Attachment style and subjective wellbeing in a sample of young and older adults, the importance of attachment pattern for satisfaction with life

Introduction

Although attachment dimensions are known to be related to wellbeing (Baird, 2002; Consedine et al., 2012; Gouin et al., 2009; La Guardia et al., 2000; Li and Zheng, 2014; Marganska et al., 2013; Marrero-Quevedo et al., 2018; Mónaco et al., 2021; Whalen, 2018), there is comparatively little research exploring the association of attachment in younger and older adulthood (Kafetsios and Sideridis, 2006). This study investigates how attachment influences life satisfaction of undergraduate university students and older adults enrolled in a continuing-education program. Adult attachment extends the infancy behavioral system into adulthood (Bowlby, 1973). In psychological research, it is a stable construct explaining how Internal Working Models (IWMs), formed through early caregiving—shape lifelong emotional regulation and interpersonal behavior throughout the lifespan (Bretherton and Munholland, 2008).

Brennan et al. (1998), and Bartholomew and Horowitz (1991) provide the theoretical framework and describe the dimensions of adult attachment styles through two orthogonal dimensions: Attachment-related anxiety (the degree to which an individual fears rejection and abandonment) and Attachment-related avoidance (the degree to which an individual feels uncomfortable with emotional intimacy). Attachment style modulates wellbeing and life satisfaction. Except for temperament, attachment style interacts more strongly in early life than any other psychological or sociocultural factor regarding subjective wellbeing (Huppert, 2009).

Wellbeing as a central aspect of mental health

Wellbeing is an individual's subjective assessment of their own life, comprising life satisfaction, frequent positive affect, and infrequent unpleasant affect (Diener, 1994). Modern mental health extends beyond the absence of disease (World Health Organization, 2007, 2010); it reflects the ability to realize individual potential, happiness, and satisfaction throughout life (Kahneman et al., 1999). This paradigm shift views wellbeing as an essential prerequisite for mental health rather than a mere absence of disorder (Diener, 1984; Ryff and Singer, 1998)—a critical perspective given today's rising mental health rates. Attachment theory through emotional regulation investigates the origins and patterns of psychological balance, providing an evolutionary perspective on wellbeing. Both concepts, wellbeing and attachment, are essential elements for understanding the genesis of life satisfaction and for addressing the challenges that mental health will present to the population in the coming decades.

Wellbeing is a complex construct that derives from two general perspectives (Ryan and Deci, 2001; Thorsteinsen and Vittersø, 2019). The hedonic approach defines wellbeing through happiness, pleasure pursuit, and suffering avoidance. Conversely, the eudaemonic perspective emphasizes meaning, self-actualization, and full functioning. This study utilizes Diener's Satisfaction With Life Scale (SWLS), a hedonic measure. According to Campbell (1976), this scale reflects both the individual's positive life assessment and the discrepancy between self-perception and aspirations.

The SWLS has been widely validated across diverse populations. These cohorts include students, nuns (McGarrahan, 1991), marines (Hindelang et al., 2004), and clinical patients (Lorenzo-Seva et al., 2019). Furthermore, researchers utilize the scale to assess wellbeing from a cross-cultural perspective (Bradshaw et al., 2021; Jang et al., 2017). Relevant studies encompass both individualistic and collectivistic societies (Ali et al., 2025; Schimmack et al., 2002; Swami et al., 2025), as well as the influence of democratic governance on wellbeing (Inglehart, 1990). In psychiatric populations, the SWLS helps analyze treatment efficacy (Arrindell et al., 2001; Friedman and Toussaint, 2006) and assess suicidal or disorder risk (Heisel and Flett, 2004). More broadly, it connects wellbeing to various psychological factors and health outcomes (Kobau et al., 2010). Most attachment research in older adulthood (Quiroga-Méndez, 2023a) and other age groups (Marganska et al., 2013) focuses on psychopathology, such as depression. These studies frequently highlight underlying emotional regulation variables (Consedine et al., 2012; Spence et al., 2020). However, very few empirical studies examine the direct relationship between attachment and wellbeing. This gap persists despite evidence that secure attachment styles define emotional regulation and life satisfaction from childhood onward (Kafetsios and Sideridis, 2006).

