Social Connections and Loneliness in OECD Countries

Report

16 October 2025

3. Inequalities in social connections: Who is least connected?

Copy link to 3. Inequalities in social connections: Who is least connected?

Abstract

Comparing social connections outcomes across socio-demographic groups reveals which portions of society are most at risk for isolation, loneliness and lack of connection. Those experiencing low income, unemployment and lower levels of education are much more likely to experience deprivations in social connections. Those belonging to minority groups, those who live alone and single people are also more likely to experience poor social connections outcomes. Young people consistently have better social connections outcomes than older age groups, however evidence of flipped age patterns from individual country surveys suggests that these age dynamics may be changing. Gender differences in social connections are generally smaller than those between other population groups.

Figure 3.1. A snapshot of inequalities in social connections outcomes across major socio-demographic groups
Copy link to Figure 3.1. A snapshot of inequalities in social connections outcomes across major socio-demographic groups

Different parts of the population have access to different social resources – goods, information, opportunities and affection provided through relationships and social networks (Webel et al., 2015[1]) – and their experiences of social connectedness vary. This chapter uses high quality data with large sample sizes to investigate inequalities in social connections for a broad range of population groups, including by socio-economic and labour market outcomes, belonging to a minority group, lifestyle characteristics including living arrangements and relationship status, age and gender (refer to Annex Table 3.A.1 for definitional details on each group). This exercise sheds light on who in society is most connected, and conversely, who is most at-risk for isolation and feeling lonely. This is not only important for understanding societal inequalities more broadly, but also for promoting social connections in the most efficient and targeted way (i.e. feelings of loneliness in youth vs. the elderly will likely best be addressed through different types of interventions). Actions to combat poor outcomes also need to consider, where possible, the ways in which risk compounds at the individual level: for example, someone who is both unemployed and living alone may be at especially heightened risk, compared to those experiencing only one of these two situations (Box 3.1).1 In order to ensure sufficiently large sample sizes for each population group, this chapter focuses on results for the OECD average, rather than for individual countries; national spotlights, using official survey data from OECD countries, are included throughout to supplement and strengthen findings from cross-country surveys.

At a high-level, the analysis shows that across a range of social connections indicators, those most at risk for poor outcomes include those with lower levels of education, those in the bottom income quintile, the unemployed, those who live alone and single people (Figure 3.1). For example, rates of feeling lonely are highest for those who live alone (14%), those in the bottom income quintile (13%), the unemployed (12%), single people (11%) and those with lower than upper secondary levels of education (10%) – conversely, those in the top income quintile are the least lonely (2%), followed by those in a relationship (3%). Patterns for social support are similar, but also reveal new groups at risk: in addition to groups with socio-economic deprivations, those who were born abroad (i.e. migrant populations) and those who live alone have some of the highest rates of lacking social support, at 13% each. Socio-economic considerations are also important for dissatisfaction with personal relationships, with the unemployed (9%) and the bottom income quintile (7%) reporting the highest dissatisfaction rates. Young people aged 16-24 outperform older age cohorts – in particular those aged 65+ – in all aspects of connection: they have the most active social lives, with over one-third (34%) getting together with friends in person on a daily basis. They are also less lonely and have higher reported social support.

The data presented in this chapter were mostly collected in 2022 and 2023, and therefore results should be viewed in the context of the tail-end of the COVID-19 pandemic, as well as the subsequent cost-of-living crisis. Confinement and social distancing policies were associated with poor social connections outcomes for all (OECD, 2021[2]), however those living alone and/or single were more affected during the most severe periods of social restrictions; similarly, cost of living pressures in recent years may have contributed to fewer financial resources that enable participation in social activities (OECD, 2021[2]; 2024[3]). However, trend data from Chapter 4 reveal that, while the pandemic undoubtedly shaped people’s social connectedness in the short term, it remains the case that single people, those who live alone and those on low incomes have historically had worse outcomes – so the challenges they face are structural and persistent rather than being unique to recent circumstances. Furthermore, groups that experienced the largest declines in outcomes over the pandemic are those that have historically not been at such high levels of risk, such as young people.

Box 3.1. Methodological insight: Overlapping risks for poor social connections outcomes

Copy link to Box 3.1. Methodological insight: Overlapping risks for poor social connections outcomes

Socio-demographic and contextual risk factors for loneliness, isolation and disconnection often overlap and intersect with one another. Analysing each risk factor individually overlooks the compounding effects and heightened vulnerabilities that can emerge for those who experience several risk factors simultaneously.

For example, studies suggest that elderly widowed men may be particularly susceptible to feelings of loneliness: not only grieving their spouse, but also becoming disconnected from the broader social network their spouse had maintained on the couple’s behalf (Dykstra, 2009[4]). Experiences of discrimination, language barriers, exposure to poverty and living in less safe neighbourhoods can also directly influence the size of one’s social circle, opportunities to socialise and feelings of inclusion – all of which may simultaneously affect people facing socio-economic deprivations, members of minority groups and recent migrant populations (Motmans et al., 2011[5]; Koelet and de Valk, 2016[6]; Kennedy, Field and Barker, 2019[7]; Taylor et al., 2024[8]).

Capturing data on these intersecting risk profiles can be difficult, even for national statistical offices. Nationally representative surveys are designed such that the sample is representative of major demographic considerations – sex, age, socio-economic status, among other factors – however once these attributes are combined, sample sizes are rarely sufficiently large to make reliable estimates of outcomes. One potential approach to address this is to tailor a survey specific to a population: for example, the Italian statistical agency developed a Social Condition and Integration among Foreign Citizens survey, designed to understand how the migrant population was integrating into broader society (Istat, 2012[9]). Beyond this, research has highlighted the methodological tools available to analysts interested in better exploring intersecting risk factors more generally – covering qualitative research design, multi-level modelling and mixed method approaches, among others (Yang, 2023[10]), with specific applications to the topic of loneliness, as well (Yang, 2023[11]).

Socio-economic and labour market outcomes

Copy link to Socio-economic and labour market outcomes

Individuals experiencing disadvantage in terms of income, education and occupation, have significantly worse outcomes than their more advantaged peers for several social connections outcomes: they feel lonelier, report fewer close friends, are more likely to lack social support and are more dissatisfied with relationships. Findings are more mixed in the case of daily socialising, reflecting the complex interacting dynamics of time poverty (which can affect both low-wage workers and high-income individuals, groups who both work long hours), enforced time off (in the case of the unemployed) as well as financial resource constraints that limit many social activities.

Education

In general, higher education levels are associated with better social connections outcomes (Figure 3.1). To illustrate the overall size of these gaps, this section compares outcomes between those with below upper secondary levels of education with those who have completed a tertiary degree (or above),2 however disaggregating by below upper secondary, at least upper secondary and tertiary levels of education typically reveals a gradient – with outcomes improving as education increases (see Figure 3.1).

Educational differences in quantitative social interactions present a mixed picture (Figure 3.2). When daily in-person socialising is considered,3 those with below upper secondary levels of education have better outcomes: for example, across 23 OECD countries, 17% see friends daily (Figure 3.2, Panel B), compared to 8% of those with a tertiary education. However, when looking at deprivations in social contact, those with a tertiary level of education are at lower risk: 3% of tertiary educated respondents report never getting together with friends in an average year, compared to 10% of those with below upper secondary. Patterns for getting together with family members are similar (Figure 3.2, Panel A). Therefore, those with below upper secondary levels of education are more likely to see friends and family both daily and never.

Figure 3.2. Those with lower levels of education are both more likely to interact with friends and family on a daily basis, *and* more likely to report never doing so
Copy link to Figure 3.2. Those with lower levels of education are both more likely to interact with friends and family on a daily basis, <em>and </em>more likely to report never doing so

This unevenness across different indicators suggests different potential constraints on spending time with others. Indeed, as is outlined in the following sections, these patterns are similar for other markers of socio-economic status, including income and employment outcomes, speaking to the often simultaneous and competing constraints of time and financial resources on socialising.

Those with lower levels of education are more likely to have lower paid jobs and may therefore have more limited financial resources to support certain types of social interactions (e.g., meeting at a bar or restaurant, hosting people at home, or traveling to meet distant friends and relatives); they may also be more likely to hold multiple jobs to make ends meet, resulting in limited time to socialise. For example, research finds that the unemployed spend less time socialising, with their additional enforced (as a result of not being in employment) time off spent on solitary, home-centred activities (Zuzanek and Hilbrecht, 2019[13]; Lobo, 1996[14]). Multiple studies have also found that young people from low income families have fewer friends, and that economic resources play a role in adolescent socialising (Hjalmarsson and Mood, 2015[15]; Cavicchiolo et al., 2022[16]).

At the other end of the scale, those with higher levels of education are more likely to have higher paid jobs, which sometimes also imply longer working hours that can constrain their available time for socialising. As an illustration of these complex forces, research using data from the American Time Use Survey finds that having a lower level of education, experiencing food insecurity and living in poverty are all associated with spending more time socialising (Davis et al., 2023[17]), while analysis of time use data in Great Britain finds that workers in both high- and low-paid (i.e., shift work, working unsocial hours) occupations are both more likely to be time deprived (Chatzitheochari and Arber, 2012[18]).

When looking at the perceived quality of social connections, those with lower levels of education fare worse across the board (Figure 3.3). On average, people with less than upper secondary levels of education are 2.5 times more likely to report feeling lonely than those with a tertiary education (10% vs. 4%), they are also less likely to report having someone to count on in times of need (10% vs. 7%), to have close friends (9% of those with a primary education have no close friends vs. 7% of tertiary) or family members (4% vs. 3%), and are twice as likely to be dissatisfied with their personal relationships (6% vs. 3%).

