Not All Loneliness Is Equal: Differential Associations of Romantic, Familial, and Social Loneliness with Mental Health in Later Life

1. Introduction

Loneliness has been rising exponentially worldwide and has even been described as an “epidemic” (Ng, 2024; Office of the Surgeon General, 2023). One major concern underlying this increase is its harmful impact on human health, as loneliness is associated with a 26% higher risk of mortality (Holt-Lunstad et al., 2015). This elevated risk is linked to multiple factors, including poorer mental health outcomes (Ng, 2024). Meta-analytic evidence indicates that individuals experiencing loneliness face an increased risk of mental health problems, most notably depression, but also anxiety and self-harm (Mann et al., 2022). The prevalence of loneliness follows a “U-shaped” distribution, with higher levels observed among both younger and older adults (Lasgaard et al., 2016). Given that the global population is aging (United Nations, 2024), loneliness among older adults is becoming an increasingly important public health concern. Given these demographic shifts, it is critical to understand how loneliness affects mental health in older populations.

Loneliness is described as the unpleasant experience that arises when there is a discrepancy between the desired and the actual state of a person’s social relations, either in terms of quantity or quality (Peplau & Perlman, 1982). While temporary feelings of loneliness can serve as an adaptive human response, chronic loneliness can be harmful with both immediate and long-term mental and physical health consequences (Cacioppo et al., 2011; Holt-Lunstad, 2024; Wang et al., 2018). Loneliness is increasingly prevalent among the elderly population worldwide (Surkalim et al., 2022). A systematic review and meta-analysis reported that in high-income countries, approximately 25.9% of older adults experience moderate loneliness, while 7.9% experience severe loneliness (Chawla et al., 2021).

Conceptually, loneliness has been framed as both a unidimensional and multidimensional construct. From the unidimensional perspective, loneliness is viewed as a single continuum of intensity, often linked to relationship failures (Russell, 1996). In contrast, the multidimensional perspective suggests that loneliness experienced in different dimensions of relationships can differ qualitatively, though they all stem from the same underlying construct. Weiss (1973) distinguished between social loneliness—the absence of broader affiliations such as friendships and workplace ties—and emotional loneliness, which arises from unmet attachment needs. Emotional loneliness can further be divided into family loneliness, associated with the absence of close family connections, and romantic loneliness, linked to intimate romantic connections (DiTommaso & Spinner, 1997). Recent research supports the multidimensional nature of loneliness, showing that its distinct dimensions—social, family, and romantic—are linked to different relational and psychological correlates (DiTommaso et al., 2004; Giletta et al., 2012; Thompson & Pollet, 2024). Notably, each dimension is closely tied to mental health outcomes, underscoring the need to consider these nuances when studying loneliness.

It is increasingly recognized that loneliness is strongly associated with negative mental health outcomes in older adults (Leigh-Hunt et al., 2017), with anxiety and depression being among the most common problems in this age group (Domènech-Abella et al., 2019; Jalali et al., 2024). A meta-analysis estimated that the global prevalence of depression among older adults is 19.2%, while the prevalence of anxiety disorders and stress is 16.5% and 13.9%, respectively (Jalali et al., 2024). These problems may be exacerbated by limited support systems, as loneliness can amplify stress responses and weaken coping capacities (Steptoe et al., 2004). Some studies suggest that emotional forms of loneliness may be especially detrimental to mental health because close relationships serve important functions related to security, emotional regulation, and support across the lifespan (Peerenboom et al., 2015). In early life, this role is fulfilled by parental attachment figures (Bowlby, 1969), but across development, the primary attachment figure often shifts to a romantic partner in young adulthood and typically remains so through midlife and later life (Hazan & Shaver, 1987). This shift becomes particularly salient in older age, when access to parental support is diminished due to advanced age or loss (Luanaigh & Lawlor, 2008). In this context, spousal or romantic relationships can take on heightened importance as a primary source of emotional security. Consistent with this perspective, a 19-year longitudinal study found that emotional loneliness increased the risk of all-cause mortality in older adults living alone (O'Súilleabháin et al., 2019).

At the same time, there are reasons to expect broader social relationships to play an increasingly important role in psychological well-being during later life. Aging is often accompanied by shrinking social networks, reduced mobility, retirement, and diminished participation in community activities (Puyané et al. 2025). These changes may increase the importance of maintaining meaningful social connections beyond the family (Dykstra & Fokkema, 2007; Tiikkainen & Heikkinen, 2005). Consequently, social loneliness may represent a distinct source of vulnerability for older adults, even when close family or romantic relationships are available. Taken together, these perspectives suggest that different dimensions of loneliness may show distinct associations with mental health outcomes, but the relative importance of romantic, familial, and social loneliness remains unclear.

Despite the growing literature on loneliness and mental health, most research has been conducted in WEIRD (Western, Educated, Industrialized, Rich, and Democratic) populations (Mann et al., 2022). Cultural contexts may shape how loneliness is experienced and managed, highlighting the need for research in underrepresented populations. Turkey provides a particularly relevant context, as its population is rapidly aging (Turkish Statistical Institute, 2023), yet research on loneliness and mental health among older adults remains limited. Existing evidence indicates that loneliness is substantial in this population, with one study reporting that 46.9% of Turkish older adults experience moderate to severe loneliness (Öztürk Haney et al., 2017).

