Life Events and Loneliness Among Older Women of Diverse Sexual Identities: Application of the Stress Process Model

Abstract

Loneliness among older adults has been a topic of interest in recent years. We analyse four waves of data from the Australian Longitudinal Study on Women’s Health. We estimate growth models to examine differences in loneliness trajectories from age 50 for women who identify as exclusively heterosexual, plurisexual (bisexual, mainly heterosexual, mainly lesbian) and exclusively lesbian. We find no significant differences in levels of loneliness across sexual identity groups at age 50. We find that while loneliness trajectories of exclusively heterosexual women trend down, levels of loneliness increase with age among plurisexual women. Adverse health events and relationship problems increase loneliness to a greater extent among plurisexual women compared to exclusively heterosexual and exclusively lesbian women. Our results suggest older lesbian women may have accumulated social or personal resources or developed coping mechanisms over the courses of their lives, while plurisexual women remain a vulnerable group.

Keywords

life events loneliness Australia women sexual identity

Introduction

Defined and theorised as the discrepancy between actual and desired social relationships, in quality or quantity (Hawkley & Cacioppo, 2010; Perlman, 2004; Weiss, 1987), loneliness increases the risk of adverse physical and mental health outcomes for older adults (Cohen-Mansfield et al., 2016; Holt-Lunstad et al., 2015; Leigh-Hunt et al., 2017; Luo et al., 2012; Perlman & Peplau, 1984; Rico-Uribe et al., 2018). Identifying the risk factors associated with loneliness among older adults is therefore an important endeavour. Such evidence can be used to inform interventions and ensure that they are targeted to those who are most at risk, at the times when they are most vulnerable. The World Health Organisation (WHO, 2021) identifies loneliness as a matter of particular concern for older women given their increased likelihood of experiencing stressful life events such as widowhood and residential relocation. Importantly, the stress process model posits that some women will be more vulnerable to the negative impacts of these events due to both contextual and intrapersonal factors (Pudrovska et al., 2005; van Gundy, 2002).

In this paper, we argue that sexual minority (i.e. lesbian, bisexual, mostly heterosexual and mostly lesbian) women are more likely to experience an increase in loneliness following stressful life events for three reasons relevant to the stress process model. First, lesbian/gay and bisexual adults report receiving less affectionate and tangible social support on average than their heterosexual peers (Kahle et al., 2020; Stinchcombe et al., 2020). Second, sexual minority women are more likely than heterosexual women to suffer poor outcomes in socioeconomic domains including education, employment and financial wellbeing (Badgett, 2018; Charlton et al., 2018; Mollborn & Everett, 2015). This applies predominantly to bisexual and other plursisexual women (i.e. women whose sexuality is oriented to more than one gender, such as mostly heterosexual and mostly lesbian women, in contrast to exclusively heterosexual or exclusively lesbian women). Third, exposure to stigma and discrimination can harm the self-esteem and self-mastery of sexual minority persons (Mereish & Goldstein, 2020; Mereish et al., 2022). Consistent with the stress process model, these inequalities in social, socioeconomic and intrapersonal resources could in turn reduce the capacities of sexual minority women to weather the impacts of stressful life events relative to heterosexual women, increasing their propensities for loneliness. To test this proposition, we analyse four waves of data from a national probability sample of Australian women aged 50 years and over.

Life Events and Loneliness

One important contributor to loneliness among older people is the experience of stressful life events. As people age, they are more likely to experience events such as widowhood, the onset of illness or an injury, being made redundant and the death of siblings or friends. There is a robust body of evidence demonstrating that the risk of feeling lonely increases following these events, often due to the disruption of social networks and ties (King et al., 2021; Shin et al., 2020; Wright et al., 2020). For example, a recent examination of people’s loneliness trajectories surrounding an experience of widowhood found a long-lasting increase in loneliness post-event not observed in the propensity-score matched control group (Bueker et al., 2020).

The Stress Process Model And Sexual Minority Women

According to the stress process model (Pearlin et al., 1981; Pearlin, 2010), an individual’s resilience in the face of stressful life events depends on factors including their levels of pre-existing chronic stress, self-concepts and available social supports. One group of older women that may be especially vulnerable to experiencing loneliness following stressful life events are sexual minority women. Older lesbian and bisexual women lived much of their early lives in a sociocultural context marked by open hostility towards people of diverse sexualities. While the lesbian, gay and bisexual (LGB) community is achieving growing visibility and support in countries such as Australia and the U.S., its members continue to face significant discrimination (Lyons et al., 2021). As we set forth below, the specific challenges older LGB people have faced throughout their lives arguably increases their vulnerability within three domains relevant to the stress process model: social support, intrapersonal resources such as self-esteem and self-mastery and socioeconomic status (Eres et al., 2021; Hughes, 2018; Pearlin, 2010).

One reason why older sexual minority women might be more vulnerable to loneliness in the face of stressful life events is a lack of social support. Being a sexually minoritized person in a heteronormative society can negatively impact the quality and quantity of a person’s social support network. Stigma, exclusion and rejection can be a common experience for sexual minority people across social contexts, including the family of origin (Carasthatis et al., 2017). While much research has focused on the rejection of LGB youth by their families upon ‘coming out’, there is evidence that family estrangement can persist across the life course. In the U.S., LGB people in older adulthood are less likely to have family support, and more likely to depend upon friend-oriented social networks (Hsieh & Wong, 2020). A study originating in the UK found that they lived at a greater distance from their mothers compared to their heterosexual peers (Green, 2016), while an analysis of an Australian sample found that LGB adults also lived further away from and had less frequent contact with their siblings (Perales & Plage, 2020). Of relevance to our study, unsupportive or estranged relationships with parents and siblings appear to be a risk factor for loneliness in older age. Among a sample of ever-widowed older adults, emotional support from siblings, positive memories of childhood relationships and current family ties were all negatively correlated with loneliness (Merz & de Jong Gierveld, 2016).

Other key sources of social support in older age are partners and children (Chen & Feeley, 2014), and LGB people may again be disadvantaged on this front. Legislation (past and present) has precluded LGB people from accessing reproductive technologies, surrogacy or adoption. Meanwhile, higher anticipated stigma upon becoming parents has been linked to lowered desires and expectations to have children among gay and lesbian people relative to heterosexual people (Gato et al., 2020; Leal et al., 2019; Shenkman, 2021). Given these barriers to parenthood, it is unsurprising that LGB older adults are more likely than heterosexual older adults to be childless (Green, 2016). Older LGB adults are also less likely to be in a relationship or, if in a relationship, to be living with their partner (Eres et al., 2021). While LGB adults foster friend-oriented social networks (families of choice), evidence of a compensatory effect is unclear. A recent literature review found that although LGB people had larger social networks, they had fewer friends living locally, and perceived less available support (Fish & Weis, 2019). This is consistent with the results of a recent study of Canadians aged 45 years and older: lesbian/gay and bisexual adults perceived less affectionate and tangible social support than heterosexual adults (Stinchcombe et al., 2020).

Lifetime experiences of stigma and discrimination may have deleterious effects on lesbian and bisexual women’s intrapersonal resources such as their personal mastery and self-esteem, thereby increasing their risk of loneliness following stressful life events (Pearlin et al., 1981). According to the minority stress model, LGB people are exposed to unique stressors in heteronormative societies including discrimination, victimisation, internalised stigma, identity concealment and expectations of rejection (Meyer, 2003). Exposure to these minority stressors is associated with decreases in self-esteem and sense of personal mastery (Mereish & Goldstein, 2020; Mereish et al., 2022).

