Dyadic Processes Underlying Loneliness and Risk of Divorce in Older Age
Loneliness in later life is frequently framed as an individual-level public health concern, associated with elevated risk for depression, morbidity, and mortality. Yet a substantial proportion of lonely older adults are married or partnered. This raises a critical relational question: how does loneliness operate within ongoing intimate partnerships, and can it shape long-term relationship stability?
In the MRI-funded project Dyadic Processes Underlying Loneliness and Risk of Divorce in Older Age, Dr. Jeffrey Stokes (University of Massachusetts Boston, in partnership with the University of Massachusetts Foundation) examines whether loneliness—experienced by one or both partners—predicts risk of “gray divorce,” defined as divorce occurring after age 50. As gray divorce rates have risen over recent decades, understanding relational processes that precede dissolution has growing demographic and clinical relevance.
Conceptually, the project advances a dyadic model of loneliness. Rather than treating loneliness as a purely intrapersonal emotional state, the study positions it as an interactional process that may reshape couple dynamics over time. Persistent feelings of disconnection can influence how partners interpret each other’s behavior, seek or withdraw support, and respond to everyday bids for closeness. In this framework, loneliness may erode relational quality not only through subjective distress, but through cumulative micro-level changes in responsiveness, warmth, and perceived mutuality.
To test these hypotheses, the project draws on longitudinal data from the Health and Retirement Study (HRS), a nationally representative panel study of older adults in the United States. Couples are followed over approximately 15 years, allowing for prospective modeling of whether one partner’s reported loneliness predicts subsequent marital dissolution or separation. The longitudinal design strengthens causal inference by establishing temporal ordering between loneliness and divorce risk.
Importantly, the study employs a dyadic analytic approach that accounts for interdependence within couples. Actor–partner models enable examination of both actor effects (how an individual’s loneliness predicts their own marital outcomes) and partner effects (how one partner’s loneliness predicts the other partner’s outcomes and shared relationship stability). This framework allows the research team to assess whether loneliness is “contagious” within couples and whether asymmetries—such as one partner feeling lonely while the other does not—carry distinct risks.
The project also investigates potential gender differences in these processes. Prior research suggests that men and women may differ in how they experience and express loneliness, as well as in their reliance on spousal support. By testing whether loneliness operates differently for husbands and wives, the study aims to clarify not only whether loneliness predicts gray divorce, but also how these pathways may vary by gendered relational roles.
The translational implications are significant. If loneliness functions as a shared relational stressor that accumulates over time, then interventions targeting older adults may benefit from focusing on dyadic connection rather than individual symptom reduction alone. Preventive efforts could emphasize strengthening responsiveness, increasing shared meaningful activities, and addressing emotional disengagement before patterns of distance become entrenched.
By situating loneliness within the interpersonal system of marriage, this project contributes to a broader shift in aging research—from viewing emotional well-being solely as an individual outcome to understanding it as embedded within enduring relational structures. In doing so, it offers a more nuanced account of how subjective disconnection can shape long-term partnership trajectories in later life.