Our relationships shape the trajectory of our lives, and the quality of relationships directly impact our sense of happiness and well-being. Relationships affect not only individual and family well-being but also mental and physical health and job performance. A romantic relationship can be the most all-encompassing and, ideally, will last a lifetime. What characteristics of individuals and their relationships matter most in relationship satisfaction? Can research help predict the future success of a relationship?
A large-scale longitudinal study set out to answer these questions. This study pulled data from 43 longitudinal datasets assessing 11,196 couples across 6 Western countries. These couples were assessed an average of 4 times across a period of four months. The longitudinal design allowed researchers to identify characteristics of relationships at baseline (before the study began) and track changes in relationship quality over time.
In the study, relationship quality was defined as “a person’s subjective perception that their relationship is relatively good vs. bad” (p. 19063). Researchers hoped to discover how relationship quality changed over time and what variables were most important in predicting relationship quality. Relationship quality was operationalized using two constructs consistently measured in relationship science: satisfaction and commitment.
The authors focused on two categories of variables that may strongly influence relationship quality: individual-difference factors and relationship-specific factors. Individual-difference factors, many of which can be considered demographic variables (age, education, religion, gender, race/ethnicity), vary from one person to the next. They encompass a range of constructs including personality, mental health, addictions, and even political beliefs. Relationship-specific constructs can be viewed from the context of interpersonal relationships and may include variables such as power, conflict, empathy, and affection.
Machine learning, or AI, was used to analyze the many variables under study. Statistical analyses lead to interesting discoveries. One’s own perceptions about the quality of the relationship (i.e., their satisfaction and commitment) proved more impactful than the partner’s perceptions of the relationship. One’s individual characteristics (e.g., personality, age, gender, mental health) were also predictive of relationship quality, but not as impactful as the relationship-specific factors. Further, self-report measures did not permit predictions about future relationship quality.
These results may seem surprising to anyone who believes individual characteristics can either enhance or detract from a happy marriage or couple relationship. Indeed, many decades of research on relationship satisfaction have focused on how individual qualities, such as personality traits or mental health, can contribute to long-term relationship satisfaction. However, this large-scale longitudinal study has called into question the importance of individual characteristics compared to our perception of the relationship.
In their concluding remarks, the authors explained: “Experiencing negative affect, depression, or insecure attachment are surely relationship risk factors. But if people nevertheless manage to establish a relationship characterized by appreciation, sexual satisfaction, and a lack of conflict—and they perceive their partner to be committed and responsive—those individual risk factors may matter little” (p. 19070).
Citation: Joel, S., Eastwick, P. W., Allison, C. J., Arriaga, X. B., Baker, Z. G., Bar-Kalifa, E., … & Wolf, S. (2020). Machine learning uncovers the most robust self-report predictors of relationship quality across 43 longitudinal couples studies. Proceedings of the National Academy of Sciences, 117(32), 19061-19071.
©Jennie Dilworth, Ph.D