Figure 3.3. Those with below upper secondary levels of education feel lonelier, less satisfied with relationships and have less support
Copy link to Figure 3.3. Those with below upper secondary levels of education feel lonelier, less satisfied with relationships and have less support

Social connections outcomes, ratio (distance from parity) between people with below upper secondary vs. tertiary education attainment, OECD 22-38, 2022/2022-2023

Income

Frequency of social interactions by income show similar patterns to those by education. Across OECD countries, those in the bottom income quintile report slightly higher rates of getting together with friends (12.4%) (Figure 3.4, Panel B) and family (12.4%) (Figure 3.4, Panel A) in person on a daily basis in an average year, when compared to respondents in the top income quintile (11.7% for friends, 8.6% for family). However, at the same time, the bottom income quintile is also more likely to report never getting together with friends and family: 9% never see friends and 4% never see family, compared to only 2% and 1%, respectively, in the top income quintile. This again alludes to time and financial resource constraints, acting in different ways for those with high and low incomes, limiting or supporting their ability to socialise.4 Discrepancies are particularly large between income quintiles for never socialising with friends.

Gaps in function and quality measures of social connections between the top and bottom income quintiles are some of the largest disparities across any socio-demographic group (Figure 3.1). On average for the OECD countries with available data, respondents in the bottom income quintile are 6.2 times lonelier (13% vs. 2%), almost twice as likely to have no one to count on in times of need (14% vs. 7%) and 3.4 times more dissatisfied with their personal relationships (7% vs. 2%) (Figure 3.5). Indeed, research has consistently shown that poverty is correlated with feeling lonely, potentially caused by factors such as reduced social participation or compounding health issues (Schnepf, d’Hombres and Mauri, 2024[21]) (WHO, 2025[22]).5 Those with lower incomes, or who recently experienced financial stress, have fewer resources to spend on social leisure activities (Klinenberg, 2016[23]; WHO, 2025[22]); relatedly, those who are married or in long-term partnerships tend to have higher joint household incomes, and being in a relationship or living with a partner is a protective factor against deprivations in social connections outcomes (Hawkley, 2008[24]) (see also the below section on lifestyle characteristics). Additionally, poverty is associated with a number of other physical and mental health issues (Macdonald, Nixon and Deacon, 2018[25]), which can themselves influence social connections outcomes (Holt-Lunstad et al., 2015[26]).

Figure 3.5. Those in the bottom income quintile are lonelier, have less support and are less satisfied with their relationships, compared to those in the top income quintile
Copy link to Figure 3.5. Those in the bottom income quintile are lonelier, have less support and are less satisfied with their relationships, compared to those in the top income quintile

Social connections outcomes, ratio (distance from parity), top vs. bottom income quintile, OECD 23-38, 2022/2022-2023

Employment

In considering social connections outcomes by labour market status, the unemployed have worse outcomes in almost all aspects of social connections (Figure 3.6). The one exception is getting together with friends on a daily basis in an average year: on average across 23 European OECD countries, 12% of unemployed respondents report doing so, compared to 9% of employed respondents. However, the unemployed also have higher rates of never getting together with friends (7%, compared to 3% of the employed). This could possibly reflect greater time constraints of the employed; however unemployment is also associated with a diminished social network and feelings of exclusion (Kunze and Suppa, 2017[27]; Pohlan, 2019[28]), and lower enjoyment of leisure activities (Krueger and Mueller, 2012).

Figure 3.6. The unemployed have worse outcomes than the employed, for almost all aspects of social connections
Copy link to Figure 3.6. The unemployed have worse outcomes than the employed, for almost all aspects of social connections

Social connections outcomes, ratio (distance from parity) between unemployed vs. employed, OECD 23-35, 2022/2017-2023

Disparities between the unemployed and employed are particularly stark for indicators of perceived relationship quality. Across OECD countries, the unemployed are 3 times more likely to feel lonely (12% vs. 4%), 1.7 times more likely to report having no one to count on (14% vs. 8%), and 2.8 times more likely to be dissatisfied with personal relationships (8% vs. 3%) (Figure 3.6). The mechanisms underpinning this may be bi-directional, and mediated by income and mental health (Morrish and Medina-Lara, 2021[29]; Barjaková, Garnero and d’Hombres, 2023[30]; Morrish, Mujica-Mota and Medina-Lara, 2022[31]; Üniversity et al., 2025[32]). For example, unemployment can lead to a decrease in social interactions, less of a sense of community and/or belonging and a subsequent increase in feelings of loneliness (Barjaková, Garnero and d’Hombres, 2023[30]). However, feelings of loneliness may also contribute to experiences of mental or physical ill-health, which can then potentially lead to unemployment or poorer future employment prospects (Barjaková, Garnero and d’Hombres, 2023[30]; Üniversity et al., 2025[32]; Morrish, Mujica-Mota and Medina-Lara, 2022[31]).

Belonging to a minority group

Copy link to Belonging to a minority group

“Belonging to a minority group” is a broad term that encompasses the experiences of very different – and in and of themselves heterogeneous – communities, based on ethnicity or skin colour, language, disability, sexual orientation or gender identity, religion or belief, migrant status or political opinion. This section brings together internationally comparative information on social connectedness for different types of minority groups where possible. However, given that such statistics are not widely available, this analysis is supplemented with deep dives into experiences of specific communities from individual OECD countries.6

Figure 3.7. Respondents who identify as a member of a minority group report worse social connections outcomes, across the board
Copy link to Figure 3.7. Respondents who identify as a member of a minority group report worse social connections outcomes, across the board

Social connections outcomes, ratio (distance from parity) between those who belong to a minority group vs. those who do not, OECD 12, 2023

Data from 12 OECD countries show that belonging to a minority group of any sort is associated with worse outcomes in all aspects of social connections. On average, respondents who self-identify as a member of any minority group are 2.2 times more likely to have felt lonely over the past four weeks in comparison to non-minority respondents (29% vs. 13%), 1.5 times more likely to have an anxious attachment style, in that they feel worried they will get hurt if they get too close to a romantic partner (46% vs. 31%), and 1.2 times less likely to have interacted with other people in person over the past week (62% vs. 74%) (Figure 3.7). These high-level findings are in line with the majority of the evidence presented in this section, in that those belonging to different minority groups tend to have worse social connections outcomes than the general population; however, the frequency of socialising varies depending on the group considered.

Place of birth

Data from 23 European OECD countries show that a proxy for migrant status – namely, those who were not born in the country where the survey is being administered – is associated with higher rates of contacting family members on a daily basis in an average year (32%), in comparison to those who were born in the country (28%) (Figure 3.8).

Figure 3.8. Those born abroad have worse overall social connections outcomes than their native-born counterparts, but are more likely to stay in regular contact with family members
Copy link to Figure 3.8. Those born abroad have worse overall social connections outcomes than their native-born counterparts, but are more likely to stay in regular contact with family members

Social connections outcomes, ratio (distance from parity) between those born abroad vs. native-born, OECD 23-31, 2022/2017-2023

This likely reflects the close ties migrants retain to their home communities, and the increasing ubiquity of digital technology tools that facilitate communication across large physical distances (Navarrete and Huerta, 20006[34]; González and Castro, 2007[35]). Migrants are meanwhile less likely to report getting together with family members on a daily basis (8% vs. 11% for those not born in another country), perhaps unsurprising given that a higher share of their family members may be based abroad. These patterns are also reflected in national data from New Zealand (Box 3.2).

When looking at functional and quality aspects of social connection, those born abroad have consistently worse outcomes: on average, they are lonelier (8% vs. 6% for the native-born) and are more dissatisfied with personal relationships (5% vs. 3.7%). Disparities in outcomes between foreign and native-born respondents are largest for having someone to count on in times of need: those born abroad are 1.7 times more likely to report having no one (13% vs. 8%) (Figure 3.8). National data from Canada, Germany and New Zealand reveal similar dynamics (Box 3.2).

A possible mechanism for poorer social connections outcomes for those born abroad may be a weaker sense of belonging to their current place of living due to cultural differences, experiences and feelings of discrimination, or the size of their local network relative to their transnational network (Motmans et al., 2011[5]; Koelet and de Valk, 2016[6]). Analysis of a survey conducted by the Italian National Statistical Institute (Istat), Social Condition and Integration among Foreign Citizens, shows that the migrant population in Italy face greater risk for feelings of loneliness and isolation due to discrimination, language barriers, worse health outcomes and higher likelihood of living in deprived neighbourhoods (Eralba and Barbiano Di Belgiojoso, 2021[36]). Similar findings have been found for migrant communities in the Netherlands (Ten Kate et al., 2020[37]), Luxembourg (Albert, 2021[38]) and Canada (De Jong Gierveld, Van der Pas and Keating, 2015[39]).

Box 3.2. National spotlight: Foreign vs. native-born social connections outcomes in Germany, Canada and New Zealand

Copy link to Box 3.2. National spotlight: Foreign vs. native-born social connections outcomes in Germany, Canada and New Zealand

Data from Germany, Canada and New Zealand highlight differences in feelings of loneliness between the foreign- and native-born populations (Figure 3.9).

Data from Germany confirm that migrants (defined as those who were not born in Germany) are lonelier than non-migrants (Figure 3.9, Panel A).

In Canada, overall rates of loneliness between migrant and non-migrant respondents are comparable, however migrants are more likely to report feeling “somewhat” lonely, and less likely to report they “never” feel lonely (Figure 3.9, Panel B).

New Zealand provides data on a range of social connections outcomes for those born in the country, long-term migrants and recent migrants (Figure 3.9, Panel C). Recent migrants have the highest rate of engaging in remote interactions with family members, while native-born respondents have the highest rates of in-person family social interactions, similar to findings from European OECD countries (Figure 3.8). Compared to other groups, recent migrants to New Zealand also have the highest rates of loneliness, and lowest rates of access to social support.

Figure 3.9. Migrant populations in Germany, Canada and New Zealand are all lonelier than their native-born counterparts
Copy link to Figure 3.9. Migrant populations in Germany, Canada and New Zealand are all lonelier than their native-born counterparts

Location of parents’ birth

These dynamics of belonging and socially connecting often extend over generations of migration to other countries. In 18 European OECD countries with data, second-generation migrants, identified as respondents who are born in their current country of residence but report that both of their parents were born abroad, are more likely to say they contact family members remotely on a daily basis in a given year (29% on average), compared to respondents with two native-born parents (26%); they are also more likely to be in daily contact with friends (Figure 3.10). Prevalence of feeling lonely for second-generation migrants, at an average of 7%, is comparable to the elevated rates found among first generation-migrants, a finding that has been replicated by other cross-European surveys (Schnepf, d’Hombres and Mauri, 2024[21]).