Attachment theory as a mechanism for emotional regulation

Bowlby (1988) identifies human attachment as an early childhood emotional regulation mechanism. This system relies on maintaining physical proximity to obtain companionship and relief. When attachment figures are sensitive and available to the child's needs, individuals develop positive world expectations and basic security. Furthermore, Bowlby (1973) termed the mental representations of self and others as “working models.” This cognitive construction guides social interaction, shapes perceptions of the environment, and influences self-efficacy. Ultimately, the resulting “secure base” serves as the most effective tool for managing anxiety and promoting wellbeing (Mikulincer and Shaver, 2016).

An individual's attachment style remains a primary resource for adult emotional regulation. This continuity occurs because early patterns of trust, mistrust, or avoidance remain stable throughout the life cycle. When secure attachment is unavailable, individuals adopt secondary emotional regulation strategies (Main, 1990). For instance, anxious insecure attachment manifests as constant worry, hypervigilance, restlessness, or demanding traits. Conversely, avoidant attachment triggers contact avoidance and compulsive self-sufficiency. Formed initially through early family interactions, these attachment orientations dictate lifelong relational and psychological functioning (Cassidy and Shaver, 2016). Ultimately, these orientations operate as mental models containing complex networks of cognitive-affective schemas, memories, and experiences projected across the lifespan (Bowlby, 1988; Mikulincer and Shaver, 2003).

From an evolutionary perspective, temperament is the sole factor preceding attachment in shaping wellbeing. This biological-genetic disposition interacts with the environment. Furthermore, it undergoes regulation through social referencing from the relationship with primary caregivers (Mónaco et al., 2021). Empirical evidence consistently links extraversion to positive affect and neuroticism to negative affect (Costa and McCrae, 1980; Diener and Biswas-Diener, 2008; Pavot and Schimmack, 2002). Twins studies suggest that 38% of responses to wellbeing are due to heredity, with the remainder attributable to environmental factors (Stubbe et al., 2005).

Attachment operates as the first psychological factor to interact with this inherited temperament, establishing critical individual differences in emotional regulation. Consequently, attachment is considered the “grand theory” of emotional regulation. This status stems from its early development, long-term stability, and emphasis on automatic, non-conscious responses. Conversely, insecure attachment styles can cause extensive psychological difficulties, significantly reducing self-perceived subjective wellbeing (Quiroga-Méndez, 2020, 2023b).

The influence of attachment on subjective wellbeing

The protective function of attachment is primary, acting as a baseline modulator for self-reported life satisfaction and wellbeing. Consequently, this childhood construct decisively influences personal perceptions before an individual even completes an assessment. Investigating the effect of attachment on wellbeing consolidates a research line rooted in early emotional regulation and its lifespan stability. Empirically, insecure attachments consistently associate with lower levels of life satisfaction (La Guardia et al., 2000; Lavy and Littman-Ovadia, 2011). Confirming the hypothesis that secure attachment promotes wellbeing would affirm the power of this initial regulatory mechanism. This outcome would support the premise that early relational experiences with attachment figures heavily influence later life satisfaction. Ultimately, achieving optimal emotional wellbeing and life satisfaction may prove difficult without establishing a secure childhood attachment (Brake et al., 2004).

Studying two distinct age groups is vital. In youth, attachment and wellbeing influence the formation of adult roles, vocational choices, and partner selection. Conversely, in old age, these concepts help buffer the numerous life stressors that strain emotional regulation mechanisms. Across both developmental stages, wellbeing enables individuals to face challenges, while attachment serves as a structural factor differentiating emotional regulation capacities.

Notably, previous studies integrated attachment and wellbeing with other variables, such as self-esteem (Landen and Wang, 2010; Li and Zheng, 2014; Wayne et al., 1997), empathy (Whalen, 2018), emotional intelligence, coping (Baird, 2002), personality (Marrero-Quevedo et al., 2018), and resilience (Karreman and Vingerhoets, 2012). To date, only one study has analyzed attachment, wellbeing, and emotional regulation as the sole research variables (Mónaco et al., 2021).

This study aims to investigate the relationships between adult attachment dimensions and subjective wellbeing within a chronologically diverse sample. To guide this investigation, we propose the following hypotheses:

• Hypothesis 1 (H1): There is a significant relationship between adult attachment dimensions and subjective emotional wellbeing across the lifespan.
• Hypothesis 2 (H2): Secure attachment is positively associated with self-perceived wellbeing.
• Hypothesis 3 (H3): Avoidant and anxious attachment styles are negatively associated with self-perceived wellbeing.
• Hypothesis 4 (H4): Sociodemographic characteristics (age, sex, marital status, and religious beliefs) account for a significant but limited proportion of variance in subjective wellbeing when considering attachment styles.