The present study examines the association between loneliness and mental health outcomes in married Turkish middle-aged and older adults (≥50 years). Loneliness is examined multidimensionally through social, familial, and romantic loneliness, while mental health outcomes include depression, anxiety, and stress. We hypothesize that higher levels of loneliness will be associated with greater psychological distress. Furthermore, guided by the attachment-based models (Bowlby, 1969; Mikulincer & Shaver, 2019) and the previous findings (O'Súilleabháin et al., 2019), we expected romantic and familial loneliness to show stronger associations with depression, anxiety, and stress than social loneliness. Focusing on married individuals provides a useful context for examining whether perceived deficits in romantic relationships remain more strongly associated with mental health than deficits in family or broader social relationships when a spouse is present.

The present study also distinguished between loneliness and social isolation. Whereas social isolation refers to the objective absence of social contacts, loneliness reflects the subjective perception that one's social relationships are insufficient (Cacioppo et al., 2014; Leigh-Hunt et al., 2017). Because loneliness can occur even in the presence of social ties (Luanaigh & Lawlor, 2008), social network size was included as a control variable to isolate the unique contribution of loneliness.

2. Method

2.1. Participants & Procedure

The sample included 484 married adults (223 women) aged 50 years and older (M = 58.54, SD = 6.39) residing in Istanbul, Turkey. The data was collected in July and August 2024 as part of a larger study on the connection of neighborhood, personality, and health. The study also included questions on demographics, mental health, loneliness, and social network. Participant selection was done by stratified random sampling method. It was aimed to reach a representative sample of Istanbul, Turkey according to municipalities and gender. To improve representativeness, more participants were recruited from larger municipalities, while fewer were selected from smaller ones. The eligibility criterion was to be aged 50 years old or above. The target sample size was 650. Participants were reached in their homes, by the survey interviewers. Participating in the study was voluntary. Before beginning the survey, participants were informed about the study. After participants provided their consent, face-to-face interviews were conducted. The interviewer asked the questions to the participants and noted verbal responses the participants provided. The interviews lasted approximately 20 minutes. The study was approved by the Intuitional Review Board of the authors university (Date: 08.05.2024, Reference Number: 1858).

2.2. Measures

2.2.1. Mental Health

The Turkish version of the short-form of the Depression Anxiety Stress Scale (Henry & Crawford, 2005; Yılmaz et al., 2017) was used to assess depression (e.g., “I felt that I had nothing to look forward to”; α = .86), anxiety (e.g., “I felt I was close to panic”; α = .86), and stress (e.g., “I found myself getting agitated”; α = .82). The scale includes a total of twenty-one items, each subscale consisting of 7-items. Participants rated their experiences over the past week using a scale ranging from 0 (did not apply to me) to 3 (applied to me very much). Mean scores were computed for each sub-scale ranging from 0 to 3. Higher scores indicated greater levels of depression, anxiety, and stress.

2.2.2. Loneliness

The Turkish version of the short-form of the Social and Emotional Loneliness Scale for Adults (Akgül, 2020; DiTommaso et al. 2004;) was used to assess social loneliness (e.g., “I do not have any friends who understand me, but I wish I did”; α = .86), familial loneliness (e.g., “I feel alone when I am with my family”; α = .87), and romantic loneliness (e.g., “I wish I had a more satisfying romantic relationship”; α = .89). The scale includes a total of fifteen items, each subscale consisting of five items. Participants rated their overall experiences using a scale ranging from 1 (strongly disagree) to 5 (strongly agree). Mean scores were computed for each sub-scale ranging from 1 to 5. Higher scores indicated greater levels of social, familial, and romantic loneliness.

2.2.3. Social Network Composition

Social network size was assessed using four adapted items from the Health and Retirement Study (Smith et al., 2017), capturing the presence of immediate family, friends, and living children. Affirmative responses were summed to create a social network size score.

2.2.4. Demographics

Participants reported their age, gender, relationship status, education level, working status, monthly family income, and subjective socioeconomic status (SES).

3. Results

Descriptive statistics and bivariate correlations for age, gender, social network, loneliness subscales (i.e., social, familial, and romantic), and mental health indicators (depression, anxiety, and stress) are presented in Table 1. Pearson correlation analyses revealed significant positive associations between the three loneliness subscales and the three mental health indicators. Higher loneliness scores were associated with higher levels of depression, anxiety, and stress, with effect sizes ranging from medium to large (r’s = .43 – .55).

To examine the unique associations between loneliness subscales and mental health indicators, a multivariate multiple regression was conducted within a structural equation modeling framework (see Figure 1 for the conceptual model). Loneliness subscales were entered as the primary predictor variables, while age, gender, and social network size were included as covariates. Mental health indicators were entered as outcome variables. Because the model was just-identified (df = 0), overall model fit indices were not informative. The analysis was conducted using the R package “lavaan” (Rosseel, 2012).