Figure 3.10. Second-generation migrants continue to report worse qualitative social connections outcomes than those with native-born parents, but are more likely to remotely contact friends and family
Copy link to Figure 3.10. Second-generation migrants continue to report worse qualitative social connections outcomes than those with native-born parents, but are more likely to remotely contact friends and family

Social connections outcomes, ratio (distance from parity) between those whose parents were born abroad vs. those whose parents are both native-born, OECD EU-EFTA 18, 2022

Indigenous, racial and ethnic minority groups

Obtaining cross-country comparable data on outcomes for racial and ethnic minorities is difficult, due to different policies and protocols on collecting this information across OECD countries. Furthermore, individuals may identify with multiple racial or ethnic groups simultaneously, making comparisons across different racial and ethnic minority groups complicated. For countries that do collect and report on this information in national surveys, social connections outcomes for those belonging to racial and ethnic minority groups, or belonging to an Indigenous community, tend to be worse than outcomes for non-minority, non-Indigenous groups, and/or worse than outcomes for the population as a whole (Box 3.3).

Box 3.3. National spotlight: Social connections outcomes for minority groups in Canada, England, New Zealand and the United States

Copy link to Box 3.3. National spotlight: Social connections outcomes for minority groups in Canada, England, New Zealand and the United States

National surveys in Canada, England, the United States and New Zealand highlight differential outcomes across a range of social connections indicators for those who belong to Indigenous, racial and ethnic minority groups, and those who do not. In Canada, rates of loneliness are higher for members of visible minority groups¹ and those with an Indigenous identity (First Nations, Inuit, Métis): the latter are 1.4 times more likely to always or often feel lonely, in comparison to non-Indigenous respondents (Figure 3.11, Panel A).

Data from England show that white respondents have the lowest deprivations in social support outcomes; they also have the lowest reported rates of loneliness, although these differences are smaller in magnitude (Figure 3.11, Panel B).

In the United States, mixed race respondents are most likely to always or often feel lonely (17%), followed by Black and Hispanic (14% for both), and white respondents (11%); Asian respondents report the lowest levels of loneliness (9%). In looking at social support, Hispanic respondents are 1.9 times more likely than white respondents to report rarely getting the social and emotional support they need, and Asian respondents are 1.7 times more likely (Figure 3.11, Panel C).

Data from New Zealand sheds light not only on qualitative social connections outcomes – such as loneliness – but also on time spent socialising (Figure 3.11, Panel D). Māori respondents have the highest rates of interacting with family members in person on at least a weekly basis (65%), followed by Pacific peoples (63%), European origin respondents (59%), with Asian respondents having the lowest rates (49%). In-person interactions with friends follow a slightly different pattern: European origin respondents have the highest rates of in-person socialising with friends on at least a weekly basis at 67%, 65% of Asian respondents, followed by 59% of Māori and Pacific peoples. In terms of loneliness, Māori respondents are 1.7 times as likely to report having felt lonely most or all of the time over the past four weeks, compared to the total population, Pacific peoples are 1.4 times more likely and Asian respondents are 1.3 times more likely.

Figure 3.11. Members of ethnic minority groups tend to have worse qualitative social connections outcomes, across Canada, England, New Zealand and the United States
Copy link to Figure 3.11. Members of ethnic minority groups tend to have worse qualitative social connections outcomes, across Canada, England, New Zealand and the United States

Sexual orientation

When data are available, social connections outcomes for people identifying as lesbian, gay, bisexual, transgender and/or intersex (LGBTI) tend to be worse than for cisgender, heterosexual respondents. Data from 11 OECD countries show that LGBTQ+ people are 2.2 times more likely to feel lonely (32% vs. 15% for non-LGBTQ+), 1.5x more likely to feel worried that they will get hurt if they get too close to their partner (50% vs. 33%), 1.3x more likely to feel uncomfortable getting close to / trusting / and depending on others (48% vs. 36%), and 1.8x more likely to strongly disagree that they have a good social support network (8% vs. 4%) (Figure 3.12). In national surveys where these data are collected, similar patterns are exhibited (Box 3.4).7 LGBTI youth were particularly affected by the COVID-19 pandemic, especially those who were required to quarantine with family members who may not be supportive – and disconnected from wider friends and support networks (OECD, 2021[2]; Ruprecht et al., 2024[46]).

Figure 3.12. LGBTQ+ respondents are more likely to report feeling lonely, worried they will be hurt if they get too close to others, and to strongly disagree that they have great social support networks
Copy link to Figure 3.12. LGBTQ+ respondents are more likely to report feeling lonely, worried they will be hurt if they get too close to others, and to strongly disagree that they have great social support networks

Social connections outcomes, ratio (distance from parity), by sexual orientation, OECD 11, 2022

Box 3.4. National spotlight: LGBTI outcomes in Canada and England

Copy link to Box 3.4. National spotlight: LGBTI outcomes in Canada and England

LGBTQ+ people in Canada are 2.3 times more likely to feel lonely always or often, in comparison to non-LGBTQ+ Canadians (Figure 3.13, Panel A). Data from England show that heterosexual or straight respondents are the least likely group to report feeling lonely often or always (7%), in comparison to gay or lesbian (12%), bisexual (16%), or other (17%) sexual orientation respondents. Similarly, straight respondents have the highest rate of reporting they chat with their neighbours more than once a month (Figure 3.13, Panel B).

Figure 3.13. LGBTI respondents have worse social connections outcomes than non-LGBTI individuals in Canada and England
Copy link to Figure 3.13. LGBTI respondents have worse social connections outcomes than non-LGBTI individuals in Canada and England

Lifestyle characteristics

Copy link to Lifestyle characteristics

Lifestyle characteristics, including structural factors but also individual choices – such as whether one lives alone or with others, is single or in a relationship, or one’s place of residence – can also influence how connected people are and feel. In general, single people and those who live alone have worse qualitative social connections outcomes than those in a relationship or who live with other people, with particularly striking gaps for feeling lonely. Available cross-country data do not pick up on significant differences in social connections outcomes between rural and urban living.

Living arrangements

Across 30 OECD countries, almost one-fifth (19%) of the population lived in single-occupancy households in 2022, a slight increase from a decade prior (15%) (Figure 3.14, Panel A). Rates of living alone are highest in Nordic countries and the Baltic states, and lowest in Latin America, East and Southern Europe. Older people are most likely to live alone, with 31% of the 65+ population living alone in 2022; all age groups, however, experienced an increase in living alone between 2010 and 2022 (Figure 3.14, Panel B). Furthermore, OECD analysis suggests that the share of single-occupancy households is only likely to increase in the coming decades, due in part to rapid population ageing (OECD, 2024[47]).

Living alone can influence how frequently one socialises – on the one hand, living alone may induce people to more pro-actively plan social events with friends and family, and indeed data from the United States has found that people who live alone spend more time with both friends and neighbours (Klinenberg, 2013[48]). On the other hand, living on one’s own may lead to more overall time spent alone – particularly during events like the COVID-19 pandemic (Fingerman et al., 2021[49]). Data from 23 European countries, collected in 2022, show this tension: those who live alone report higher levels of never interacting with friends and family, however their rates of daily in-person interactions are either not significantly different from those who live with others, or slightly higher (Figure 3.15). Differences in contacting friends and family show similar patterns (not pictured).8

When considering a range of functional and qualitative social connections indicators, those who live alone have worse outcomes, and often by a large magnitude (Figure 3.15). While there is no significant difference between those who live alone and those who live with others in terms of having close friends, on average in 17 European countries those who live alone have fewer close relationships with family members. 11% report having no close family members, compared to only 2% of respondents who live with others.9 People who live alone are also 3.9 times more likely than those who live with others to report having been lonely most or all of the time over the past four weeks (14% vs. 4%), 1.5 times more likely to report having no one to count on in times of need (13% vs. 9%), and 1.8 times more likely to be dissatisfied with their personal relationships (6% vs. 3%). The links between living alone and experiencing feelings of loneliness are not straightforward, in that spending time alone or solitary living do not inherently imply an undesirable situation (Klinenberg, 2013[48]). However, living alone – as opposed to momentary experiences of solitude – may be particularly detrimental for older people (Pauly et al., 2016[50]; O’Súilleabháin, Gallagher and Steptoe, 2019[51]; Yeh and Lo, 2004[52]), those facing mobility challenges, and anyone, regardless of age, who has little access to social infrastructure that provides spaces and opportunities for socialising outside of the home (Klinenberg, 2016[53]) (see Chapter 5 for an extended discussion).

Figure 3.14. Across 30 OECD countries, close to 1 in 5 people live alone; older adults are more likely to live alone, but rates have been rising for all ages since 2010
Copy link to Figure 3.14. Across 30 OECD countries, close to 1 in 5 people live alone; older adults are more likely to live alone, but rates have been rising for all ages since 2010

Figure 3.15. Those who live alone have worse social connections outcomes than those who live with others
Copy link to Figure 3.15. Those who live alone have worse social connections outcomes than those who live with others

Social connections outcomes, ratio (distance from parity) between those who live alone vs. those who live with others, OECD 17-38, 2022/2022-2023

Relationship status

In comparison to people in a relationship, single people are more likely to frequently socialise with friends. In 2022, across 23 European OECD countries, 18% of single people reported they get together with friends on a daily basis in an average year, compared to 7% of people in a relationship; similarly, 38% of single people contact their friends daily, vs. 22% of those in a relationship (Figure 3.16).10 Differences in socialising with family members are much smaller in magnitude. Single people, however, are four times more likely than people in a relationship to report feeling lonely (11% vs. 3%), slightly more likely to have no one to count on (10% vs. 9%), and twice as likely to be dissatisfied with their personal relationships (6% vs. 3%).