Materials and methods

Method

This study follows a cross-sectional, observational design based on non-probabilistic convenience sampling. The primary aim was to examine the association between attachment dimensions—anxiety and avoidance, as assessed by the Experiences in Close Relationships (ECR) questionnaire—and subjective wellbeing, measured using the Satisfaction With Life Scale (SWLS). Prior to inferential analyses, the distribution of attachment styles was examined across sociodemographic variables (sex, age group, marital status, and religious beliefs) to identify potential differences in attachment profiles within the sample. Subsequently, the relationship between attachment and wellbeing was analyzed using correlational and regression-based approaches.

Participants

We recruited 984 participants between April 2021 and August 2024 at Universidad Pontificia de Salamanca, Salamanca, Spain. The study was approved by the Ethics Committee of the Universidad Pontificia de Salamanca under protocol number Acta 28/10/2021, and adhered to the Declaration of Helsinki and complementary biomedical regulations.

The sample comprised undergraduate students in Computer Science and Psychology (n = 503) and older adults enrolled in the “Universidad de la Experiencia” (n = 473). The latter is a continuing education program aimed at older adults seeking to continue learning and actively participating in the cultural and academic life at Universidad Pontificia de Salamanca. The participants were young undergraduate students and older adults enrolled in a university program for seniors. All subjects who volunteered to participate in both groups were included. The participants were contacted in classroom announcements and a week later the questionnaire was presented to those who wanted to do so.

Participants met the following inclusion criteria: being adult (> 18 years old), signing the informed consent form, and being enrolled in the degrees of Computer Science or Psychology, or in the “Universidad de la Experiencia.” In summary, most participants were women (62%), the age distribution was almost balanced, and the most frequent marital status was single (58.7%), widowed (8.7%), separated (5.0%). Approximately six out of ten participants declared themselves as religious believers. These sociodemographic characteristics provide the necessary context for interpreting the subsequent comparisons of attachment and wellbeing, as presented in Table 1.

Table 1. Sociodemographic characteristics of the sample.

Percentages are calculated based on valid cases; differences from the total (100%) are due to missing data (< 3% per variable). The information presented in Table 1 is descriptive in nature and is provided exclusively to characterize the study sample. Therefore, the reported frequencies and percentages should not be interpreted as formal inferential comparisons between age groups.

Procedure

The study was carried out using a cross-sectional design with non-probabilistic convenience sampling. For the sample size calculation, recommendations for this type of study were considered, which suggested recruiting between 5 and 10 participants per item, with a minimum of 205 participants in this case due to the existence of 41 items (Argimon Pallás and Jiménez Villa, 2004).

Data collection was conducted during the academic period by a qualified researcher. All participants completed the Spanish version of the SWLS (Pons et al., 2000), and the Spanish version of the ECR questionnaire (Alonso-Arbiol et al., 2007). Participants were then administered the battery of self-report measures. These measures were completed individually by participants at the psychology practice site, using pencil and paper.

Measures

Satisfaction with life scale (SWLS)

The SWLS (Diener et al., 1985; Spanish adaptation by Atienza et al. (2000) is a five-item instrument that captures global cognitive judgements of life satisfaction. Items are rated on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). In its original validation, the SWLS showed excellent internal consistency (α = 0.87) and good test–retest reliability, as well as convergent coefficients of r = 0.50–0.75 with other subjective wellbeing measures. It has been extensively employed in health-psychology research, where scores correlate with mental-health indicators and have been linked to outcomes such as suicide attempts (Pavot and Diener, 2008). In the present sample, the SWLS achieved high reliability (α = 0.84).