Importantly, bisexual and other plurisexual people are exposed to double discrimination (being stereotyped and rejected by both heterosexual and lesbian/gay communities) and the invisibility and erasure of bisexual identities (Doan Van et al., 2019). Negative stereotypes of bisexual women include that they are confused, unstable, irresponsible, promiscuous and automatically consenting to sexual activity (Dyar et al., 2019; Flanders et al., 2019). In a national probability sample in the U.S., almost one in five people agreed (somewhat to strongly) with the statements that ‘Bisexual women would have sex with just about anyone’ and ‘I think bisexuality is just a phase for women’ (Dodge et al., 2016). Collectively, these stereotypes frame bisexual and other plurisexual women as undesirable partners to be in a relationship with and create barriers to finding acceptance and support within sexual minority communities (Dyar et al., 2019; Doan Van et al., 2019). In addition, this bisexual-specific stigma has been directly linked to increased lifetime experiences of sexual violence, internalised bi-negativity and sexual identity uncertainty among bisexual and other plurisexual women (Dyar & London, 2018; Flanders et al., 2019). It is highly plausible that the added stigma and minority stressors that plurisexual women experience erodes their interpersonal resources, such as their self-esteem and self-mastery, increasing their vulnerability to feeling lonely in the wake of stressful life events.

Finally, labour market discrimination (Drydakis, 2015; Mize, 2016), along with the negative impacts of minority stress on physical and mental health (Flentje et al., 2020; Pitoňák, 2017), could conceivably result in lower lifetime economic and educational attainment among some sexual minority women compared to their exclusively heterosexual peers. Bisexual and other plurisexual women again seem particularly disadvantaged in these regards. Evidence from the U.S. finds that bisexual women are significantly less likely to have completed high school, and significantly more likely to be unemployed, uninsured and living in poverty than exclusively heterosexual women (Badgett, 2018; Charlton et al., 2018; Mollborn & Everett, 2015). Meanwhile, lesbian women are no more likely to be poor than heterosexual women (Badgett, 2018). Lower socioeconomic status is associated with an increased risk of experiencing chronic stress (Baum et al., 1999), which in turn lowers resilience in the face of stressful life events (Pearlin et al., 1981; Pearlin, 2010).

The Resilience Perspective

While the above suggests that older sexual minority women – especially bisexual and other plurisexual women – will be more vulnerable to the negative impacts of stressful life events, there is a counterargument. When considering pathways to health and wellbeing outcomes, the health equity promotion model emphasises both vulnerability and resilience in the lives of LGB people (Fredriksen-Goldsen et al., 2014). Resilience perspectives focus attention on a strength-oriented understanding of minority experience of adverse social conditions (Colpitts & Gahagan, 2016). Resilience is developed through supportive environments, protective relationships and individual characteristics such as positive self-esteem, cognitive ability and proactive coping (Colpitts & Gahagan, 2016). Group resources within identity-based communities, such as counter-structures and values, social support and the experience of environments free of stigma, can provide a significant buffer against stress and may aid in the development of resilience (Meyer, 2003). In line with this, we may observe no differences in loneliness upon life events across sexual identity groups, should sexual minority women shore up social and personal resources or coping mechanisms across the life course (Aneshensel & Avison, 2015).

Methods

Data

The aim of our research was to test whether stressful life events differentially impact the loneliness of older women according to their sexual identities. To achieve this, we used data from a national probability sample of Australian women born between 1946 and 1951 taking part in the Australian Longitudinal Study on Women’s Health (ALSWH). Women were recruited to the study in 1996 through stratified random sampling of the Australian Medicare database (Brown et al., 1998). Medicare Australia is the national health-insurance scheme covering all citizens and permanent residents. In Wave 1 (1996), there were 13,714 women from the 1946-1951 cohort who were randomly selected and agreed to participate in the study. Comparisons between this sample and national census data on a range of sociodemographic variables indicated that women in the study were broadly representative of Australian women of the same ages. Since 1996, data have been collected from the women approximately every three years via self-completed questionnaires. In this study, we utilise data from four waves: Wave 3 (2001, when sexual identity data were collected), Wave 4 (2004), Wave 5 (2007) and Wave 6 (2010, the last wave in which data on life events were collected). In Wave 3, 11,226 women from the original sample participated, a response rate of 82%. Although the ALSWH also contains a cohort of women born between 1921 and 1926, no data on sexual orientation were ever collected from these women and we were therefore unable to include them in our analyses.

Measures

Loneliness

The outcome variable for all our analyses was a measure of how often women had felt lonely over the past week. Possible responses ranged from 1 (rarely or none of the time) to 4 (most or all the time). The distribution of responses to this item was highly skewed, with women scoring 1 in approximately 70% of observations. Thus, to reduce skew, we used a log transformation of the outcome variable in our analyses. Data on loneliness was missing in 2.3% of person-year observations.

Sexual identity

Sexual identity was our level-2 (time-invariant) explanatory variable¹. In Wave 3 (2001) of the survey, when women were aged 50–55 years, they were asked the following question: ‘Which of these most closely describes your sexual identity?’ The seven possible responses were: (i) ‘I am exclusively heterosexual’, (ii) ‘I am mainly heterosexual’, (iii) ‘I am bisexual’, (iv) ‘I am mainly homosexual (lesbian)’, (v) ‘I am exclusively homosexual (lesbian)’, (vi) ‘I don't know’ and (vii) ‘I don't want to answer’. From these responses we created a set of three time-invariant dummy variables distinguishing between women who identified as (1) exclusively heterosexual, (2) mainly heterosexual, bisexual, or mainly lesbian (responses ii-iv collapsed to create a plurisexual group) and (3) exclusively lesbian. Women who responded ‘I don’t know’, or ‘I don’t want to answer’ were coded as missing.

Life Events

At level 1, we used a set of 13 time-variant dummy variables as predictors. These dummy variables captured whether women had experienced the following life events over the previous 12 months: major personal illness or injury; major decline in the health of partner, other close family member or friend; death of a partner or child; death of another close family member or friend; serious problems in their relationship with their partner (e.g. infidelity, separation or divorce); moving house; a significant decline in income; a child leaving home; significant changes in work type/hours (e.g. starting a new job, retiring); significant changes to partner’s work type/hours; victim of assault (physical or sexual); victim of other crime or adversity (e.g. robbery, natural disaster); and a positive event (e.g. birth of a grandchild, starting a new relationship, a personal achievement). Less than 1% of person-year observations were missing data on life events.

Age in years was our level-1 variable capturing time. To facilitate interpretation of the intercept, we centred age on 50 years (the youngest age of women in Wave 3).

Covariates

Time-invariant controls used in adjusted models included country of birth (Australia, other English-speaking country, European non-English-speaking country, Asian non-English-speaking country, other non-English-speaking country), highest educational attainment (no formal qualifications, secondary certificate, higher secondary certificate, post-school non-university qualification, university qualification), and an indicator variable for whether the respondent ever had children. Time-variant controls included perceived ability to manage on current income (impossible, always difficult, sometimes difficult, not too bad, easy), as actual income is not consistently measured in the survey; employment status (unpaid employment, paid employment, not in the labour market/unemployed), geographic location (major city, inner regional, outer regional, remote, overseas), availability of support from others when needed (all the time, most of the time, some of the time, little/none of the time), marital status (married, de facto, separated, divorced, widowed, never married) and an indicator variable for whether the respondent lives alone. Missingness was no more than 1.6% on any of the control variables, and for most of the variables it was less than 1%.