Having a partner, whether through marriage or cohabitation, is one of the strongest predictors of reduced loneliness (Arpino et al., 2022[55]). Among other things, having a partner increases the frequency of having in-person contact and, in most cases, social support. However, the quality of the relationship with said spouse or partner is crucial; having a spousal confidant is found to be negatively associated with loneliness, while experiencing marital stress may be positively associated with loneliness (Hawkley, 2008[24]; Hawkley and Kocherginsky, 2017[56]). Indeed, data show that individuals in unsatisfactory relationships are more likely to be lonely than those not in a relationship (Schnepf, d’Hombres and Mauri, 2024[21]).11

Figure 3.16. Single people spend more time interacting with friends than people in a relationship, but single people are also more likely to feel lonely, lacking in support and dissatisfied with relationships
Copy link to Figure 3.16. Single people spend more time interacting with friends than people in a relationship, but single people are also more likely to feel lonely, lacking in support and dissatisfied with relationships

Social connections outcomes, ratio (distance from parity) between single vs. partnered, OECD 22-38, 2022/2022-2023

Place of residence – urban/rural

Existing data do not show significant differences in social connections outcomes for urban vs. rural residents across OECD countries, with the exception of a slightly higher likelihood for rural residents to frequently get together with friends in person (13% vs 11%) (Figure 3.17). This may in part be due to the insufficient granularity of “urban” and “rural” designations in the underlying datasets: because of small sample sizes, and/or inconsistent answer categorisations across different countries participating in the surveys used for this analysis, outcomes for residents in large urban areas are combined with those in smaller towns or suburbs.

Aside from somewhat limited granularity, however, research has found limited evidence for the impact of urban characteristics such as housing type, city size or area density on loneliness or isolation (Bower et al., 2023[57]). This may in part be explained by the fact that people who choose to live in different areas have different preferences for social contact. For example, a case study of suburban towns in the metro area of Warsaw shows that residents who move to these areas do so because they prefer the relative tranquillity of being outside of the city and closer to nature, and prefer to socialise with others either in the home or in private gardens (Kępkowicz and Mantey, 2016[58]). Other research has found a U-shaped curve for loneliness, in particular, with respondents in both the least (i.e., rural) and most (i.e., urban) densely populated areas having the highest rates of loneliness (Schnepf, d’Hombres and Mauri, 2024[21]).

Figure 3.17. Social connections outcomes do not significantly differ across urban vs. rural residents
Copy link to Figure 3.17. Social connections outcomes do not significantly differ across urban vs. rural residents

Social connections outcomes, ratio (distance from parity) between rural vs. urban residents, OECD 21-22, 2022

Age

Copy link to Age

Cross-country data from European OECD countries show that younger people tend to have better social connections outcomes than older age cohorts: they are more likely to interact with friends, they are more likely to report having someone to count on and they are less likely to report feeling lonely. However, data on age patterns from national surveys in several non-European OECD countries reveal that in Canada, England, New Zealand and the United States, young people are the loneliest age group (Box 3.5): this could reveal country-level differences in age-related social connections outcomes, or a reflection or worsening trends for young people that are further along in some countries than in others (see Chapter 4 on trends).

Age differences in social interactions are most apparent when looking at relationships with friends. While there is little variation in how often people of different ages interact with family members – either in person, or remotely – social interactions with friends present striking age patterns (Figure 3.18). 2022 data from 23 European OECD countries show that on average, young people aged 16 to 24 are four times more likely to get together with friends on a daily basis in a typical year, compared to older people aged 65 and up (34% vs. 8%); furthermore, young people are also four times more likely to contact their friends daily (64%, compared to 15% for those 65+). Evidence from many national data sources corroborates these findings: younger people spend the most amount of time with their friends, and older people spend the least (Goebel et al., 2019[40]; Insee, 2022[59]; ONS, 2024[60]; CBS, 2023[61]).

Figure 3.18. Compared to other age groups, young people have significantly higher rates of contacting friends and getting together with them in person
Copy link to Figure 3.18. Compared to other age groups, young people have significantly higher rates of contacting friends and getting together with them in person

Share of respondents who _____ with friends or family on a daily basis in an average year, OECD EU-EFTA 23, 2022

Longitudinal research has corroborated the finding that time spent with friends declines with age: during middle-age and adulthood, life events such as marriage, divorce and/or childbirth diminish peoples’ time with friends, and in old age the decline is linked to deteriorating physical health and the death of others in one’s social circle, including the death of a spouse (Augustsson et al., 2025[62]; Wrzus et al., 2013[63]; Pahl and Pevalin, 2005[64]; Sander, Schupp and Richter, 2017[65]). Time spent with family, however, is more or less stable as people age (Sander, Schupp and Richter, 2017[65]).

The same evidence from 23 European OECD countries shows that in 2022, older people were the loneliest age group, and that the prevalence of feeling lonely increases with age: 8.4% of those 65 years and up reported having felt lonely most or all of the time over the past week, on average (Figure 3.19). This average pattern is reversed in Denmark, Sweden, Switzerland and Ireland, where the youngest age cohort reports the highest levels of feeling lonely. Indeed, this reflects age patterns seen in other international surveys, as well as national data sources (Box 3.5). On the one hand, this could be a reflection of different age patterns in different countries, cultural groups or regions – that is, in certain places younger people may be more vulnerable to feelings of loneliness. On the other hand, this could reflect underlying structural changes that are making young people everywhere more at-risk to feeling lonely, with these trends already visible in some countries and not yet in others, though they may arise in future. On-going monitoring of trends, and longer time series, are needed to better understand these dynamics (see Chapter 4 for a longer discussion on trends in youth outcomes).

Figure 3.19. In almost all European OECD countries, older people are the loneliest age group
Copy link to Figure 3.19. In almost all European OECD countries, older people are the loneliest age group

Share of respondents who felt lonely most or all of the time over the past 4 weeks, OECD EU-EFTA 23, 2022

Box 3.5. National spotlight: Lonely young people in Canada, England, New Zealand and the United States

Copy link to Box 3.5. National spotlight: Lonely young people in Canada, England, New Zealand and the United States

Data from national surveys in Canada, the United States, England and New Zealand all show visible age gradients in loneliness outcomes (Figure 3.20).

Figure 3.20. In Canada, England, New Zealand and the United States, young people are much lonelier than older people
Copy link to Figure 3.20. In Canada, England, New Zealand and the United States, young people are much lonelier than older people

In all four countries, in recent years younger age groups have reported the highest levels of loneliness, with a general pattern of declining loneliness for each subsequent age cohort. (There are slight deviations from this pattern in England, where the 75+ age group is lonelier than the 65-74 group, and the difference between the 16 to 24 and 25 to 34 age groups are not statistically different from one another; in New Zealand age cohorts do not follow a linear descending pattern. In both instances, however, younger age groups are lonelier than older ones).

Differences in outcomes between the old and young are particularly striking in Canada and the United States. In 2024, young people aged 15 to 24 in Canada were almost twice likely as older people 65+ to feel lonely always or often (Figure 3.20, Panel A). Similarly, and in the same year, in the United States 18-to-29-year-olds were at least twice as likely to feel lonely compared to older groups, starting at age 50, with the age gap particularly striking vis-a-vis those aged 80+ (Figure 3.20, Panel D).

Pooled annual data from 2017-2023 for all OECD countries show that, on average, younger people are the age group most likely to report having social support, with only 5% saying they have no one to count on in times of need (Figure 3.21). Conversely, 11% of those aged 50 to 64, and 10% of older people aged 65 and up say they have no social support available. Once again, however, more recent data from the United States suggests that young people are the least likely to report having the social support they need, with the oldest age cohorts reporting the best outcomes (U.S. Census Bureau, 2024[45]).

Figure 3.21. Young people are least likely to say they have no one to count on in times of need
Copy link to Figure 3.21. Young people are least likely to say they have no one to count on in times of need

Share of respondents who have no one to count on in times of need, by age, OECD, 2017-2023

Gender

Copy link to Gender

Gender patterns in social connections outcomes are less clear than comparisons between other population groups, and vary by the outcome considered. Across OECD countries, women tend to be in contact with others at higher rates than men, and get together in person with family members more frequently than men do. However, women report seeing friends less frequently than men – and some data suggest women say they have fewer close friends. Compared to men, women also report very slightly higher rates of feeling lonely, and are more likely to be dissatisfied with their romantic relationships. Still, men are less likely to report having various types of social support. All of these differences are statistically significant, however the magnitudes are small in size compared to the other socio-economic and demographic differences in social connections outcomes discussed in the previous sections.

Women have closer relationships with family members, and interact with family more frequently than men do – both in person, and remotely. Data from European OECD countries reveal that in 2022, 11% of women got together with family members who do not live in the same household on at least a daily basis in an average year, compared to 9% of men; furthermore, 34% of women contacted family members daily, compared to 23% of men (Figure 3.22, Panel A). These findings align with other evidence from European OECD countries showing that women report having 4.9 close family members, compared to 4.6 for men (Figure 3.22, Panel B). The size of these gender differences is not large, but illustrates a consistent pattern of slightly stronger family bonds for women (and perhaps also family responsibilities – which may feed into the higher rates of feeling lonely among women discussed later on).

