Retirement villages are proliferating across the western world (Jeste and Childers, Reference Jeste and Childers2017). Retirement villages can involve social renting, leaseholds, or mixed tenure (Gray and Worlledge, Reference Gray and Worlledge2018), and can be further characterized by the extent to which they are isolated or segregated within the broader community. A main motivating factor in relocating to a retirement village is social connection (Roy et al., Reference Roy, Dubé, Després, Freitas and Légaré2018). However, as Boyd and colleagues (Reference Boyd2020) found in their assessment of retirement village residents, levels of loneliness in this population remain high, and likely higher than those found in the broader older population (Gray, Reference Gray2015; Taylor et al., Reference Taylor, Wang and Morrow-Howell2018).

Understanding why loneliness is higher among retirement village residents, and how to intervene is therefore a priority. As well as being unpleasant in its own right, loneliness is associated with a multitude of negative health consequences such as early mortality and cognitive decline (Boss et al., Reference Boss, Kang and Branson2015; Holt-Lunstad et al., Reference Holt-Lunstad, Smith, Baker, Harris and Stephenson2015). Specifically among retirement housing residents, loneliness is associated with increased levels of anxiety and depression (Bekhet and Zauszniewski, Reference Bekhet and Zauszniewski2012).

As a component of a broader, three-phase intervention study, Boyd and colleagues sought to measure loneliness and its correlates among residents of 33 retirement villages in NZ using a standardized assessment tool and a companion survey. Participants were recruited using both randomized and non-randomized sampling methods. In multivariate analyses, loneliness as a binary outcome was more likely among residents who were widowed, divorced, separated, or never married (relative to those partnered), who had a poor or fair quality of life, who reported moving to the retirement village to seek social connection, and who were using Selective Serotonin Reuptake Inhibitors (SSRIs) or had a higher risk of depression. Loneliness was not associated with family relationships, gender, pain, or informal support receipt.

Boyd’s study (hereafter called the “original study”) is not without limitations. The cited registered intervention protocol does not mention loneliness, meaning that the current study may have been opportunistic, occluding the decision process in analytic design and hypothesis choice (see Gelman and Loken [Reference Gelman and Loken2013]). As the authors acknowledge, there are further limitations associated with using single-item loneliness measures. As well as the issue with bias caused by stigma, cited by the authors, single-item measures are unsuitable because loneliness is likely to be multidimensional (Weiss, Reference Weiss1973), thus not measurable with a single item. Longitudinal extension of this research is planned by the research team, which will overcome the limitation of the cross-sectional nature of the research (since loneliness is thought to have a bidirectional association with many of the correlates they explored). Methodologically, an absence of correction for multiple comparisons, particularly given the number of bivariate tests used, is likely to have inflated Type I error (as the authors acknowledge), an error type further inflated by the absence of preregistration of the study (Gelman and Loken, Reference Gelman and Loken2013).

Despite these limitations, results are thought-provoking, and some key questions arise from the original study. Participant mean age was 81.3, and 57% of participants were aged 80 or over. An inflection point in the relationship between age and loneliness may exist around this point in life (Pinquart and Sorensen, Reference Pinquart and Sorensen2001) such that these older individuals may be more likely to be lonely than their younger counterparts. As such it may be that retirement village residents are lonelier than their ordinary-housing counterparts because they are on average older (because of the non-randomized recruitment of some participants, it is not possible to meaningfully compare rates of loneliness in this study to those found in the broader aging population). A developmental perspective on loneliness in the retirement village context, taking into account age and life stage, would certainly yield further insights.

Another issue arising from the original study concerns the association between loneliness and motivations to relocate. Individuals who relocated to retirement villages in order to seek social connection (over 1/3 of respondents) were most likely to be lonely. Qualter and colleagues (Reference Qualter2015) have suggested that such social re-engagement (“reaffiliation”) is a key motivation among those wishing to offset the adverse experience of loneliness. Empirically, however, loneliness predicts later social disengagement in older adults (McHugh Power et al., Reference McHugh Power, Steptoe, Kee and Lawlor2019), creating a feedback loop (Hawkley and Cacioppo, Reference Hawkley and Cacioppo2010). Individuals who are motivated to relocate to a retirement village, however, may be particularly motivated to reaffiliate and thus avoid this feedback loop. In this scenario, it may take time for the relocation to reduce loneliness, particularly since transitions (such as moving) can cause temporary increases in loneliness (Lawson et al., Reference Lawson, Vines, Wilson, Barnett and Barreto2014). This explanation is corroborated by a longitudinal demonstration of a decrease in loneliness during the year post-relocation to a retirement village (Crisp et al., Reference Crisp, Windsor, Butterworth and Anstey2015). It would be of interest to explore (likely curvilinear) loneliness trajectories among those who cite seeking social connection as a motivation to relocate to a retirement village relative to those who did not (almost 2/3 of Boyd’s sample), in the context of theoretical work by Qualter and Cacioppo.