Statistical Analyses

Due to low levels of missingness (<3% for each variable) we used complete-case analysis. The number of observations contributed by each woman in our analytic sample ranged from 1 to 4, with an average of 3.5. To test the associations between sexual identity, life events and women’s loneliness trajectories we estimated a series of three nested growth curve models within the multilevel framework. All analyses were conducted using the mixed command in Stata 16. In Model 1, we included our time-invariant predictor, sexual identity and the cross-level interactions between sexual identity and age. This allowed us to assess if the intercepts and/or rates of change for women’s loneliness trajectories differed according to their sexual identities. In Model 2, we added life events and sociodemographic covariates. This enabled us to examine if recent life events impacted women’s levels of loneliness over and above their underlying trajectories, which were conditioned on sexual identity. It also allowed us to test if life events and sociodemographic variables partly attenuated associations between sexual identity and loneliness. In our final analysis (Model 3), we included interaction terms between sexual identity and each of the life events. This allowed us to establish if life events differentially impacted women’s loneliness according to their sexual identities.

Given the evidence that bisexual and other plurisexual women tend to be more disadvantaged than both exclusively heterosexual and exclusively lesbian women, we made plurisexual women the reference group in all our models. Model intercepts therefore represent the predicted log of loneliness for plurisexual women at age 50, while the coefficients for heterosexual and lesbian women show the differences in their predicted log of loneliness at age 50 compared to plurisexual women. Likewise, the model rates of change represent loneliness slopes for plurisexual women between the ages 50 and 65. Meanwhile, the rate of change coefficients for heterosexual and lesbian women show the differences in their loneliness slopes compared to the slope of plurisexual women. In Model 3, the main effects for each life event represents the impact of that event on the loneliness of plurisexual women, while the coefficients for heterosexual and lesbian women show the difference in impact on their loneliness compared to that of plurisexual women.

While the loneliness variable in our data was measured on an ordinal scale, we chose to treat it as continuous in our analyses. This approach is common in the literature, and generally makes for models that are easier to fit and results that are more intuitive to interpret. However, we acknowledge that there are potential pitfalls to this approach. Therefore, to test the robustness of our findings, we also estimated random-effects ordered logistic regression models using the original variable (i.e. prior to log transformation). We present the full results of these models in Appendix 2 and refer to them briefly in our results section.

Results

Descriptive statistics

Descriptive statistics for the sample are displayed in Appendix 1 (Table A1). We conducted omnibus tests, followed up when significant with post-hoc pairwise comparisons, to assess if the women in our sample differed significantly on loneliness, occurrence of life events or socio-demographic characteristics according to their sexual identity. The mean level of loneliness across observations was significantly lower for exclusively heterosexual (M = 1.4, SD = 0.7) and exclusively lesbian women (M = 1.4, SD = 0.7) than for plurisexual women (M = 1.6, SD = 0.9). There were few differences in the occurrence of life events across sexual identity groups. Plurisexual and exclusively lesbian women were more likely to report problems in their relationship with their partner (7.5% and 7.3%, respectively) compared to exclusively heterosexual women (4.1%). Exclusively lesbian women were also more likely to report the death of a close family member/friend (25.8%) and being the victim of a crime/adversity (15.5%) than exclusively heterosexual and plurisexual women (10.9% and 11.6%, respectively).

Not unexpectedly, women’s socio-demographic characteristics varied according to their sexual identity. Plurisexual women were less likely than exclusively heterosexual women, but more likely than exclusively lesbian women, to be married and have children. Plurisexual women were twice as likely as exclusively heterosexual women to be separated or divorced, and approximately four times as likely to be single (never married). In contrast to prior evidence from the U.S. (Mollborn & Everett, 2015), plurisexual and exclusively lesbian women were more likely than exclusively heterosexual women to have a university qualification. Interestingly, plurisexual women were more likely than exclusively heterosexual women to say that it was ‘impossible’ to manage on their current income, but also more likely to say that it was ‘easy’. Plurisexual women were the group least likely to report that social support was available to them ‘all of the time’, and the group most likely to report that it was available ‘little/none of the time’.

Results from Growth Curve Models

Table 1 displays results from our first two growth models. Estimates from Model 1 indicate no significant differences between plurisexual, exclusively heterosexual and exclusively lesbian women in average levels of loneliness at age 50 (i.e. the intercept). However, rates of change in loneliness differed significantly between plurisexual and exclusively heterosexual women. While loneliness increased with age among plurisexual women, among exclusively heterosexual women it decreased slightly. For exclusively lesbian women, average levels of loneliness changed little over time. Loneliness trajectories by sexual identity, based on estimates from Model 1, are displayed visually in Figure 1. Overall, our findings are consistent with prior, cross-sectional studies that have identified bisexual and other plurisexual women as a particularly vulnerable group with regards to their psychosocial outcomes (Doan Van et al., 2019; Puckett et al., 2016; Vrangalova & Savin-Williams, 2014).

Table 1.

Results from growth curve models of log of loneliness.

	Model 1		Model 2
	Est.	SE	Est.	SE
Intercept	0.294^***	0.038	0.228***	0.040
Sexual identity (ref=plurisexual)
Exclusively heterosexual	−0.044	0.039	0.010	0.035
Exclusively lesbian	−0.074	0.061	−0.040	0.055
Rate of change	0.010^*	0.004	0.011**	0.004
Sexual identity (ref=plurisexual)
Exclusively heterosexual	−0.012^**	0.004	−0.012**	0.004
Exclusively lesbian	−0.009	0.007	−0.010	0.006
Life events – previous 12 months
Major illness, injury, surgery or accident			0.024***	0.006
Major decline in the health of spouse, close family or friend			0.016***	0.004
Death of partner or child			0.205***	0.018
Death of other close family or friend			−0.003	0.005
Problems in relationship with partner			0.136***	0.010
Positive life event			−0.022***	0.004
Victim of assault (physical or sexual)			0.087***	0.014
Victim of other crime/adversity			0.022***	0.006
Changes to working life			0.006	0.005
Partner retired or made redundant			0.002	0.008
Moved to a new house			0.021**	0.006
Significant decline in income			0.012*	0.005
Child or other family member moved out			0.020**	0.007
Sociodemographic covariates	No		Yes
Model fit statistics
AIC	31,017.04		26,425.93
BIC	31,101.77		26,858.08
Deviance	30,997.04		26,323.94

Notes. Australian Longitudinal Study on Women’s Health. Women born 1956-1951. Data from waves 3 (2001), 4 (2004), 5 (2007), and 6 (2010). N = 35,362 observations from 10,133 women. Controls in adjusted models include financial status, employment status, geographic location, social support, marital status, lives alone, country of birth, educational attainment and has children. Statistical significance: * p < .05, ** p < .01, *** p < .001.

Figure 1.

Older women’s linear growth trajectories of loneliness conditioned on sexual identity (from Model 1).