In recent years, particularly in the United States and United Kingdom, much has been made of a “male friendship crisis”, or “male friendship recession” (Hirsch, 2022[66]; Pearson, 2022[67]; Wollaston, 2023[68]). Academic research in these countries has pointed to men having fewer close friends than in the past, with implications for their physical, mental and social well-being. Indeed, across all OECD countries, where trend data are available, social connections outcomes for men have been declining – especially in more recent years (from 2018/2019 to 2022) – and have deteriorated more than those for women (refer to OECD (2024[3]) as well as Chapter 4). Nevertheless, data from European OECD countries do not indicate that levels of social contact, or numbers of close friends, are lower for men than for women, a priori: for example, in 2022, 13% of men reported getting together with friends on a daily basis in an average year, similar or slightly higher than the share of women (10%) (Figure 3.22, Panel A).12 And, on average in 22 European OECD countries, men report having 4.7 close friends compared to 3.7 close friends for women (Figure 3.22, Panel B).13

The small gender differences – in favour of men – with regard to spending time with friends may also at least in part reflect the larger responsibility for childcare and domestic tasks that fall on women, which could preclude their having time available to spend with friends, rather than family (OECD, 2021[69]; Craig and Mullan, 2013[70]; OECD, 2020[71]). Furthermore, these data were collected in 2022, when the effects of the pandemic were still being felt, especially for parents of school-aged children who were dealing with the impacts of remote schooling, the social and behavioural effects that had on young people, all while balancing their own work and family obligations – with women more likely to take on care duties than men (OECD, 2021[2]; 2021[72]).

Figure 3.22. Data from European OECD countries show that women have closer relationships with family, but men are more likely to see friends in person and report a higher number of close friends
Copy link to Figure 3.22. Data from European OECD countries show that women have closer relationships with family, but men are more likely to see friends in person and report a higher number of close friends

In addition, despite reporting more interaction with friends, and a higher number of close friends, men report lower levels of social support, suggesting their needs may not be (fully) met in these friendships. Across OECD countries in 2022-2023, women were on average slightly (but significantly) more likely to report having social support than men: 91% of women say they have friends or family to count on in times of need, compared to 90% of men (Figure 3.23, Panel A). Data from 22 European OECD countries further underscore the gender difference in social support (Figure 3.23, Panel B). On average, men are more likely than women to report having no one to provide three of four different types of social support: having no one to help if you were sick and confined to bed (13% of men report having no one, compared to 10% of women), no one with whom to share private worries and fears (9% men vs. 7% women) and no one to do something enjoyable with (5% men vs 4% women). This mismatch between feelings of loneliness and social support – with more women reporting feeling lonely, but more men lacking social and emotional support – has been echoed in research findings elsewhere (Goddard and Parker, 2025[73]).

Figure 3.23. Women are more likely to say they have friends or family to count on in times of need; in European OECD countries, men are more likely to report having no one to provide all types of social support
Copy link to Figure 3.23. Women are more likely to say they have friends or family to count on in times of need; in European OECD countries, men are more likely to report having no one to provide all types of social support

In survey data used in this report, women report very slightly higher levels of feeling lonely than men.14 For example, in 25 OECD countries with comparable data from 2022-2023, on average 7% of women report having felt lonely most or all of the time over the past four weeks, compared to 5% of men (Figure 3.24).

Figure 3.24. Women are slightly more likely to report feeling lonely in almost all OECD countries with comparable data
Copy link to Figure 3.24. Women are slightly more likely to report feeling lonely in almost all OECD countries with comparable data

Share of respondents who felt lonely “all” or “most of the time” over the past 4 weeks, OECD 25, 2022/2023

National data from the United States, England, Canada and Colombia also show small but consistently higher rates of loneliness for women (U.S. Census Bureau, 2024[45]; ONS, 2024[60]; Statistics Canada, 2025[41]; DANE, n.d.[75]). This may reflect genuine gender differences in outcomes for men vs. women, but could also be capturing response behaviours that vary by gender: for example, some research has shown that men have stronger perceptions of the stigma of loneliness than women, and are more reluctant to self-report loneliness, especially via direct questions (Langenkamp and Schobin, 2024[76]). Regardless, not all countries follow this pattern: data from Japan show that men report slightly higher levels of loneliness than women, and the problem of lonely and socially isolated young men is a significant policy issue (Box 3.6).

Box 3.6. National spotlight: Loneliness and social isolation for Japanese men

Copy link to Box 3.6. National spotlight: Loneliness and social isolation for Japanese men

Social isolation among men in Japan has been a policy concern for years. Hikikomori are individuals who withdraw from society, isolating themselves in their rooms for months or years at a time, without attending school or work. Hikikomori spend their days watching television, gaming or online, spending little time speaking to others and thus lose contact with friends as time passes. Many are affected by mental ill-health, including anxiety and mental exhaustion.

Data from 2022 estimate that 2% of the Japanese population are affected, with young and retired men over-represented. In response, in 2024 the Japanese government introduced a law recognising loneliness and isolation as national priorities. To reintegrate hikikomori into society, more specifically, the government has rolled out public awareness campaigns, created hikikomori support centres and established an online platform, the hikikomori voice station.

Source: OECD (2025[77]), “Supporting Japanese people affected by severe social isolation: A case study”, OECD Publishing: Paris. https://www.oecd.org/en/blogs/2025/03/supporting-opportunities-insights-from-communities/supporting-japanese-people-affected-by-severe-social-isolation-a-case-study.html.

When it comes to evaluations of the quality of their relationships, women are slightly more likely than men to report feeling satisfied: on average, across 26 OECD countries with comparable data, women rate satisfaction with their relationships with family, friends, neighbours and other people they know an 8.1 on a scale from 0 (not at all satisfied) to 10 (completely satisfied), compared to 8.0 for men (Figure 3.25, Panel A). The difference, while small in size, is statistically significant. However, when focusing only on romantic partners, specifically, women report worse outcomes than men. Data from 22 European OECD countries show that women report their satisfaction with their partners a 7.5 on a 1-10 scale, compared to 7.8 for men (Figure 3.25, Panel B).

Figure 3.25. Women are more likely to feel satisfied with their personal relationships, but more likely to feel dissatisfied with romantic relationships, specifically
Copy link to Figure 3.25. Women are more likely to feel satisfied with their personal relationships, but more likely to feel dissatisfied with romantic relationships, specifically

Care work responsibility

It is well documented that women generally take on a larger share of care responsibilities – spending more time on childcare and taking care of other dependents (including elderly, or ill, relatives), in addition to spending more time on household chores (OECD, 2021[69]; Craig and Mullan, 2013[70]; OECD, 2020[71]). Research has also consistently found that care work is associated with increased levels of loneliness and social isolation (Brimblecombe and Cartagena Farias, 2022[79]; Levine, 1999[80]; Sun, Finkelstein and Ouchida, 2019[81]; Nowland et al., 2021[82]; U.S. Surgeon General, 2025[83]). Data from New Zealand shed light on how the responsibility of childcare, disaggregating by types of parenting arrangements, can impact feelings of loneliness (Box 3.7).

Box 3.7. National spotlight: New Zealand data on the role of parenting in childcare and loneliness outcomes

Copy link to Box 3.7. National spotlight: New Zealand data on the role of parenting in childcare and loneliness outcomes

Evidence from New Zealand highlights the complex dynamics of parenting and loneliness (Figure 3.26). On average, male and female parents are slightly less lonely than their non-parent counterparts: 3.5% of female parents are lonely compared to 4.7% of non-parent females, and 2.6% of male parents are lonely vs. 3.7% of male non-parents. However, parenting in a couple is an important mediating factor. Single parents are lonelier than all other groups, with 8.3% reporting they felt lonely most or all of the time over the past four weeks.

Figure 3.26. Single parents are lonelier than both parents in a relationship, and non-parents
Copy link to Figure 3.26. Single parents are lonelier than both parents in a relationship, and non-parents

Share of respondents who felt lonely most or all of the time over the past 4 weeks, by parent type, New Zealand, 2023

References

[38] Albert, I. (2021), “Perceived loneliness and the role of cultural and intergenerational belonging: the case of Portuguese first-generation immigrants in Luxembourg”, European Journal of Ageing, Vol. 18/3, pp. 299-310, https://doi.org/10.1007/S10433-021-00617-7/.

[55] Arpino, B. et al. (2022), “Loneliness before and during the COVID-19 pandemic—are unpartnered and childless older adults at higher risk?”, European Journal of Ageing, Vol. 19/4, pp. 1327-1338, https://doi.org/10.1007/s10433-022-00718-x.

[62] Augustsson, E. et al. (2025), “Friends and trends: Friendship across life phases and cohorts”, Archives of Gerontology and Geriatrics, Vol. 135, p. 105872, https://doi.org/10.1016/J.ARCHGER.2025.105872.

[33] AXA (2023), Mind Health Report 2023, AXA Global Healthcare, https://www.axaglobalhealthcare.com/en/wellbeing/emotional/mind-health-report/.

[85] Balestra, C. and L. Fleischer (2018), “Diversity statistics in the OECD: How do OECD countries collect data on ethnic, racial and indigenous identity?”, ECD Statistics Working Papers 2018/09, https://doi.org/10.1787/89bae654-en (accessed on 4 June 2021).

[30] Barjaková, M., A. Garnero and B. d’Hombres (2023), “Risk factors for loneliness: A literature review”, Social Science & Medicine, Vol. 334, p. 116163, https://doi.org/10.1016/j.socscimed.2023.116163.

[57] Bower, M. et al. (2023), “The impact of the built environment on loneliness: A systematic review and narrative synthesis”, Health & Place, Vol. 79, p. 102962, https://doi.org/10.1016/j.healthplace.2022.102962.

[79] Brimblecombe, N. and J. Cartagena Farias (2022), “Inequalities in unpaid carer’s health, employment status and social isolation”, Health and Social Care in the Community, Vol. 30/6, pp. e6564-e6576, https://doi.org/10.1111/HSC.14104;PAGE:STRING:ARTICLE/CHAPTER.

[16] Cavicchiolo, E. et al. (2022), “Adolescents’ characteristics and peer relationships in class: A population study”, International Journal of Environmental Research and Public Health 2022, Vol. 19, Page 8907, Vol. 19/15, p. 8907, https://doi.org/10.3390/IJERPH19158907.

[61] CBS (2023), Sociale samenhang en welzijn, Centraal Bureau voor de Statistiek, https://www.cbs.nl/nl-nl/onze-diensten/methoden/onderzoeksomschrijvingen/korte-onderzoeksomschrijvingen/sociale-samenhang-en-welzijn.

