups, with moderate effect size d = 0.462, 95% CI = 0.133–0.790.

Finally, a significant difference was found when comparing the groups according to whether they had children or not, with higher levels of conflict found in the group of couples with children (Mdn = 7, n = 194) U = 37137.000, z = 4.795, p < 0.001, r = 0.18 and higher scores on Satisfaction in the group without children (Mdn = 13, n = 498) U = 38926.500, z = 4.134, p < 0.001, r = 0.16.

An exploratory factor analysis was conducted to determine the behavior of the conflict questionnaire items and their construct validity. A confirmatory factor analysis was conducted with all other variables included in the model and all the scales retained the original item and factor structure. After the first analysis, the relationship between PEI and Conflict (H2) b = 0.05, p = 0.240 and EF and Satisfaction (H6) b = 0.000, p = 0.992 were found to be insignificant and therefore dropped from the model, so H2 and H6 were not supported. The definitive, simplified SEM along with the standardized structural coefficients is shown in Figure 6. The SEM revealed a good fit with the data χ²(2, N = 692) = 3.373, p < 0.001; RMSEA = 0.059, 95% CI: 0.052, 0.065; NFI = 0.923; TLI = 0.931; CFI = 0.944; GFI = 0.941 and AGFI = 0.918. The direct effect of PEI on PCR was found to be positive and significant b = 0.032, p < 0.001, supporting H1. The direct effect of PCR on Conflict was found to be negative and significant b = −0.711, p < 0.001, supporting H3. The direct effect of Conflict on EF was found to be positive and significant b = 4.465, p < 0.001, supporting H4. Finally, the direct effect of Conflict on Satisfaction was found to be negative and significant b = −0.791, p < 0.001, supporting H5.

In summary, all hypotheses except H2 and H6 were supported and a significant indirect effect was found between PEI and Conflict, EF and Satisfaction through the mediators (Table 3), explaining 70% of the variance in EF and 34% of the variance in Satisfaction.

The study’s primary participants are individuals who identify as heterosexual, are in monogamous, committed relationships—including marriage—and have been in relationships for a significant length of time. Although this selection criterion was motivated by the fact that most previous research on intimate relationships has focused on the experiences of heterosexual couples, caution should be exercised in generalizing the results. Additionally, the vast majority of the sample had higher education levels and the age sample was unbalanced, which may limit the generalisability of the findings to other social groups. To provide a more comprehensive understanding of the implications of the relationships found in the model, a longitudinal study could be conducted to analyze the functioning of couples over time and observe changes in the variables of interest. Future research could also explore the role as potential risk or protective factors in intimate partner violence.