In Model 2 we added life events and sociodemographic covariates. This improved model fit as indicated by decreases to the AIC, BIC and deviance statistic. Adding life events and covariates to the model did not attenuate differences in the rate of change in loneliness between plurisexual and exclusively heterosexual women. Consistent with prior research (Barlow et al., 2015; King et al., 2021; Vozikaki et al., 2018), the following life events were associated with an increase in loneliness (listed in order of magnitude of the coefficient): death of a partner or child, problems in relationship with partner, victim of assault, major illness or injury, victim of other crime/adversity, residential relocation, child or other family member moved out, major decline in health of partner/other family/close friend, and significant decline in income. In the other direction, experiencing a positive life event such as the birth of a grandchild or starting a new relationship was associated with a significant decrease in loneliness.

Table 2 displays the results of our third growth model, which includes interactions between sexual identity and each of the life events (Model 3). This is our main model of interest, as it tests the proposition that sexual minority women are more vulnerable to increased loneliness following stressful life events. Overall, we found more support for the resilience perspective than enhanced vulnerability, with the impacts of most life events on loneliness consistent across sexual identity groups. However, there were a few exceptions. In two instances, plurisexual women showed greater vulnerability to the impacts of stressful life events than their exclusively heterosexual and lesbian peers. The increase in loneliness following a major illness or injury was significantly greater among plurisexual women than exclusively heterosexual women, while the increase in loneliness associated with problems in the partner relationship was significantly greater among plurisexual women than exclusively lesbian women. Yet, in another two instances plurisexual women were less vulnerable to the negative impacts of life events. Among plurisexual women, a major decline in the health of their partner/close family/friend and moving to a new house were associated with significant decreases in loneliness, but this was not the case for the exclusively heterosexual women in our sample.

Table 2.

Results from growth curve models of log of loneliness with interactions between sexual identity and life events.

	Model 3
	Est.	SE
Intercept	0.260^***	0.047
Sexual identity (ref = plurisexual)
Exclusively heterosexual	−0.023	0.042
Exclusively lesbian	−0.069	0.067
Rate of change	0.012^**	0.004
Sexual identity (ref = plurisexual)
Exclusively heterosexual	−0.013^**	0.004
Exclusively lesbian	−0.013^*	0.007
Life events – previous 12 months
Major illness, injury, surgery or accident (main effect)	0.192^***	0.052
x Exclusively heterosexual	−0.171^**	0.052
x Exclusively lesbian	−0.100	0.076
Major decline in the health of partner/close family/friend (main effect)	−0.079^*	0.036
x Exclusively heterosexual	0.097^**	0.037
x Exclusively lesbian	0.052	0.057
Death of partner or child (main effect)	0.717	0.368
x Exclusively heterosexual	−0.511	0.368
x Exclusively lesbian	−0.691	0.412
Death of other close family or friend (main effect)	−0.018	0.040
x Exclusively heterosexual	0.015	0.040
x Exclusively lesbian	0.063	0.060
Problems in relationship with partner/spouse (main effect)	0.250^***	0.060
x Exclusively heterosexual	−0.115	0.061
x Exclusively lesbian	−0.209^*	0.095
Positive life event (main effect)	−0.071^*	0.034
x Exclusively heterosexual	0.049	0.035
x Exclusively lesbian	0.095	0.055
Victim of assault (physical or sexual) (main effect)	0.086	0.106
x Exclusively heterosexual	0.001	0.107
x Exclusively lesbian	0.017	0.162
Victim of other crime/adversity (main effect)	−0.062	0.051
x Exclusively heterosexual	0.085	0.051
x Exclusively lesbian	0.056	0.076
Changes to working life (main effect)	0.022	0.039
x Exclusively heterosexual	−0.015	0.039
x Exclusively lesbian	−0.014	0.062
Partner retired or made redundant (main effect)	−0.095	0.076
x Exclusively heterosexual	0.098	0.076
x Exclusively lesbian	0.134	0.125
Moved to a new house (main effect)	−0.115^*	0.053
x Exclusively heterosexual	0.138^**	0.053
x Exclusively lesbian	0.079	0.088
Significant decline in income (main effect)	0.033	0.043
x Exclusively heterosexual	−0.022	0.043
x Exclusively lesbian	0.009	0.068
Child or other family member moved out (main effect)	0.006	0.061
x Exclusively heterosexual	0.014	0.062
x Exclusively lesbian	0.039	0.095
Sociodemographic covariates	Yes
Model fit statistics
AIC	26,430.28
BIC	27,082.73
Deviance	26,276.28

Notes. Australian Longitudinal Study on Women’s Health. Women born 1956-1951. Data from waves 3 (2001), 4 (2004), 5 (2007), and 6 (2010). N = 35,362 observations from 10,133 women. Controls in include financial status, employment status, geographic location, social support, marital status, lives alone, country of birth, educational attainment and has children. Statistical significance: * p < 0.05, ** p < 0.01, *** p < 0.001.

Robustness Check

To test the robustness of our results, we estimated random-effects ordered logistic regression models using the original variable (i.e. prior to log transformation). The results of these models, reported in full in Appendix 2, were highly consistent with the results of our growth curve models. For example, there were no significant differences between plurisexual and exclusively heterosexual/lesbian women in the marginal probabilities of scoring 4 on the loneliness variable (‘I felt lonely most or all the time’) at age 50. However, there was a significant interaction between age and sexual identity, such that the predicted probability of scoring 4 stayed relatively constant for the exclusively heterosexual and lesbian women as they aged, but increased markedly for the plurisexual women. Consistent with the results of our growth models, plurisexual women were more likely than exclusively heterosexual or lesbian women to have a higher loneliness score following an adverse health event. However, in contrast to the results of our growth curve model, there were no significant differences between the three groups when problems in the partner relationship occurred.

Discussion

Loneliness is associated with a range of poor mental health outcomes (Cohen-Mansfield et al., 2016; Holt-Lunstad et al., 2015). Understanding variation in loneliness and gaining deeper understandings of groups that are more vulnerable to loneliness are therefore important tasks. In this paper, we drew on a dataset containing nuanced measures of sexual identity to generate novel findings on loneliness among older women of diverse sexual identities. Building on incipient evidence, we looked beyond a single moment in time and examined longitudinal trends in loneliness from age 50. We then considered whether the impacts of adverse life events on older women’s loneliness systematically varied across sexual identity groups.

Despite dramatic decreases in levels of structural stigma over the past two decades, discrimination against people who belong to a sexual minority group persists (Lyons et al., 2021). In addition, earlier life contexts and experiences can have lingering effects on older LGB people in the form of internalised stigma and identity concealment, poorer socioeconomic conditions and less support from families of origin. We therefore expected that older sexual minority women would exhibit higher levels of loneliness than exclusively heterosexual women. Yet, this was not uniformly the case. At age 50, the starting point for our analyses, levels of loneliness among plurisexual and exclusively lesbian women were not significantly different to those of exclusively heterosexual women. However, for the plurisexual women in our sample, disparities emerged over time. While levels of loneliness decreased slightly as exclusively heterosexual women aged, for plurisexual women they increased. This is consistent with prior cross-sectional evidence that bisexual, mostly heterosexual and mostly lesbian women are vulnerable to poorer psychosocial outcomes and mental health compared to exclusively heterosexual (and in many cases exclusively lesbian) women (Doan Van et al., 2019; Perales, 2019; Vrangalova & Savin-Williams, 2014).

When we examined the impacts of time-variant life events on the loneliness of older women of diverse sexual identities, we found more support for resilience than enhanced vulnerability. This is in line with the view that many sexual minority women may have accumulated personal and social resources or developed coping mechanisms over the life course, with the effects of life events similar regardless of sexual identity in most cases. However, some exceptions to this were detected among plurisexual women. Adverse health events and problems in the partner relationship increased the loneliness of plurisexual women to a greater extent than that observed among exclusively heterosexual and exclusively lesbian women. Thus, not only did the plurisexual women in our sample have uniquely upward trajectories of loneliness in older age, but they were also particularly susceptible to the impacts of certain stressful life events.

The mechanisms giving rise to the enhanced vulnerability of plurisexual women in our sample is an open question. One possible explanation is that, while bisexual and other plurisexual women are still vulnerable to minority stressors, they might be less likely than lesbian women to connect with and gain support from the broader LGBTQAI+ community due to the phenomenon of double discrimination (Doan Van et al., 2019). Negative stereotypes of plurisexual people endorsed by both lesbian/gay and heterosexual people include that they are confused, attention-seeking, untrustworthy, promiscuous, immature, unstable and low on warmth and competence (Dyar et al., 2019; Flanders et al., 2019; Mize, 2016). Using a probability sample of adults in the U.S., Herek (2002) found that bisexual people were rated less favourably than all other groups asked about except injecting drug users. This stigmatisation of bisexuality is not without consequences. Plurisexual women who reported increased exposure to bisexual-specific sigma also reported increased minority stress in the form of sexual violence victimisation, identity concealment and internalised bi-negativity (Flanders et al., 2019; Dyar & London, 2018). Further research is needed to ascertain the impact this has on plurisexual women’s intrapersonal (e.g. self-esteem, self-mastery) and social resources. For example, the plurisexual women in our sample had relatively high rates of being single (never married), separated or divorced, and they reported the lowest levels of social support of all identity groups. Gaining a greater understanding of the life courses and experiences of these women is a task of paramount importance going forwards.

Though our findings have contributed to existing literatures, our study is not without limitations. Unfortunately, the very small numbers of women who identified as bisexual or mainly lesbian (34 women, 0.3% of our sample) meant that we had to combine them with mainly heterosexual women into a single plurisexual group to ensure adequate statistical power and minimise Type II errors. Further research that examines these groups separately would be informative. In addition, the proportion of our sample who identified as exclusively lesbian was also small (∼1%), which may have limited our ability to detect significant differences when compared to exclusively heterosexual women (who comprised almost 98% of our sample). This is an unavoidable limitation of using probability samples to study minority populations and reinforces the need to complement studies such as ours with research using purposively recruited samples of sexual minority people (e.g. Fredriksen-Goldsen et al., 2017). Other directions for future research include testing whether our findings replicate among older adults of different genders, including men, non-binary people, transgender people and other gender minorities; and examining how gender and sexual identities intersect in shaping loneliness among older people.

In conclusion, the results of our study underscore the value of considering vulnerability from a longitudinal perspective, with sexual identity-based disparities in loneliness only emerging among women in our sample after age 50 and worsening with time. Longitudinal analyses are also important for examining the effects of relatively infrequent adverse life events (such as health events or relationship problems), which our results suggest may compound systemic disadvantages faced by sexually minoritized groups in some instances. Building on existing evidence (e.g. Fredriksen-Goldsen et al., 2017), our study further highlights bisexual and other plurisexual people as a vulnerable group of older people deserving attention. Concerted efforts to challenge negative stereotypes and raise awareness of the lived experiences of bisexual and other plurisexual people is long overdue. Such efforts must span not only the public domain, but also target those working with older people such as health and aged care professionals. Community-based programmes and interventions can also play a role helping to build the social and intrapersonal resources of bisexual and other plurisexual women. Though we did not find significant differences in loneliness between lesbian and heterosexual women, which is consistent with the resilience perspective, we suggest further research is needed to understand the potential costs of resilience. For example, research has shown that positive adaption and resilience in the face of adversity can come with a physical cost (Rudd et al., 2021). Future research that explores whether such costs are also observed for sexual minority people, particularly over the longer term, would be valuable.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Australian Research Council Centre of Excellence for Children and Families over the Life Course (Project ID CE140100027). The research on which this paper is based was conducted as part of the Australian Longitudinal Study on Women’s Health (ALSWH) by the University of Queensland and the University of Newcastle. We are grateful to the Australian Government Department of Health for funding and to the women who provided the survey data. The ALSWH survey programme has ongoing ethical approval from the Human Research Ethics Committees (HRECs) of the Universities of Newcastle and Queensland (approval numbers H-076-0795 and 2004000224, respectively, for the 1973-78, 1946-51 and 1921-26 cohorts; and, H-2012-0256 and 2012000950, for the 1989-95 cohort). We would also like to acknowledge the research assistance of Catherine Dickson and Ellen Nixon.

ORCID iD

Jack Lam

Note

Author Biographies

Jack Lam is a research fellow at the Institute for Social Science Research, University of Queensland, Australia, and the ARC of Excellence on Children and Families over the Life Course. His research focuses on family dynamics, health, and ageing.

Alice Campbell is Postdoctoral Research Fellow in the Australian Research Council Centre of Excellence for Children and Families over the Life Course at the Institute for Social Science Research. Alice's research expertise falls under the broad themes of gender and sexuality, family dynamics, social inequalities, and longitudinal and life-course studies.

Appendix 1

Table A1.

Descriptive statistics for full sample, and by sexual identity.

	Full sample	Exclusively heterosexual (a)	Plurisexual (Bi+) (b)	Exclusively Lesbian (c)	Omnibus test: statistical significance
	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Omnibus test: statistical significance
Loneliness (1-4)	1.4 (0.7)	1.4 (0.7)^b	1.6 (0.9)^a	1.4 (0.7)	^***
Age in years (level-1 time variable)	56.9 (3.7)	56.9 (3.7)	56.7 (3.7)	57.0 (3.8)
Key explanatory variables	%	%	%	%
Sexual identity (level 2)
Exclusively heterosexual	97.6
Plurisexual	1.4
Exclusively lesbian	1.0
Life events – previous 12 months (level 1)
Major illness, injury, surgery or accident	11.6	11.6	10.8	14.1
Major decline in the health of spouse, close family or friend	34.1	34.1	30.6	36.7
Death of partner or child	1.1	1.1	0.2	1.2
Death of other close family or friend	20.0	20.0^c	20.4	25.8^a	^*
Problems in relationship with partner	4.1	4.1^bc	7.5^a	7.3^a	^***
Positive life event	27.3	27.2	30.5	30.8
Victim of assault (physical or sexual)	1.8	1.8	2.2	2.9
Victim of other crime/adversity	10.9	10.9^c	11.6	15.5^a	^*
Changes to working life	24.4	24.3	27.1	24.0
Partner retired or made redundant	5.9	6.0	4.5	3.8
Moved to a new house	9.4	9.4	10.0	8.5
Significant decline in income	19.4	19.3	20.8	19.4
Child or other family member moved out	8.9	9.0	7.1	7.3
Controls
Time-variant (level 1)
Ability to manage on income					^***
Impossible	1.7	1.6^b	3.3^a	2.1
Always difficult	9.6	9.5^bc	13.2^a	14.7^a
Sometimes difficult	24.9	25.0^b	20.2^a	22.0
Not too bad	44.9	45.0^b	40.1^a	39.9
Easy	19.0	18.9^b	23.2^a	21.4
Employment status					^*
Employed, not paid	10.8	10.8	12.8	11.1
Employed, paid	59.3	59.2^c	61.9	64.8^a
Not in labour force/unemployed	29.9	30.1^bc	25.3^a	24.0^a
Geographic location					^***
Major cities of Australia	36.7	36.5^b	43.0^a	39.9
Inner regional Australia	40.2	40.2	40.1	37.8
Outer regional Australia	19.4	19.5^b	14.5^a	17.0
Remote Australia	3.7	3.7	2.4	4.1
Overseas	0.1	0.1^c	0.0^c	1.2^ab
Social support availability					^***
All the time	51.9	52.1^b	37.1^ac	54.8^b
Most of the time	28.1	28.0^bc	33.8^ac	22.6^ab
Some of the time	14.1	14.0^b	18.1^a	13.8
Little/none of the time	6.0	5.9^bc	11.0^a	8.8^a
Marital status					^***
Married	75.1	75.9^bc	48.7^ac	34.6^ab
De facto	5.3	4.9^bc	13.2^ac	34.3^ab
Separated	3.3	3.3^b	6.1^a	4.4
Divorced	9.7	9.5^bc	19.3^ac	14.1^ab
Widowed	4.4	4.4	3.3	2.6
Never married	2.2	2.0^bc	9.4^a	10.0^a
Lives alone	12.6	12.4^bc	21.2^a	21.7^a	^***
Time-invariant (level 2)
Country of birth					^***
Australia	78.3	78.4^b	71.9^ac	78.9^b
Other English-speaking country	14.2	14.2	13.6	17.6
European non-English-speaking country	5.3	5.3^bc	7.7^ac	2.3^ab
Asian non-English-speaking country	1.6	1.5^b	5.7^ac	1.2^b
Other non-English-speaking country	0.6	0.6	1.2^c	0.0^b
Highest educational attainment					^***
No formal qualifications	10.3	10.4^b	6.5^a	9.4
Secondary certificate (year 10)	25.6	25.8^bc	13.0^ac	20.2^ab
Higher secondary certificate (year 12)	19.6	19.8^bc	13.4^a	14.4^a
Post-school non-university qualification	24.2	24.1^bc	29.9^ac	18.5^ab
University qualification	20.3	19.9^bc	37.3^a	37.5^a
Ever had children	92.2	92.6^bc	82.7^ac	68.6^ab	^***
Observations	35,362	34,512	509	341
Individuals	10,133	9890	146	97

Notes. Australian Longitudinal Study on Women’s Health. Women born 1956-1951. Waves 3 (2001)–6 (2010). Omnibus tests (chi-squared for categorical variables and ANOVA tests for continuous variables) showed significant differences on all control variables, and on some life events variables, by sexual identity. Post-hoc pairwise comparisons were conducted (two-sample proportions z-tests for categorical variables and Tukey comparisons of means for continuous variables). Superscript letters denote groups with significantly different proportions/means (p < .05).

Appendix 2

Table A4.

Odds ratios from random-effects ordered logistic models of loneliness with interactions between sexual identity and life events

	Model 3
	OR	95% CIs
Age (centred at 50 years)	1.09^**	1.02, 1.16
Sexual identity (ref = plurisexual)
Exclusively heterosexual	.83	.44, 1.57
Exclusively lesbian	.55	.18, 1.64
Sexual identity x age interactions
Exclusively heterosexual x age	.91^**	.85, .97
Exclusively lesbian x age	.91	.82, 1.02
Life events – previous 12 months
Major illness, injury, surgery or accident (main effect)	3.62^***	1.71, 7.69
x Exclusively heterosexual	.33^**	.15, .70
x Exclusively lesbian	.53	.17, 1.69
Major decline in the health of partner/close family/friend (main effect)	.65	.37, 1.13
x Exclusively heterosexual	1.82^*	1.03, 3.22
x Exclusively lesbian	1.11	.43, 2.85
Death of partner or child (main effect)	25.1	.36, 1736.4
x Exclusively heterosexual	.12	<.001, 8.66
x Exclusively lesbian	.04	<.001, 4.76
Death of other close family or friend (main effect)	.82	.45, 1.49
x Exclusively heterosexual	1.18	.65, 2.15
x Exclusively lesbian	1.92	.73, 5.07
Problems in relationship with partner/spouse (main effect)	3.99^***	1.80, 8.82
x Exclusively heterosexual	.55	.25, 1.23
x Exclusively lesbian	.38	.10, 1.49
Positive life event (main effect)	.57^*	.34, .98
x Exclusively heterosexual	1.43	.84, 2.45
x Exclusively lesbian	2.09	.84, 5.16
Victim of assault (physical or sexual) (main effect)	1.53	.39, 5.96
x Exclusively heterosexual	1.13	.29, 4.49
x Exclusively lesbian	.94	.12, 7.57
Victim of other crime/adversity (main effect)	.623	.29, 1.34
x Exclusively heterosexual	1.932	.89, 4.19
x Exclusively lesbian	1.454	.42, 4.98
Changes to working life (main effect)	1.24	.68, 2.25
x Exclusively heterosexual	.86	.47, 1.57
x Exclusively lesbian	.92	.33, 2.55
Partner retired or made redundant (main effect)	.53	.15, 1.89
x Exclusively heterosexual	1.98	.55, 7.19
x Exclusively lesbian	3.08	.35, 27.18
Moved to a new house (main effect)	.46	.21, 1.02
x Exclusively heterosexual	2.59^*	1.15, 5.83
x Exclusively lesbian	2.0	.49, 8.14
Significant decline in income (main effect)	1.23	.65, 2.34
x Exclusively heterosexual	.88	.46, 1.69
x Exclusively lesbian	1.07	.37, 3.09
Child or other family member moved out (main effect)	1.03	.41, 2.58
x Exclusively heterosexual	1.15	.46, 2.91
x Exclusively lesbian	1.55	.36, 6.65
Sociodemographic covariates	Yes

References

Aneshensel

C. S.

Avison

W. R.

(2015). The stress process: An appreciation of leonard I. Pearlin. Society and Mental Health, 5(2), 67–85. https://doi.org/10.1177/2156869315585388

Badgett

M. V. L.

(2018). Left out? Lesbian, gay, and bisexual poverty in the US. Population Research and Policy Review, 37(5), 667–702. https://doi.org/10.1007/s11113-018-9457-5

Barlow

M. A.

Liu

S. Y.

Wrosch

(2015). Chronic illness and loneliness in older adulthood: The role of self-protective control strategies. Health Psychology, 34(8), 870–879. https://doi.org/10.1037/hea0000182

Baum

Garofalo

J. P.

Yali

A. M.

(1999). Socioeconomic status and chronic stress: does stress account for SES effects on health? Annals of the New York Academy of Sciences, 896(1), 131–144. https://doi.org/10.1111/j.1749-6632.1999.tb08111.x

Brown

W. J.

Bryson

Byles

J. E.

Dobson

A. J.

Lee

Mishra

Schofield

(1998). Women's health Australia: Recruitment for a national longitudinal cohort study. Women and Health, 28(1), 23–40. https://doi.org/10.1300/j013v28n01_03

Buecker

Denissen

J. J.

Luhmann

(2020). A propensity-score matched study of changes in loneliness surrounding major life events. Journal of Personality and Social Psychology, 121(3), 669–690. https://doi.org/10.1037/pspp0000373

Burns

R. A.

Butterworth

Jorm

A. F.

(2018). The long-term mental health risk associated with non-heterosexual orientation. Epidemiology and Psychiatric Sciences, 27(1), 74–83. https://doi.org/10.1017/S2045796016000962

Carastathis

G. S.

Cohen

Kaczmarek

Chang

(2017). Rejected by family for being gay or lesbian: Portrayals, perceptions, and resilience. Journal of Homosexuality, 64(3), 289–320. https://doi.org/10.1080/00918369.2016.1179035

Charlton

B. M.

Gordon

A. R.

Reisner

S. L.

Sarda

Samnaliev

Austin

S. B.

(2018). Sexual orientation-related disparities in employment, health insurance, healthcare access and health-related quality of life: A cohort study of US male and female adolescents and young adults. BMJ Open, 8(6), e020418. https://doi.org/10.1136/bmjopen-2017-020418

10.