[18] Chatzitheochari, S. and S. Arber (2012), “Class, gender and time poverty: A time-use analysis of British workers’ free time resources”, The British Journal of Sociology, Vol. 63/3, pp. 451-471, https://doi.org/10.1111/J.1468-4446.2012.01419.X.

[70] Craig, L. and K. Mullan (2013), “Parental Leisure Time: A Gender Comparison in Five Countries”, Social Politics: International Studies in Gender, State & Society, Vol. 20/3, pp. 329-357, https://doi.org/10.1093/SP/JXT002.

[75] DANE (n.d.), Encuesta Pulso Social (database), Departamento Administrativo Nacional de Estadística, https://www.dane.gov.co/index.php/estadisticas-por-tema/encuesta-pulso-social.

[86] Data for Good at Meta (2022), Social Connections Survey, https://data.humdata.org/dataset/social-connections-survey.

[17] Davis, J. et al. (2023), “Socialization, relaxation, and leisure across the day by social determinants of health: Results from the American Time Use Survey, 2014–2016”, Healthcare 2023, Vol. 11, Page 1581, Vol. 11/11, p. 1581, https://doi.org/10.3390/HEALTHCARE11111581.

[44] DCMS (2024), Community Life Survey 2023/24: Loneliness and support networks, Department for Culture, Media & Sport, https://www.gov.uk/government/statistics/community-life-survey-202324-annual-publication/community-life-survey-202324-loneliness-and-support-networks--2.

[39] De Jong Gierveld, J., S. Van der Pas and N. Keating (2015), “Loneliness of older immigrant groups in Canada: Effects of ethnic-cultural background”, Journal of Cross-Cultural Gerontology, Vol. 30/3, pp. 251-268, https://doi.org/10.1007/S10823-015-9265-X/.

[93] DHOMBRES, B. et al. (2024), Loneliness and social media use in the European Union, https://publications.jrc.ec.europa.eu/repository/handle/JRC135806 (accessed on 1 April 2025).

[4] Dykstra, P. (2009), “Older adult loneliness: Myths and realities”, European Journal of Ageing, Vol. 6/2, pp. 91-100, https://doi.org/10.1007/S10433-009-0110-3/.

[36] Eralba, C. and E. Barbiano Di Belgiojoso (2021), “Loneliness among migrants in Italy: Risks and protectors”, Migration Letters, Vol. 18/6, pp. 637-648, https://migrationletters.com/index.php/ml/article/view/1330 (accessed on 9 March 2025).

[19] European Commission and Joint Research Centre (JRC) (2024), EU Loneliness Survey, Joint Research Centre (JRC) [Dataset] PID:, https://doi.org/10.2760/2083329.

[12] Eurostat (n.d.), European Union Statistics on Income and Living Conditions (EU-SILC) – Scientific Use File (SUF) (database), https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and-living-conditions.

[91] Eurostat (n.d.), European Union Statistics on Income and Living Conditions (EU-SILC) – Scientific Use File (SUF) (database), https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and-living-conditions (accessed on 10 June 2022).

[49] Fingerman, K. et al. (2021), “Living alone during COVID-19: Social contact and emotional well-being among older adults”, The Journals of Gerontology: Series B, Vol. 76/3, pp. e116-e121, https://doi.org/10.1093/GERONB/GBAA200.

[20] Gallup (n.d.), Gallup World Poll, https://www.gallup.com/analytics/318875/global-research.aspx.

[73] Goddard, I. and K. Parker (2025), Men, Women and Social Connections, https://www.pewresearch.org/social-trends/2025/01/16/men-women-and-social-connections/?utm_source=Pew+Research+Center&utm_campaign=b793eaa086-SDT_2025_02_20&utm_medium=email&utm_term=0_-0056404212-401039990.

[40] Goebel, J. et al. (2019), “The German socio-economic panel (SOEP)”, Jahrbücher für Nationalökonomie und Statistik, Vol. 239/2, pp. 345-360, https://doi.org/10.1515/JBNST-2018-0022.

[35] González, V. and L. Castro (2007), “Keeping strong connections to the homeland via web-based tools: The case of Mexican migrant communities in the United States”, Journal of Community Informatics, Vol. 3/3, https://www.researchgate.net/publication/220536767_Keeping_Strong_Connections_to_the_Homeland_via_Web-based_Tools_The_Case_of_Mexican_Migrant_Communities_in_the_United_States.

[24] Hawkley, L. (2008), “From social structural factors to perceptions of relationship quality and loneliness: The Chicago health, aging, and social relations study”, The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, Vol. 63/6, pp. S375-S384, https://doi.org/10.1093/geronb/63.6.s375.

[56] Hawkley, L. and M. Kocherginsky (2017), “Transitions in loneliness among older adults: A 5-year follow-up in the National Social Life, Health, and Aging Project”, Research on Aging, Vol. 40/4, pp. 365-387, https://doi.org/10.1177/0164027517698965.

[66] Hirsch, A. (2022), “Men have fewer friends than ever. Why?”, Vox, https://www.vox.com/the-highlight/23323556/men-friendship-loneliness-isolation-masculinity.

[15] Hjalmarsson, S. and C. Mood (2015), “Do poorer youth have fewer friends? The role of household and child economic resources in adolescent school-class friendships”, Children and Youth Services Review, Vol. 57, pp. 201-211, https://doi.org/10.1016/J.CHILDYOUTH.2015.08.013.

[26] Holt-Lunstad, J. et al. (2015), “Loneliness and social isolation as risk factors for mortality: A meta-analytic review”, Perspectives on Psychological Science, Vol. 10/2, pp. 227-237, https://doi.org/10.1177/1745691614568352.

[87] INEGI (2021), Encuesta Nacional de Bienestar Autorreportado: ENBIARE 2021, Instituto Nacional de Estadística y Geografía (INEGI), https://www.inegi.org.mx/contenidos/programas/enbiare/2021/doc/enbiare_2021_presentacion_resultados.pdf.

[59] Insee (2022), L’enquête Statistiques sur les ressources et conditions de vie, Institut national de la statistique et des études économiques, https://www.insee.fr/fr/information/4230346.

[9] Istat (2012), Social condition and integration of foreign citizens, Istituto Nazionale di Statistica, https://www.istat.it/en/microdata/social-condition-and-integration-of-foreign-citizens/.

[7] Kennedy, L., O. Field and K. Barker (2019), Barriers to belonging: An exploration of loneliness among people from Black, Asian and Minority Ethnic backgrounds, British Red Cross, https://www.redcross.org.uk/about-us/what-we-do/we-speak-up-for-change/barriers-to-belonging.

[58] Kępkowicz, A. and D. Mantey (2016), “Suburbs as none-places? The need for gathering places: A case study of the Warsaw suburbs, Poland”, Architektura Krajobrazu, Vol. 1, pp. 44-67, https://bibliotekanauki.pl/articles/1191236 (accessed on 10 March 2025).

[53] Klinenberg, E. (2016), “Social isolation, loneliness, and living alone: Identifying the risks for public health”, American Journal of Public Health, Vol. 106/5, p. 786, https://doi.org/10.2105/AJPH.2016.303166.

[23] Klinenberg, E. (2016), “Social Isolation, Loneliness, and Living Alone: Identifying the Risks for Public Health”, American Journal of Public Health, Vol. 106/5, p. 786, https://doi.org/10.2105/AJPH.2016.303166.

[48] Klinenberg, E. (2013), Going Solo: The extraordinary rise and surprising appeal of living alone, Penguin Books, https://www.penguinrandomhouse.com/books/305485/going-solo-by-eric-klinenberg/ (accessed on 18 September 2025).

[6] Koelet, S. and H. de Valk (2016), “Social networks and feelings of social loneliness after migration: The case of European migrants with a native partner in Belgium”, Ethnicities, Vol. 16/4, pp. 610-630, https://doi.org/10.1177/1468796816638398.

[27] Kunze, L. and N. Suppa (2017), “Bowling alone or bowling at all? The effect of unemployment on social participation”, Journal of Economic Behavior & Organization, Vol. 133, pp. 213-235, https://doi.org/10.1016/J.JEBO.2016.11.012.

[76] Langenkamp, A. and J. Schobin (2024), “Revealing loneliness: Disentangling the interaction of gender and community stigma”, Social Science Quarterly, Vol. 105/5, pp. 1529-1543, https://doi.org/10.1111/ssqu.13434.

[80] Levine, C. (1999), “The Loneliness of the Long-Term Care Giver”, New England Journal of Medicine, Vol. 340/20, pp. 1587-1590, https://doi.org/10.1056/NEJM199905203402013.

[14] Lobo, F. (1996), “Les effets sur le style de vie du chômage en fin de carrier”, Loisir et Societe, Vol. 19/1, pp. 169-199, https://doi.org/10.1080/07053436.1996.10715517.

[88] Lykes, V. and M. Kemmelmeier (2014), “What predicts loneliness? Cultural difference between individualistic and collectivistic societies in Europe”, Journal of Cross-Cultural Psychology, Vol. 45/3, pp. 468-490, https://doi.org/10.1177/0022022113509881.

[92] Macdonald, S., J. Nixon and L. Deacon (2018), “‘Loneliness in the city’: examining socio-economics, loneliness and poor health in the North East of England”, Public Health, Vol. 165, pp. 88-94, https://doi.org/10.1016/J.PUHE.2018.09.003.

[25] Macdonald, S., J. Nixon and L. Deacon (2018), “‘Loneliness in the city’: Examining socio-economics, loneliness and poor health in the North East of England”, Public Health, Vol. 165, pp. 88-94, https://doi.org/10.1016/J.PUHE.2018.09.003.

[89] Maes, M. et al. (2019), “Gender differences in loneliness across the lifespan: A meta-analysis”, European Journal of Personality, Vol. 33/6, pp. 642-654, https://doi.org/10.1002/PER.2220.

[74] Ministerio de Desarrollo Social y Familia (2021), Encuesta de Bienestar Social, http://observatorio.ministeriodesarrollosocial.gob.cl/encuesta-bienestar-social.

