Marital quality over the life course: a hierarchical linear model of duration and cohort effects

Abstract

Considerable debate over the trajectory of marital quality over the marital course has encamped into two major schools of thought: U-shaped or linear decline. Access to longitudinal data that extends into the later years of marriage and analytical techniques that allow tracking of changes in marital quality is often cited as a remedy for the limitations that plague much of this research. The sample, from the UCLA 1971-1997 Longitudinal Study of Generations, consisted of six waves of data on marital quality from three separate historical cohorts in intact, first marriages. Analyses were conducted using hierarchical linear modeling, a technique particularly suited for analyzing change across time in panel data, to determine duration and cohort effects on marital quality in marriages ranging from one to 69 years. The cohort married during the years 1945-1954 exhibited the familiar U-curve of both positive and negative marital quality. Results for the youngest cohort, married between 1964 and 1984, replicated similar research for a linear decline in positive marital quality and an increase in negative marital quality. While cohort effects were inconsistent, this study, unlike other longitudinal studies, replicated the U-shaped curve of marital quality found in previous cross-sectional studies in a cohort followed from the third through fifth decades of marriage.