Experiences in close relationships questionnaire (ECR)

The ECR (Brennan et al., 1998); Spanish adaptation (Alonso-Arbiol et al., 2007) comprises 36 items that assess two continuous attachment dimensions: Anxiety (18 items) and Avoidance (18 items). Higher scores indicate greater insecurity on each dimension. To enhance accessibility for older respondents, we used a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree) rather than the original 7-point format. The ECR-R has consistently demonstrated strong construct validity and internal consistency (typically α ≥ 0.85). In the current study, the total scale yielded α = 0.86; the Anxiety subscale showed excellent reliability (α = 0.89) and the Avoidance subscale showed acceptable reliability (α = 0.71). These indices confirm the suitability of the ECR-R for evaluating attachment styles in both younger and older Spanish adults.

For classification purposes, a median split procedure was applied to both attachment dimensions. Individual scores were computed as the mean of the corresponding items, and participants were classified as having low or high levels on each dimension based on the sample medians (avoidance: 1.67; anxiety: 1.72). Based on this classification, three dichotomous variables were created: avoidant attachment, anxious attachment, and secure attachment. Secure attachment was defined as the combination of low levels of both anxiety and avoidance. This approach allows for the identification of individuals with relatively secure relational profiles.

Statistical analysis and results

The main objective of this analysis is to examine the relationship between attachment orientations and wellbeing. To do this, we started with the two main dimensions assessed by the ECR questionnaire (anxious attachment and avoidant attachment), in addition to calculating a secure attachment profile. Wellbeing was measured using the SWLS, widely used in assessing people's overall judgment of their satisfaction with life.

Although the analyses are based on the continuous dimensions of anxious and avoidant attachment, the distinction between secure and insecure attachment is used for interpretative purposes, in accordance with established theoretical frameworks.

Before exploring this relationship, it is essential to describe the distribution of attachment styles according to sociodemographic variables such as sex, age group, marital status, and religious beliefs, in order to identify possible structural differences in attachment profiles according to these characteristics (Table 2). This provides a solid starting point for understanding how different attachment styles are distributed in the sample before examining their link to wellbeing.

Table 2. Distribution of attachment styles.

Values are presented as n (%). Percentages are calculated within the total sample. Group differences were evaluated using Pearson's chi-square tests. ^*p < 0.05.

Table 2 presents the distribution of attachment styles across sociodemographic variables. Avoidant attachment appears to show greater variability across groups, with relatively higher frequencies among older adults, men, and individuals with certain marital statuses and religious beliefs. In contrast, anxious and secure attachment styles display a more homogeneous distribution across the categories analyzed, with less pronounced differences between groups.

Chi-square tests revealed significant associations between avoidant attachment and all sociodemographic variables analyzed [sex: χ²(1) = 5.654, p = 0.017; age group: χ²(1) = 27.233, p < 0.001; marital status: χ²(3) = 15.893, p = 0.001; religious beliefs: χ²(1) = 7.129, p = 0.008]. This style was more prevalent among older adults (27.5 vs. 20.6% in young people), married and widowed individuals (15.0 and 5.0%, respectively), religious believers (30.9%), and men (20.0%). In contrast, anxious attachment and secure attachment showed homogeneous distributions across the sociodemographic variables considered, with no statistically significant associations (all p > 0.05). These chi-square results are presented in Table 2, while the complete descriptive statistics for these variables can be found in Table 3A.

Table 3A. Descriptive statistics of sociodemographic variables and attachment styles.

The means confirmed the previous pattern: avoidant attachment was higher in men (M = 1.72, SD = 0.53), older adults (M = 1.73, SD = 0.50), married/separated individuals, and believers. The differences between subgroups in anxious attachment were minimal, although the highest scores appeared in women, young people, and non-believers, as further supported by the age-group comparisons presented in Table 3B.

Table 3B. Age-group comparisons of attachment dimensions.

Differences between age groups were evaluated using independent-samples t-tests. For avoidant attachment, Welch's t-test was reported because the assumption of homogeneity of variances was violated (Levene's test: p < 0.001). Cohen's d represents the standardized effect size.

The average SWLS scores ranged from 3.13—observed in separated individuals—to 3.57—in married individuals—, indicating a moderately high level of life satisfaction in the sample as a whole. Descriptive analyses revealed (a) a non-significant tendency toward higher wellbeing among younger participants (M = 3.54, SD = 0.86) compared to older participants (M = 3.41, SD = 0.88); (b) higher mean scores among married participants compared to single, widowed, and especially separated individuals, although not all differences reached statistical significance and (c) minimal differences according to sex and religious affiliation (Table 4).

Table 4. Descriptive statistics of sociodemographic variables and subjective well being.