Social loneliness was significantly associated with depression (β = .29, p < .001), anxiety (β = .20, p = .001), and stress (β = .35, p < .001). Familial loneliness was significantly associated with depression (β = .27, p = .001) and anxiety (β = .19, p = .017), but not with stress (β = .09, p = .271). Romantic loneliness was not significantly associated with depression (β = .04, p = .638), anxiety (β = .15, p = .067), or stress (β = .06, p = .468).

Among the control variables, age was positively associated with depression (β = .09, p = .010), anxiety (β = .10, p = .004), and stress (β = .09, p = .011); gender was associated with anxiety (β = −.18, p < .001); and social network size was associated with stress (β = .06, p = .044). The model explained 33% of the variance in depression and anxiety and 26% in stress.

4. Discussion

The present study investigated the associations between multidimensional loneliness and mental health outcomes among married middle-aged and older adults in Turkey. Consistent with previous research across diverse cultural contexts (Dahlberg et al., 2022), the findings indicate that loneliness is an important correlate of mental health in the aging Turkish population, explaining a substantial proportion of variance in depression, anxiety, and stress. Bivariate correlations showed that higher levels of social, familial, and romantic loneliness were each associated with greater depression, anxiety, and stress, with large effect sizes. To further examine the unique contributions of each dimension, a multivariate multiple regression analysis was conducted, controlling for their shared variance. Contrary to expectations derived from the attachment framework, more proximal relationship domains did not show stronger associations with mental health outcomes. Instead, social loneliness emerged as the most consistent and robust predictor across outcomes, even after accounting for social network size.

Unexpectedly, romantic loneliness was not significantly associated with any mental health outcomes when other dimensions of loneliness were taken into account. This diverges from the theoretical expectations that the closest relational context exerts the strongest psychological impact— namely, that humans are fundamentally oriented toward emotional connection, and that close relationships serve as primary sources of security, support, and affect regulation (Bowlby, 1969; Mikulincer & Shaver, 2019).

One possible explanation is that, within a sample of married individuals, romantic relationships may be relatively stable and basic attachment needs are at least partially met, resulting in lower variability and reduced explanatory power for mental health outcomes. Under such circumstances, differences in psychological well-being may depend less on the availability of a romantic attachment figure and more on the quality and availability of broader social connections. In midlife and older adulthood, friendships and wider social networks may become increasingly important sources of companionship, support, and social engagement, and have been shown to contribute substantially to emotional well-being and mental health (Ng et al., 2021). Consistent with this perspective, social loneliness emerged as the strongest and most consistent correlate of depression, anxiety, and stress in the present study. This finding aligns with previous research linking social support and friendship quality to psychological well-being in later life (Santini et al., 2020).

The prominence of social loneliness is particularly noteworthy given that the study was conducted in Turkey, where family relationships are often viewed as a central source of social support and belonging (Ayçiçeği-Dinn & Caldwell-Harris, 2013; Kalaycıoğlu & Rittersberger-Tılıç, 2001). Despite the availability and cultural importance of family ties, broader social connections emerged as the strongest correlate of depression, anxiety, and stress. This finding suggests that meaningful engagement beyond the family may play an important role in psychological well-being during later life. Friendships and wider social networks may provide opportunities for companionship, reciprocity, and social participation that are not fully captured by family and marital relationships alone. They may also offer a distinct space for autonomy and identity expression, allowing individuals to engage in social roles and relationships that extend beyond family-based expectations and responsibilities (Adams & Blieszner, 1995). When these social connections are absent, older adults may lose vital outlets for self-expression and mutual support, potentially exacerbating feelings of isolation and diminishing self-worth. Taken together, these findings suggest that interventions in Turkey should not only strengthen family and marital ties but also foster opportunities for meaningful friendships, which may serve as vital protective resources in later life. Future research should incorporate cross-cultural comparisons to examine how cultural norms shape experiences of different forms of loneliness, as these patterns may vary across societies.

This study contributes to the growing literature on multidimensional loneliness by demonstrating that social, familial, and romantic loneliness suggest distinct associations with depression, anxiety, and stress in later life. By examining these relationships in a Turkish sample, it also extends existing evidence beyond the predominantly WEIRD populations in which much of the loneliness literature has been developed. Future research should investigate the mechanisms through which social loneliness influences psychological well-being, including the roles of social support quality, coping resources, and resilience. A key strength of the study is its representative urban sample from Istanbul, enhancing the generalizability of the findings to urban Turkey. However, the findings may not extend to rural populations, where community structures and family dynamics may differ (Sunar & Fişek, 2005). In addition, the cross-sectional design precludes causal inferences. Longitudinal studies are needed to clarify the directionality of associations between loneliness dimensions and mental health.

5. Conclusion

Overall, the findings suggest that loneliness should not be treated as a unitary construct in research on aging and mental health. Among married middle-aged and older adults, social loneliness emerged as a more consistent correlate of depression, anxiety, and stress than either familial or romantic loneliness. These findings highlight the importance of maintaining broader social integration in later life and suggest that the presence of a spouse alone may not be sufficient to protect against psychological distress when meaningful social connections beyond the family are lacking.