Another issue highlighted in the original study findings are the particularly at-risk group of widows and widowers. Residents of retirement villages are more likely than the general population to be bereaved, and a curvilinear trend between time since widowhood and loneliness was in evidence. Individuals who were widowed between 2 and 5 years previously were likeliest to be lonely. With these two points in mind, it is necessary to consider whether retirement village transitions are likely to help these individuals in their bereavement, and whether this is likely to, over time, reduce loneliness levels. Tailored support for these individuals is likely to be necessary in retirement villages.

A critical issue mentioned in the original study is the conflation of loneliness and social isolation. While social engagement may for some individuals help to ameliorate loneliness, this is not likely to be the case for all, and a significant minority of individuals experience loneliness even in the absence of social isolation. In fact, retirement villages themselves offer a unique opportunity to study loneliness in individuals who are not objectively socially isolated (Morlett Paredes et al., Reference Morlett Paredes2020). Living near members of one’s cohort, with increased opportunities for incidental social support, should theoretically reduce social isolation and increase social engagement. It does not necessarily follow that loneliness levels will also fall, since loneliness results from many more proximal causes (e.g. genetic and developmental factors, cognitive biases, and personality traits; (McHugh Power et al., Reference McHugh Power, Dolezal, Kee and Lawlor2018). It should be clear, then, that reducing social isolation will not necessarily have any impact on loneliness, and as such, more nuanced approaches to intervention are required.

Such interventions are likely, if successful, to be multifactorial (Saito et al., Reference Saito, Kai and Takizawa2012). While improving opportunities for social engagement will not itself reduce loneliness, more specific opportunities such as offering a wide range of social activities, fostering ongoing informal contact between neighbors, offering counseling for extremely isolated individuals, and offering group counseling on rebuilding personal friendships and social skills may offer productive ways forward to reduce loneliness in retirement village residents (Gray and Worlledge, Reference Gray and Worlledge2018; Morlett Paredes et al., Reference Morlett Paredes2020). Provision of tailored activities is likeliest to result in the shared empathic understanding that may alleviate loneliness most (van Staden and Coetzee, Reference van Staden and Coetzee2010).

Further research into loneliness in a retirement village context could consider these above points, acknowledged by the study authors, as well as a few additional avenues of exploration. For instance, such research should explore the role of nationality and culture on loneliness among residents of retirement villages. Nationality is a major predictor of loneliness levels (Yang and Victor, Reference Yang and Victor2011), and culture may impact the extent to which an individual experiences loneliness given constant social isolation levels (Rokach, Reference Rokach2007).

Further research would also be well advised to consider loneliness as a multidimensional phenomenon. Specifically, research into loneliness delineates two subtypes: social and emotional (Weiss, Reference Weiss1973), which are likely to require completely different interventions. Further exploration of rates of both among retirement village residents is warranted, particularly given the high levels of emotional loneliness likely to be found among widowed individuals. Another factor worth considering is the role of intergenerational friendships in preventing loneliness (Gaggioli et al., Reference Gaggioli2014; Jeste and Childers, Reference Jeste and Childers2017; O’Luanaigh and Lawlor, Reference O’Luanaigh and Lawlor2008) – if separated from the broader community, retirement villages have the potential to segregate older residents from younger generations.

Ultimately, the aim of research such as that conducted by Boyd and colleagues is to design effective interventions to reduce loneliness among those living in retirement villages. Existing theoretical perspectives on loneliness such as those described offer valuable bases on which to model new hypotheses and further empirical study. By designing and conducting theoretically informed studies on loneliness in retirement village residents, future researchers can inform the design of effective interventions, and provide new insight into the nature of loneliness and its causes.