Values are presented as mean, standard deviation (SD), and median of the SWLS. Differences between two groups (sex, age group, and religious beliefs) were evaluated using independent-samples t-tests. Differences according to marital status were evaluated using one-way analysis of variance (ANOVA). η² represents the effect size for the ANOVA model. ^*p < 0.05.

Although the SWLS is typically interpreted using a total score, item-level analyses were conducted to explore whether specific components of life satisfaction show differential associations with sociodemographic variables, thereby providing a more detailed characterization of subjective wellbeing. Results based on the total SWLS score are reported in subsequent analyses.

The Spearman coefficients between the sociodemographic variables and the five SWLS items were, in all cases, of low magnitude (|ρ| ≤ 0.17). Age was negatively correlated with items B2 (“My living conditions are close to my ideal”) and B5 (“If I could live my life over again, I would change almost nothing”). Marital status showed a single negative correlation with item B2, while religious beliefs were negatively associated with items B3 (“I am satisfied with my life”) and B4 (“So far, I have obtained the important things I want in life”). No significant relationship was observed between sex and any of the SWLS items.

These descriptive trends were corroborated by inferential tests. Comparison of means using the t-test indicated that younger participants ( ≤ 30 years) reported significantly higher levels of wellbeing (M = 3.54, SD = 0.86) than older participants (≥ 60 years; M = 3.41, SD = 0.88), p = 0.026. With regard to marital status, the ANOVA analysis was significant, p = 0.010, η² = 0.01; Tukey's contrast showed that married people scored higher than separated people (p = 0.012). No statistically significant differences were found based on sex (p = 0.753) or religious beliefs (p = 0.316).

Regarding attachment styles, avoidant attachment was negatively associated with all five items of the SWLS (ρ = −0.15 to −0.24, ps < 0.001), while anxious attachment showed weaker negative correlations (ρ = −0.08 to −0.14, ps ≤ 0.012); In contrast, secure attachment was positively related to life satisfaction (ρ = 0.13 to 0.23, ps < 0.001). Mean comparisons confirmed this pattern. Participants with avoidant attachment reported significantly lower wellbeing (M = 3.28) than those without avoidant attachment (M = 3.67), based on independent samples t-tests (p < 0.001). Similarly, participants with anxious attachment showed lower levels of wellbeing (M = 3.36) compared to those without anxious attachment (M = 3.61), according to independent samples t-tests (p < 0.001). In contrast, participants with secure attachment reported significantly higher wellbeing (M = 3.81) than those without secure attachment (M = 3.36), also based on independent samples t-tests (p < 0.001). Taken together, these results confirm that insecure attachment styles—especially avoidance—are linked to lower levels of life satisfaction, while secure attachment acts as a clear protective factor.

To estimate the incremental contribution of attachment styles and sociodemographic variables to subjective wellbeing, a three-step hierarchical linear regression was performed The results of these models are presented in Table 5 show that sociodemographic variables initially account for 1.0% of the variance in life satisfaction (R² = 0.010). The inclusion of avoidant and anxious attachment significantly increased the explained variance by an additional 6.7% (ΔR² = 0.067; ΔF = 34.400; p < 0.001). Conversely, the subsequent addition of secure attachment only increased the explained variance by 0.3% (ΔR² = 0.003), failing to reach statistical significance (ΔF = 2.589; p = 0.108).

Table 5. Hierarchical regression models predicting subjective well being.

β = standardized coefficients. Dependent variable: life satisfaction (Media_SWL). ΔR² represents the incremental variance explained relative to the preceding model. ΔF represents the significance test associated with the change in R². ^*p < 0.05; ^***p < 0.001; ns = not significant.

These findings indicate that attachment dimensions, particularly avoidant and anxious attachment, provide a significant incremental contribution to the explanation of subjective wellbeing after controlling for sociodemographic factors. Consequently, the revised model structure offers a more appropriate assessment of the incremental validity of attachment, yielding a theoretically more robust interpretation of its contribution to subjective wellbeing.

In Model 1, the sociodemographic variables (age group, marital status, and religious beliefs) were entered as predictors of life satisfaction. This initial model explained 1.0% of the variance in subjective wellbeing (R² = 0.010; adjusted R² = 0.007). Among these variables, only religious beliefs reached statistical significance (β = −0.071, p < 0.05), whereas age group and marital status were not significantly associated with life satisfaction.