[29] Morrish, N. and A. Medina-Lara (2021), “Does unemployment lead to greater levels of loneliness? A systematic review”, Social Science & Medicine, Vol. 287, p. 114339, https://doi.org/10.1016/J.SOCSCIMED.2021.114339.

[31] Morrish, N., R. Mujica-Mota and A. Medina-Lara (2022), “Understanding the effect of loneliness on unemployment: Propensity score matching”, BMC Public Health, Vol. 22/1, https://doi.org/10.1186/S12889-022-13107-X.

[5] Motmans, J. et al. (2011), “Equal is not enough: Challenging differences and inequalities in contemporary societies”, Policy Research Centre on Equal Opportunities (PRCEO), Vol. Conference proceedings, https://lib.ugent.be/catalog/rug01:001696428.

[34] Navarrete, C. and E. Huerta (20006), “Building virtual bridges to home: The use of the Internet by transnational communities of immigrants”, International Journal of Communications, Law and Policy, https://papers.ssrn.com/abstract=949626.

[82] Nowland, R. et al. (2021), “Experiencing loneliness in parenthood: A scoping review”, Perspectives in Public Health, Vol. 141/4, pp. 214-225, https://doi.org/10.1177/17579139211018243/.

[77] OECD (2025), “Supporting Japanese people affected by severe social isolation: A case study”, https://www.oecd.org/en/blogs/2025/03/supporting-opportunities-insights-from-communities/supporting-japanese-people-affected-by-severe-social-isolation-a-case-study.html.

[3] OECD (2024), How’s Life? 2024: Well-being and Resilience in Times of Crisis, OECD Publishing, Paris, https://doi.org/10.1787/90ba854a-en.

[54] OECD (2024), OECD Affordable Housing Database - indicator HM1.4 Living arrangements by age groups, https://oe.cd/ahd.

[47] OECD (2024), Society at a Glance 2024: OECD Social Indicators, OECD Publishing, Paris, https://doi.org/10.1787/918d8db3-en.

[69] OECD (2021), Bringing Household Services Out of the Shadows: Formalising Non-Care Work in and Around the House, Gender Equality at Work, OECD Publishing, Paris, https://doi.org/10.1787/fbea8f6e-en.

[72] OECD (2021), “Caregiving in crisis: Gender inequality in paid and unpaid work during COVID-19”, Tackling Coronavirus (COVID-19): Contributing to a global effort, https://doi.org/10.1787/3555d164-en.

[2] OECD (2021), COVID-19 and Well-being: Life in the Pandemic, OECD Publishing, Paris, https://doi.org/10.1787/1e1ecb53-en.

[71] OECD (2020), How’s Life? 2020 - Measuring Well-being, OECD Publishing, Paris, https://doi.org/10.1787/9870c393-en.

[78] OECD (n.d.), How’s Life? Well-being Database, http://data-explorer.oecd.org/s/fu.

[60] ONS (2024), Community Life Survey 2023/24 annual publication, Office for National Statistics, https://www.gov.uk/government/statistics/community-life-survey-202324-annual-publication (accessed on 10 April 2025).

[51] O’Súilleabháin, P., S. Gallagher and A. Steptoe (2019), “Loneliness, living alone, and all-cause mortality: The role of emotional and social loneliness in the elderly during 19 years of follow-up”, Psychosomatic Medicine, Vol. 81/6, pp. 521-526, https://doi.org/10.1097/PSY.0000000000000710.

[64] Pahl, R. and D. Pevalin (2005), “Between family and friends: A longitudinal study of friendship choice”, The British Journal of Sociology, Vol. 56/3, pp. 433-450, https://doi.org/10.1111/J.1468-4446.2005.00076.X.

[50] Pauly, T. et al. (2016), “How we experience being alone: Age differences in affective and biological correlates of momentary solitude”, Gerontology, Vol. 63/1, pp. 55-66, https://doi.org/10.1159/000450608.

[67] Pearson, C. (2022), “Why Is It So Hard for Men to Make Close Friends?”, The New York Times, https://www.nytimes.com/2022/11/28/well/family/male-friendship-loneliness.html.

[28] Pohlan, L. (2019), “Unemployment and social exclusion”, Journal of Economic Behavior & Organization, Vol. 164, pp. 273-299, https://doi.org/10.1016/J.JEBO.2019.06.006.

[46] Ruprecht, M. et al. (2024), ““Being queer, it was really isolating”: Stigma and mental health among lesbian, gay, bisexual, transgender, and queer (LGBTQ+) young people during COVID-19”, Health Education & Behavior, Vol. 51/4, pp. 521-532, https://doi.org/10.1177/10901981241249973.

[65] Sander, J., J. Schupp and D. Richter (2017), “Getting together: Social contact frequency across the life span”, Developmental Psychology, Vol. 53/8, pp. 1571-1588, https://doi.org/10.1037/DEV0000349.

[21] Schnepf, S., B. d’Hombres and C. Mauri (2024), Loneliness in Europe: Determinants, risks and interventions, Springer Cham, https://link.springer.com/book/10.1007/978-3-031-66582-0.

[41] Statistics Canada (2025), Loneliness by gender and other selected sociodemographic characteristics, https://open.canada.ca/data/dataset/277e3275-5b97-4b2b-bf59-59af72541bd7.

[43] Statistics Canada (2021), Visible minority of person, https://www23.statcan.gc.ca/imdb/p3Var.pl?Function=DEC&Id=45152.

[90] Statistics Norway (2023), Lavere livskvalitet blant ikke-heterofile, https://www.ssb.no/sosiale-forhold-og-kriminalitet/levekar/artikler/lavere-livskvalitet-blant-ikke-heterofile.

[42] Stats NZ (2024), Wellbeing statistics: 2023, https://www.stats.govt.nz/information-releases/wellbeing-statistics-2023/.

[81] Sun, A., E. Finkelstein and K. Ouchida (2019), “Social isolation, loneliness, and caregiver burden among paid and unpaid caregivers of homebound older adults”, Innovation in Aging, Vol. 3/Suppl 1, p. S967, https://doi.org/10.1093/GERONI/IGZ038.3505.

[8] Taylor, H. et al. (2024), “Racial/ethnic differences in loneliness among older adults: The role of income and education as mediators”, Innovation in Aging, Vol. 8/8, https://doi.org/10.1093/GERONI/IGAE068.

[37] Ten Kate, R. et al. (2020), “A closer look at loneliness: Why do first-generation migrants feel more lonely than their native Dutch counterparts?”, The Gerontologist, Vol. 60/2, pp. 291-301, https://doi.org/10.1093/GERONT/GNZ192.

[45] U.S. Census Bureau (2024), Household Pulse Survey, https://www.census.gov/programs-surveys/household-pulse-survey.html.

[83] U.S. Surgeon General (2025), Parents Under Pressure: The U.S. Surgeon General’s Advisory on the Mental Health & Well-Being of Parents, Office of Health and Human Services (HHS), https://www.hhs.gov/sites/default/files/parents-under-pressure.pdf.

[32] Üniversity, A. et al. (2025), “Breaking the cycle: Exploring the bidirectional link between loneliness and unemployment”, Journal of Life Economics, Vol. 12/1, pp. e2697-e2697, https://doi.org/10.15637/JLECON.2697.

[1] Webel, A. et al. (2015), “Social resources, health promotion behavior, and quality of life in adults living with HIV”, Applied Nursing Research, Vol. 30, p. 204, https://doi.org/10.1016/J.APNR.2015.08.001.

[22] WHO (2025), From Loneliness to Social Connection: Charting a Path to Healthier Societies, World Health Organization (WHO) Commission on Social Connection, https://www.who.int/publications/i/item/978240112360.

[68] Wollaston, S. (2023), “‘We don’t do deep emotional discussions’: Why men lose their friends – and how they can make more”, The Guardian, https://www.theguardian.com/lifeandstyle/2023/may/25/we-dont-do-deep-emotional-discussions-why-men-lose-their-friends-and-how-they-can-make-more.

[63] Wrzus, C. et al. (2013), “Social network changes and life events across the life span: A meta-analysis”, Psychological Bulletin, Vol. 139/1, pp. 53-80, https://psycnet.apa.org/buy/2012-13785-001.

[10] Yang, K. (2023), Analysing intersectionality : A toolbox of methods, Sage, https://uk.sagepub.com/en-gb/eur/analysing-intersectionality/book279536.

[11] Yang, K. (2023), “Loneliness and intersectionality: A progressive conditional approach”, Current Research in Behavioral Sciences, Vol. 5, pp. 2666-5182, https://doi.org/10.1016/J.CRBEHA.2023.100122.

[84] Yang, K. and C. Victor (2011), “Age and loneliness in 25 European nations”, Ageing & Society, Vol. 31/8, pp. 1368-1388, https://doi.org/10.1017/S0144686X1000139X.

[52] Yeh, S. and S. Lo (2004), “Living alone, social support, and feeling lonely among the elderly”, Social Behavior and Personality, Vol. 32/2, pp. 129-138, https://doi.org/10.2224/SBP.2004.32.2.129.

[13] Zuzanek, J. and M. Hilbrecht (2019), “Enforced leisure: Time use and its well-being implications”, Time & Society, Vol. 28/2, pp. 657-679, https://doi.org/10.1177/0961463X16678252.

Annex 3.A. Technical annex

Copy link to Annex 3.A. Technical annex

Annex Table 3.A.1 provides details on how each of the socio-demographic breakdowns presented in this chapter are defined by each source survey. Recall from the Reader’s Guide that results are only reported if sample sizes are sufficiently large enough to do so, defined as having at least 100 observations per population group / country / year grouping.