Chen

Feeley

(2014). Social support, social strain, loneliness, and well-being among older adults: An analysis of the Health and Retirement Study. Journal of Social and Personal Relationships, 31(2), 141–161. DOI:10.1177/0265407513488728

11.

Cohen-Mansfield

Hazan

Lerman

Shalom

(2016). Correlates and predictors of loneliness in older-adults: A review of quantitative results informed by qualitative insights. International Psychogeriatrics, 28(4), 557–576. https://doi.org/10.1017/S1041610215001532

12.

Colpitts

Gahagan

(2016). The utility of resilience as a conceptual framework for understanding and measuring LGBTQ health. International Journal for Equity in Health, 15(1), 60–68. https://doi.org/10.1186/s12939-016-0349-1

13.

Doan Van

E. E.

Mereish

E. H.

Woulfe

J. M.

Katz-Wise

S. L.

(2019). Perceived discrimination, coping mechanisms, and effects on health in bisexual and other non-monosexual adults. Archives of Sexual Behavior, 48(1), 159–174. https://doi.org/10.1007/s10508-018-1254-z

14.

Dodge

Herbenick

Friedman

M. R.

Schick

T. J.

Bostwick

Bartelt

Muñoz-Laboy

Pletta

Reece

Sandfort

T. G.

(2016). Attitudes toward bisexual men and women among a nationally representative probability sample of adults in the United States. PloS one, 11(10), e0164430. https://doi.org/10.1371/journal.pone.0164430

15.

Drydakis

(2015). Sexual orientation discrimination in the United Kingdom’s labour market: A field experiment. Human Relations, 68(11), 1769–1796. https://doi.org/10.1177/0018726715569855

16.

Dyar

Feinstein

Davila

(2019). Development and validation of a brief version of the anti-bisexual experiences scale. Archives of Sexual Behavior, 48(3), 1–15. https://doi.org/10.1007/s10508-018-1157-z

17.

Dyar

London

(2018). Longitudinal examination of a bisexual-specific minority stress process among bisexual cisgender women. Psychology of Women Quarterly, 42(3), 342–360. https://doi.org/10.1177/0361684318768233

18.

Eres

Postolovski

Thielking

Lim

M. H.

(2021). Loneliness, mental health, and social health indicators in LGBTQIA+ Australians. American Journal of Orthopsychiatry, 91(3), 358–366. https://doi.org/10.1037/ort0000531

19.

Fish

Weis

(2019). All the lonely people, where do they all belong? An interpretive synthesis of loneliness and social support in older lesbian, gay and bisexual communities. Quality in Ageing, 20(3), 130–142. https://doi.org/10.1108/QAOA-10-2018-0050

20.

Flanders

LeBreton

Robinson

(2019). Bisexual Women’s Experience of Microaggressions and Microaffirmations: A Community-Based, Mixed-Methods Scale Development Project. Archives of Sexual Behavior, 48(1), 143–158. https://doi.org/10.1007/s10508-017-1135-x

21.

Flentje

Heck

N. C.

Brennan

J. M.

Meyer

I. H.

(2020). The relationship between minority stress and biological outcomes: A systematic review. Journal of Behavioral Medicine, 43(5), 673–694. https://doi.org/10.1007/s10865-019-00120-6

22.

Fredriksen-Goldsen

K. I.

Shiu

Bryan

A. E.

Goldsen

Kim

H. J.

(2017). Health equity and aging of bisexual older adults: Pathways of risk and resilience. The Journals of Gerontology: Series B, 72(3), 468–478. https://doi.org/10.1093/geronb/gbw120

23.

Fredriksen-Goldsen

K. I.

Simoni

J. M.

Kim

H. J.

Lehavot

Walters

K. L.

Yang

Hoy-Ellis

C. P.

Muraco

(2014). The health equity promotion model: Reconceptualization of lesbian, gay, bisexual, and transgender (LGBT) health disparities. American Journal of Orthopsychiatry, 84(6), 653–663. https://doi.org/10.1037/ort0000030

24.

Gato

Leal

Coimbra

Tasker

(2020). Anticipating parenthood among lesbian, gay, bisexual, and heterosexual young adults without children in Portugal: Predictors and profiles. Frontiers in Psychology, 11, 1058. https://doi.org/10.3389/fpsyg.2020.01058

25.

Green

(2016). Do the companionship and community networks of older LGBT adults compensate for weaker kinship networks? Quality in Ageing and Older Adults, 17(1), 36–49. https://doi.org/10.1108/QAOA-07-2015-0032

26.

Hawkley

L. C.

Cacioppo

J. T.

(2010). Loneliness matters: A theoretical and empirical review of consequences and mechanisms. Annals of Behavioral Medicine, 40(2), 218–227. https://doi.org/10.1007/s12160-010-9210-8

27.

Herek

G. M.

(2002). Heterosexuals attitudes toward bisexual men and women in the United States. Journal of sex Research, 39(4), 264–274. https://doi.org/10.1080/00224490209552150

28.

Holt-Lunstad

Smith

T. B.

Baker

Harris

Stephenson

(2015). Loneliness and social isolation as risk factors for mortality: A meta-analytic review. Perspectives on Psychological Science, 10(2), 227–237. https://doi.org/10.1007/s12160-010-9210-8

29.

Hsieh

Wong

J. S.

(2020). Social networks in later life: Similarities and differences between sexual-minority and heterosexual older adults. Socius, 6, 1–10. https://doi.org/10.1177/2378023120977731

30.

Hughes

(2018). Health and well being of lesbian, gay, bisexual, transgender and intersex people aged 50 years and over. Australian Health Review, 42(2), 146–151. https://doi.org/10.1071/AH16200

31.

Kahle

E. M.

Veliz

McCabe

S. E.

Boyd

C. J.

(2020). Functional and structural social support, substance use and sexual orientation from a nationally representative sample of US adults. Addiction, 115(3), 546–558. https://doi.org/10.1111/add.14819

32.

King

Carr

D. C.

Taylor

M. G.

(2021). Loneliness following widowhood: The role of the military and social support. The Journals of Gerontology: Series B, 76(2), 403–414. https://doi.org/10.1093/geronb/gbz164

33.

Leal

Gato

Tasker

(2019). Prospective parenting: Sexual identity and intercultural trajectories. Culture, Health & Sexuality, 21(7), 757–773. https://doi.org/10.1080/13691058.2018.1515987

34.

Leigh-Hunt

Bagguley

Bash

Turner

Turnbull

Valtorta

Caan

(2017). An overview of systematic reviews on the public health consequences of social isolation and loneliness. Public Health, 152, 157–171. https://doi.org/10.1016/j.puhe.2017.07.035

35.

Luo

Hawkley

L. C.

Waite

L. J.

Cacioppo

J. T.

(2012). Loneliness, health, and mortality in old age: A national longitudinal study. Social Science & Medicine, 74(6), 907–914. https://doi.org/10.1016/j.socscimed.2011.11.028

36.

Lyons

Alba

Waling

Minichiello

Hughes

Barrett

Fredriksen-Goldsen

Edmonds

Blanchard

(2021). Recent versus lifetime experiences of discrimination and the mental and physical health of older lesbian women and gay men. Ageing & Society, 41(5), 1072–1093. https://doi.org/10.1017/S0144686X19001533

37.

Mereish

E. H.

Goldstein

C. M.

(2020). Minority stress and cardiovascular disease risk among sexual minorities: Mediating effects of sense of mastery. International Journal of Behavioral Medicine, 27(6), 726–736. https://doi.org/10.1007/s12529-020-09919-z

38.

Mereish

E. H.

Parra

L. A.

Watson

R. J.

Fish

J. N.

(2022). Subtle and intersectional minority stress and depressive symptoms among sexual and gender minority adolescents of color: Mediating role of self-esteem and sense of mastery. Prevention Science, 23(1), 142–153. https://doi.org/10.1007/s11121-021-01294-9

39.

Merz

E.-M.

de Jong Gierveld

(2016). Childhood memories, family ties, sibling support and loneliness in ever-widowed older adults: Quantitative and qualitative results. Ageing & Society, 36(3), 534–561. https://doi.org/10.1017/S0144686X14001329

40.

Meyer

I. H.

(2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychological Bulletin, 129(5), 674–697. https://doi.org/10.1037/0033-2909.129.5.674

41.

Mize

T. D.

(2016). Sexual orientation in the labor market. American Sociological Review, 81(6), 1132–1160. https://doi.org/10.1177/0003122416674025

42.

Mock

S. E.

Eibach

R. P.

(2012). Stability and change in sexual orientation identity over a 10-year period in adulthood. Archives of Sexual Behavior, 41(3), 641–648. https://doi.org/10.1007/s10508-011-9761-1

43.

Mollborn

Everett

(2015). Understanding the educational attainment of sexual minority women and men. Research in Social Stratification and Mobility, 41, 40–55. https://doi.org/10.1016/j.rssm.2015.04.004

44.

Pearlin

L. I.

(2010). The life course and the stress process: Some conceptual comparisons. The Journals of Gerontology: Series B, 65B(2), 207–215. https://doi.org/10.1093/geronb/gbp106

45.

Pearlin

L. I.

Menaghan

Lieberman

Mullan

(1981). The stress process. Journal of Health and Social Behavior, 22(4), 337–356. https://doi.org/10.2307/2136676

46.

Perales

(2019). The health and wellbeing of Australian lesbian, gay and bisexual people: A systematic assessment using a longitudinal national sample. Australian and New Zealand Journal of Public Health, 43(3), 281–287. https://doi.org/10.1111/1753-6405.12855

47.

Perales

Plage

(2020). Sexual orientation, geographic proximity, and contact frequency between adult siblings. Journal of Marriage and Family, 82(5), 1444–1460. https://doi.org/10.1111/jomf.12669

48.

Perlman

(2004). European and Canadian studies of loneliness among seniors. Canadian Journal on Aging/Le Revue Canadienne Du Vieillissement, 23(1), 181–188. https://doi.org/10.1353/cja.2004.0025

49.

Perlman

Peplau

L. A.

(1984). Loneliness research: A survey of empirical findings. In Peplau

L. A.

Goldston

S. E.

(Eds.), Preventing the harmful consequences of severe and persistent loneliness (pp. 13–46). U.S. Department of Health and Human Services, Public Health Service, Alcohol, Drug Abuse, and Mental Health Administration, National Institute of Mental Health.

50.

Pitoňák

(2017). Mental health in non-heterosexuals: Minority stress theory and related explanation frameworks review. Mental Health and Prevention, 5, 63–73. https://doi.org/10.1016/j.mhp.2016.10.002

51.

Puckett

J. A.

Surace

F. I.

Levitt

H. M.

Horne

S. G.

(2016). Sexual orientation identity in relation to minority stress and mental health in sexual minority women. LGBT Health, 3(5), 350–356. https://doi.org/10.1089/lgbt.2015.0088

52.

Pudrovska

Schieman

Pearlin

L. I.

Nguyen

(2005). The sense of mastery as a mediator and moderator in the association between economic hardship and health in late life. Journal of Aging and Health, 17(5), 634–660. https://doi.org/10.1177/0898264305279874

53.

Rico-Uribe

L. A.

Caballero

F. F.

Martín-María

Cabello

Ayuso-Mateos

J. L.

Miret

(2018). Association of loneliness with all-cause mortality: A meta-analysis. PloS one, 13(1), e0190033. https://doi.org/10.1371/journal.pone.0190033

54.

Rudd

K. L.

Doan

S. N.

Yates

T. M.

(2021). The physical health costs of positive adaptation to childhood adversity. Journal of Health Psychology, 26(9), 1324–1338. https://doi.org/10.1177/1359105319873961

55.

Shenkman

(2021). Anticipation of stigma upon parenthood impacts parenting aspirations in the LGB community in Israel. Sexuality Research & Social Policy: A Journal of the NSRC, 18(3), 753–764. https://doi.org/10.1007/s13178-020-00498-y

56.

Shin

Park

Amano

Kwon

Kim

(2020). Nature of retirement and loneliness: The moderating roles of social support. Journal of Applied Gerontology, 39(12), 1292–1302. https://doi.org/10.1177/0733464819886262

57.

Stinchcombe

Hammond

N. G.

Wilson

(2020). Differential effects of social support by sexual orientation: A study of depression symptoms among older Canadians in the CLSA. Research on Aging, 42(9–10), 251–261. https://doi.org/10.1177/0164027520923111

58.

van Gundy

(2002). Gender, the assertion of autonomy, and the stress process in young adulthood. Social Psychology Quarterly, 65(4), 346–363. https://doi.org/10.2307/3090107

59.

Vozikaki

Papadaki

Linardakis

Philalithis

(2018). Loneliness among older European adults: Results from the survey of health, aging and retirement in Europe. Journal of Public Health, 26(6), 613–624. https://doi.org/10.1007/s10389-018-0916-6

60.

Vrangalova

Savin-Williams

R. C.

(2014). Psychological and physical health of mostly heterosexuals: A systematic review. The Journal of Sex Research, 51(4), 410–445. https://doi.org/10.1080/00224499.2014.883589

61.

Weiss

R. S.

(1987). Reflections on the present state of loneliness. Journal of Social Behavior and Personality, 2(2), 1–16.

62.

World Health Organization . (2021). Social isolation and loneliness among older people: Advocacy brief. WHO. https://apps.who.int/iris/bitstream/handle/10665/343206/9789240030749-eng.pdf?sequence=1

63.

Wright

M. R.

Hammersmith

A. M.

Brown

S. L.

Lin

I-Fen

. (2020). The roles of marital dissolution and subsequent repartnering on loneliness in later life. The Journals of Gerontology: Series B, 75(8), 1796–1807. https://doi.org/10.1093/geronb/gbz121