In Model 2, the insecure attachment dimensions (avoidant attachment and anxious attachment) were added to the sociodemographic variables. This resulted in a significant increase in explained variance (ΔR² = 0.067; ΔF = 34.400, p < 0.001), raising the total variance explained to 7.7% (R² = 0.077; adjusted R² = 0.072). Both avoidant attachment (β = −0.211, p < 0.001) and anxious attachment (β = −0.138, p < 0.001) emerged as significant negative predictors of subjective wellbeing, whereas the effect of religious beliefs remained significant.

Finally, Model 3 incorporated secure attachment. This addition produced only a minimal increase in explained variance (ΔR² = 0.003; ΔF = 2.589, p = 0.108), resulting in a final R² of 0.080 (adjusted R² = 0.074). Secure attachment did not make a statistically significant contribution to the model, nor did age group or marital status. Religious beliefs remained a significant predictor (β = −0.078, p = 0.021), indicating that non-believers reported slightly lower levels of subjective wellbeing compared to believers. Overall, the findings indicate that insecure attachment dimensions, particularly avoidant attachment, account for the largest proportion of explained variance in life satisfaction after controlling for sociodemographic characteristics.

In summary, the regression models indicate that insecure attachment styles are the variables most strongly associated with life satisfaction in this sample. Avoidant attachment showed the strongest negative association with subjective wellbeing (β = −0.211), followed by anxious attachment (β = −0.138), whereas secure attachment did not contribute significantly once insecure attachment dimensions were included in the model. The contribution of sociodemographic variables was comparatively small; among them, only religiosity remained a significant predictor (β = −0.078), indicating slightly lower levels of subjective wellbeing among non-believers compared to believers. Neither age group nor marital status showed significant effects. Thus, these findings indicate that insecure attachment styles—particularly avoidant attachment—are the primary factors associated with subjective wellbeing in the present sample.

Discussion

The results of this study suggest that attachment style is an important factor associated with life satisfaction and wellbeing, consistent with theoretical perspectives that highlight the role of emotional regulation processes across the life span. As expected, insecure styles were negatively associated with SWLS, while secure attachment showed the expected positive relationship, aligning with previous empirical evidence (Kafetsios and Sideridis, 2006; Karreman and Vingerhoets, 2012; La Guardia et al., 2000; Lavy and Littman-Ovadia, 2011).

Avoidant attachment showed the strongest association with lower levels of wellbeing. This finding suggests that persistent suppression of proximity seeking—a distinctive feature of avoidance—not only involves physiological stress responses such as increased IL-6 and decreased NK activity (Gouin et al., 2009; Picardi et al., 2007), but also a more negative subjective assessment of life. This pattern contrasts with Mónaco et al. (2021), who did not observe decreased wellbeing in avoidant individuals, suggesting that emotional deactivation and reduced emotional awareness inherent to avoidance may skew self-reported outcomes depending on the contextual characteristics of the sample.

Anxious attachment was also negatively associated with wellbeing, although its relative impact was more moderate than that of avoidant attachment. The hypervigilance, emotional dysregulation, and persistent dependence on relational approval that characterize anxious attachment likely explain its significant negative association with life satisfaction.

Importantly, any observed variations in attachment scores or wellbeing levels across different age cohorts within this study must be interpreted with strict caution. While descriptive data showed slight fluctuations in satisfaction levels among older and younger participants with anxious or avoidant tendencies, these trends did not reach statistical significance. Crucially, this study was not designed to test the formal moderating effects of age, nor did it perform statistical interaction analyses between age groups and attachment styles. Consequently, these descriptive patterns should not be interpreted as verified cohort differences, but rather as exploratory trends that warrant future longitudinal investigation with larger, stratified samples.

Secure attachment showed a positive association with life satisfaction; however, its statistical contribution diminished considerably when controlling for insecure dimensions in the hierarchical regression models, replicating prior findings (Karreman and Vingerhoets, 2012). This suggests that variability in subjective wellbeing within adult populations may be more robustly explained by the presence and severity of insecure attachment strategies (avoidance and anxiety) than by the presence of secure attachment per se.

Sociodemographic variables provided minor additional context but demonstrated very limited predictive power once attachment dimensions were controlled for. Marital status was descriptively related to higher satisfaction scores among married participants, but this effect did not remain significant in multivariate models. Regarding religious beliefs, the analysis revealed a small but significant effect (β = −0.08); this negative coefficient explicitly indicates that non-believers reported slightly lower levels of wellbeing compared to believers. Neither sex nor age were significantly associated with life satisfaction in the final models. These findings are consistent with previous literature suggesting relatively small or inconsistent direct effects of sociodemographic factors on subjective wellbeing (Blanchflower and Oswald, 2008). Nevertheless, the inclusion of older adults who were actively enrolled in university programs represents a specific sample characteristic that may limit the generalizability of these sociodemographic findings to the broader elderly population. The overall proportion of explained variance in the hierarchical regression models was modest (approximately 7%), underscoring that attachment is only one of multiple psychological components underlying subjective wellbeing. A substantial portion of variance is likely accounted for by unmeasured variables, such as stable personality traits (e.g., neuroticism, extraversion), physical and mental health status, or socioeconomic conditions. Therefore, these findings must be contextualized within a comprehensive, multifactorial framework of human wellbeing.

Overall, the data are consistent with theoretical perspectives suggesting a convergence between the emotional regulation described by hedonic psychology (Quoidbach et al., 2010) and the attachment strategies proposed by Bowlby, and subsequent developments (Mikulincer et al., 2003). Previous studies, like this one, find evidence that individuals differ in psychological wellbeing depending on their attachment style (Karreman and Vingerhoets, 2012). Early emotional security has been proposed as a relevant factor in emotional functioning, potentially contributing to better stress regulation and more positive evaluations of life. In contrast, higher levels of avoidance and anxiety have been associated with lower levels of wellbeing. Thus, insecure attachment—especially avoidance—appears to show a stronger association with life satisfaction than age, sex, marital status, and religiosity within this sample. This highlights the robust link between attachment-related processes and wellbeing when examined across a highly age-diverse population, independent of variations in age itself.

Limitations

This study is not without limitations, among which we can highlight the use of self-perception instruments, with the corresponding biases inherent in self-administered questionnaires. Future studies should incorporate methods that provide greater objectivity in the recording of attachment and wellbeing. The ECR scale was modified from its original 7-point Likert format to a 5-point format to facilitate comprehension among elderly participants. Questionnaire completion can represent a cognitively demanding task in older populations, and response quality may decrease when survey formats are complex (Schneider et al., 2022). Previous research on aging populations also highlights the importance of simplifying response formats to improve data quality and reduce cognitive burden (Schneider et al., 2024). Psychometric research indicates that reducing the number of response categories from seven to five has minimal impact on reliability or validity (Matell and Jacoby, 1971; Preston and Colman, 2000), although this modification should be considered when replicating or comparing these findings with studies using the original version.

Crucially, while our sample encompasses a wide age range, our analytical approach treated age as a main effect covariate rather than a moderator. Consequently, our design demonstrates that the attachment-wellbeing link holds after controlling for age, but it does not formally test whether the strength or direction of this relationship varies across different age cohorts. Future research should employ formal moderation or multigroup analyses to explore potential interactive effects.

We also consider it important to introduce longitudinal studies to analyze variables over time, as well as to expand the sample to generalize these findings from a cross-cultural perspective. Finally, although the sample is large, it comes from a single academic and cultural context, which limits the generalizability of the findings. It will be necessary to replicate this study with probability samples from different sociocultural settings to verify the sociological validity of the conclusions obtained.

Conclusions

In summary, the findings of this study suggest that attachment dimensions, particularly avoidance, are associated with subjective wellbeing, whereas sociodemographic variables appear to play a more limited role. These results align with theoretical perspectives emphasizing the importance of attachment processes in emotional functioning and life satisfaction.

Nevertheless, the proportion of explained variance was modest, and the cross-sectional design limits causal interpretations. Future research using longitudinal approaches would be necessary to clarify the directionality and mechanisms underlying these associations.

Overall, the findings highlight the potential relevance of incorporating attachment-related processes into the understanding of wellbeing, demonstrating that this relationship remains robust and meaningful when examined within an age-diverse population.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Matthew Cole, Lawrence Technological University, United States

Reviewed by: Sunay Güngör, Gumushane University, Türkiye

Anggi Mayangsari, Jakarta State University, Indonesia