Population group	Definition used for EU-SILC Survey data	Definition used for Gallup World Poll data	Definition used for EU Loneliness Survey data	Definition used for AXA Mind Health Survey data
Education	Education categories correspond to ISCED levels 0-2 for “below upper secondary” level (i.e. less than primary, primary and lower secondary); 3-4 for “upper secondary” level (i.e. secondary and post-secondary non-tertiary education); and 5-8 for “tertiary” level	“Primary”, “secondary” and “tertiary” levels of education are pre-coded by Gallup. As per the Gallup codebook, primary education refers to having completed elementary education or less (up to eight years of basic education); secondary refers to having completed some secondary education and up to three years of tertiary education (9-15 years of education); and tertiary refers to having completed four years of education beyond “high school”, or having received a four-year college degree Unless otherwise noted, Gallup World Poll inequalities use 3-year pooled averages to ensure sufficiently large sample sizes	“Primary” refers to less than primary, primary and lower secondary education; “secondary” refers to upper secondary and post-secondary non-tertiary education; “tertiary” refers to tertiary education	NA
Income	Income quintiles are determined by dividing total annual household disposable income by 12, and dividing the resulting total into quintiles	Income quintiles are pre-cleaned by Gallup	NA	NA
Labour market	“Employed” refers to any form of employment (including self-employment). EU-SILC measures self-declared current ‘main activity status’. Respondents can consider themselves being employed irrespective of their official labour market status, working time or kind of income from employment. They can also be looking for another job in parallel. Also, other categories can apply to them as long as they consider employment to be their main activity. Persons who would choose another main activity status can also be in employment. For instance, many people who would regard themselves as full-time students or mainly fulfilling domestic tasks can have a job. In that case they can assign themselves to the corresponding category. Respondents helping in the family business, even if it is unpaid, can consider themselves as employed. Outcomes for retired and inactive respondents are not included in these comparisons	“Employed” refers to any form of employment (including self-employment and part-time employment) but does not include underemployment. Outcomes for those out of the workforce are not included Labour market inequalities use 7-year pooled averages to ensure sufficiently large sample sizes	NA	NA
Minority group	NA	NA	NA	Respondents are asked whether they consider themselves to be a part of a minority group, for any reason, including but not limited to one’s ethnicity or skin colour, language, disability, sexual orientation or gender identity, religion or belief, migrant status or political opinion. “Minority” refers to those who answer yes, and “non-minority” those who answer no.
Place of birth	The country of birth of an individual is defined as the country of usual residence (in its current boundaries) of the individual’s mother at the time of delivery. Information on the country of birth is used to distinguish between native-born (born in the reporting country) and born abroad (born in a country other than the reporting country) residents.	“Born abroad” refers to respondents whose place of birth is different from the country where the interview is being conducted; “born domestically” refers to respondents who were born in the country of the survey. Place of birth inequalities use 7-year pooled averages to ensure sufficiently large sample sizes	NA	NA
Place of birth of parents	Outcomes are only included for respondents who were themselves born in the country of the survey (i.e., no first-generation migrants). The variables country of birth of the father/mother report on the country of birth of the person’s father/mother, i.e., the country of usual residence (in its current borders, if the information is available) of the person’s father/mother at the time of the delivery, or, failing this, the country (in its current borders, if the information is available) in which the birth of the person’s father/mother took place. This information determines whether the person’s father/mother is native-born (born in the reporting country) or foreign-born (born in a country other than the reporting country). Outcomes are compared for respondents whose parents were both born abroad, and whose parents are both native-born; outcomes are not included for those who have a single parent born abroad.	NA	NA	NA
LGBTI	NA	NA	NA	Respondents who self-identify as LGBTQ+ (lesbian, gay, bisexual, transgender, queer and/or questioning) and those who do not
Living arrangement	“Live alone” refers to respondents who live in a single-person household; “live with others” refers to respondents in households with more than one person providing themselves with the essentials for living	“Live alone” refers to respondents who live in a single occupancy household; “live with others” refers to respondents in households with more than one person	“Live alone” refers to respondents who live in a single occupancy household; “live with others” refers to respondents in households with more than one person	NA
Relationship status	“In a relationship” refers to respondents who are married or in a consensual union (on either a legal, or non-legal basis); “single” refers to respondents who are separated, divorced, widowed or never married and are not in a consensual union	“In a relationship” refers to respondents who are married or have a domestic partner; “single” refers to those who are single / have never been married, are separated, divorced or widowed	“In a relationship” refers to respondents who are “married or cohabitating” or “in a relationship”; “single” refers to respondents who are single, separated or divorced or widowed	NA
Place or residence – urban/rural	“Urban” refers to cities, towns and suburbs, “rural” refers to rural areas	NA	“Urban” refers to a small or medium-sized town (50 000 people or less) or a large town/city (over 50 000 people); “rural” refers to a rural area or village	NA
Age	For the purposes of this publication, several age categories are created by grouping respondent self-reported age into the following groups: 16-24, 25-49, 50-64 and 65+	Age categories are created by grouping respondent self-reported age into the following groups: 16-24, 25-49, 50-64 and 65+ Age inequalities use 7-year pooled averages to ensure sufficiently large sample sizes	NA	NA
Gender	Respondent self-reported “male” vs. “female”	Respondent self-reported “male” vs. “female”	Respondents are asked “which of the following describes how you think of yourself?”	Refers to gender at birth, for which options are “male” or “female”

Notes

Copy link to Notes

← 1. Sample sizes in international household surveys often preclude investigating intersecting risk profiles –individuals who belong to two at-risk socio-demographic groups, such as being both unemployed and a first-generation migrant, or being elderly and living alone – especially when only a small share of the population is exposed to each risk. Nevertheless, large-scale national datasets often enable more granular assessments – see Box 3.1.

← 2. Sample sizes are sufficiently large to analyse outcomes for those with below upper secondary levels of education for the EU-SILC survey (23% of respondents in 23 European OECD countries have achieved less than upper secondary education in 2022). However, outcomes from both the 2022 EU Loneliness Survey and Gallup World Poll combine primary, lower and upper secondary educational attainment into a single group, given smaller sample sizes in these surveys.

← 3. Refer to the Chapter 2 Annex for a discussion of answer groupings for the variables relating to the frequency of in-person vs. remote social interactions with friends and family in a given year. In this chapter, for each socio-demographic group, outcomes are shown for those answering both that they socialise with friends or family “daily” in a given year, as well as “never”.

← 4. Future analysis could disaggregate income quintiles (and/or employment status) by age group, to understand whether retirement is at least partly affecting these results. Sample sizes within individual countries do not allow for this level of granularity in this report.

← 5. While the relationship between poverty, loneliness and social isolation is well documented, the (causal) mechanisms underpinning the relationship are less well understood, in part, perhaps, because financial outcomes are often used as control variables in research studies rather than the direct outcome measured (Barjaková, Garnero and d’Hombres, 2023[30]).

← 6. Cross-country statistics on minority groups are not widely available, given that what is deemed a minority in one country may not be in another; in addition, some OECD countries have explicit policies against measuring certain socio-demographic characteristics such as race and ethnicity in official surveys (Balestra and Fleischer, 2018[85]). For these outcomes, findings from individual OECD countries that do capture this information are presented. The most commonly collected information in a cross-country comparable way is a proxy for migration status: whether the respondent was born in the current country of residence, or born abroad. Findings based on first- and second-generation migration status are included here for European OECD countries, with supplemental findings from two non-European countries with official data.

← 7. In Norway, the annual Quality of Life survey allows for the disaggregation of well-being indicators by sexual orientation; findings for loneliness have not been analysed, but published analysis finds that non-heterosexuals in Norway report lower life satisfaction, have more negative emotions, and experience less meaning and engagement in comparison to heterosexuals, possibly due to socio-demographic factors including rates of cohabitation (Statistics Norway, 2023[90]).

← 8. Data from other sources – also collected in 2022 but with smaller samples and using different questions to capture social interactions – show less nuanced patterns. The 2022 Gallup State of Global Connections survey finds that people across OECD countries who live alone are more likely than those who live with others to have not interacted with friends or family who live nearby over the past week (9% vs. 4%) (Data for Good at Meta, 2022[86]); similarly, the 2022 EU Loneliness Survey finds that people who live alone are more likely to report never meeting family members in person (8% vs. 5%), or to speak with them over the phone (6% vs. 4%). However, similar to the findings in Figure 3.15, the EU Loneliness Survey shows that people who live alone are more likely to both never contact and get together with friends, as well as to report doing so weekly – that is, this survey also shows clustering at both extremes for different “types” of one-person households (European Commission and Joint Research Centre (JRC), 2024[19]).

← 9. Survey data on household composition do not necessarily specify with whom the respondent lives – nuclear family, extended family, friends or roommates.

← 10. These high rates of socialising with friends echo social connections outcomes for young people, described later in this chapter. This might also be partly driven by the fact that single people in the EU-SILC dataset are, on average, eight years younger than those in a relationship.

← 11. Data from the Mexican statistical office’s ENBIARE well-being survey supports this. In 2021, data show that among 71.5% of adults who report being in a relationship, 91.1% qualify their relationship as “good” (indicating their relationship fulfils at least one of the following quality criteria: admiration, mutual recognition and teamwork), however almost 1 in 10 (8.9%) report their relationship fulfils none of the criteria. There is a gender gap: women are less likely to be partnered (68.1% vs. 75.3% for men) as well as more likely to report a poor quality relationship (11.3% vs. 6.4%) (INEGI, 2021[87]).

← 12. Data from New Zealand (not pictured) does not show any significant gender differences in rates of face-to-face contact with friends at least once a week.

← 13. The difference in the share of men vs. women who report having no close friends, or no close family members, is not statistically significant.

← 14. Evidence from the academic literature finds inconclusive results in terms of loneliness and gender – and comparisons across studies suggest that the measurement tool used to measure loneliness may affect which gender is more likely to have a higher prevalence. Some studies using indirect measures of loneliness, such as the UCLA scale, have found men to be lonelier than women, while other studies that use a direct loneliness measure (such as the indicator shown in Figure 3.24) find either that women are lonelier, or find no significant differences (Schnepf, d’Hombres and Mauri, 2024[21]; Maes et al., 2019[89]; Lykes and Kemmelmeier, 2014[88]; Yang and Victor, 2011[84]).

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Countries

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Publications

Publications

Featured publications

Data

Data

Featured data

News & events

News & events

Featured events

About OECD

About

Engage with us

Work with us

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment