No open-access location is known for this record yet.
No open-access location is known for this record yet.
Old age is often characterised by an increasing incidence of negative life events (Kok et al., 2021), with particularly common and impactful events being the onset of disability and widowhood. These events can substantially reduce life satisfaction among older adults (Ghose et al., 2025; Puvill et al., 2019; Switsers et al., 2021). However, individuals differ markedly in their life satisfaction trajectories surrounding these events: some maintain stable life satisfaction, others recover after a temporary decline, while some experience persistent reductions (Galatzer-Levy et al., 2018). Understanding which factors account for these individual differences is essential for developing targeted interventions to support well-being in later life. Prior research has linked a wide range of variables to late-life well-being and adaptation to negative life events, including: - sociodemographic characteristics such as sex, age, education, wealth, income, and region of residence (e.g., Dabergott, 2022; Prati, 2022; Puvill et al., 2019); - personality dispositions such as neuroticism, optimism, control beliefs, generalised trust and life goals (“values”) (e.g., Asselmann et al., 2021; Chopik et al., 2015; Gerstorf et al., 2014; Lee et al., 2022; Poulin & Haase, 2015); - social participation such as contact with friends and family, and religious and civic engagement (e.g., Berg et al., 2006; Sullivan & Infurna, 2020; Windsor et al., 2015); - and health and lifestyle factors such as subjective health, physical activity, and sleep satisfaction (e.g., Dumitrache et al., 2017; Lücke et al., 2022; Wang et al., 2023). However, most studies have examined these variables in isolation or small sets while assuming strict linearity of effects (Rolandi et al., 2024), leaving open questions about their relative importance and how potentially non-linear combinations of factors may jointly shape life satisfaction. Some evidence suggests that psychosocial factors interact. For instance, social ties may buffer the impact of lower wealth on well-being in older adults (cf., Niedzwiedz et al., 2016), while a combination of high neuroticism and low perceived control may represent a particularly strong risk for poor adaptation to negative events (Gerstorf et al., 2014). Identifying such constellations of psychosocial variables requires statistical methods that can accommodate many predictors without a predefined order of importance, and can explore potential interactions among them. To this end, we will use Structural Equation Model (SEM) trees – a statistical method combining theory-driven SEM with data-driven decision trees (Brandmaier et al., 2013). SEM trees start with a baseline SEM. In our case, the baseline SEM is a piecewise latent basis growth model of life satisfaction surrounding negative life events, capturing potentially non-linear anticipation and adaptation phases (as in Infurna & Luthar, 2017). The SEM tree algorithm then repeatedly fits this latent growth curve model to subsets of participants who share similar constellations of psychosocial characteristics measured prior to the event, identifying subgroups with distinct growth curve parameters – as indicated by improved model fit when this subgroup structure is introduced. This yields subgroups characterised by specific pre-event psychosocial profiles and distinct life satisfaction trajectories (as, e.g., in Brandmaier et al., 2017). Compared with latent mixture models or standard decision trees, SEM trees offer the advantages of using observed variables to define and describe subgroups that do not need to be pre-specified, and produce interpretable parameters (e.g., slope differences) describing subgroup trajectories. We will conduct separate SEM tree analyses for two events – disability and widowhood – as well as for a matched control group for each event, comprising older adults who did not experience the respective event. Namely, psychosocial variables may influence life satisfaction differently depending on whether a negative event is experienced, and if so, which event is experienced. For instance, high perceived control has been found to predict greater life satisfaction in the absence of adversity [Gerstorf et al., 2014], and in the context of disability [Joshi et al., 2025; Stewart et al., 2012], but to relate to poorer life satisfaction following widowhood [Bisconti et al., 2006; Specht et al., 2011]). Control participants will be matched to event participants on relevant covariates to ensure comparable baseline propensities for experiencing negative events (cf., Buecker et al., 2021). Life satisfaction trajectories of event participants will be modelled relative to the timing of the event. Trajectories of control participants will be centred around the corresponding time point of their matched event participant, as if they had experienced an event at that time point. This approach allows us to examine whether psychosocial variables show differential importance in the context of specific negative life events compared to their general associations with life satisfaction trajectories. Thus, the current study aims to identify combinations of psychosocial variables associated with distinct life satisfaction trajectories surrounding the onset of disability and widowhood in later life (here defined as 65 years and older), and to examine whether these combinations differ from those associated with general life satisfaction trajectories among matched controls. Candidate psychosocial variables include sociodemographic characteristics, personality dispositions, social participation indicators, and health and lifestyle factors. By uncovering psychosocial constellations linked to resilience or vulnerability surrounding negative life events, this study seeks to contribute to both theoretical understanding of successful ageing and the design of tailored interventions to support well-being in older adulthood.
Loading…
Imported from OpenAlex (CC0) — nothing here is generated.